Split (1)

train · ~9.11M rows (showing the first 93.5k)

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W2613806235	https://journals.umcs.pl/k/article/download/3686/3619	English	null	A contemporary religious political movement: LDS Church’s nation-wide political campaigns	Annales Universitatis Mariae Curie-Skłodowska. Sectio K, Politologia	2,017	cc-by	7,920	A N N A L E S UNIVE R SITAT IS MAR IAE C UR IE -SKŁ ODOW SKA LUBLIN – POLONIA . XXIII, 1 SECTIO K 2 2016 VOL. XXIII, 1 VOL. XXIII, 1 aculty of International and Political Studies, Department of Political Systems, University of Łódź, Poland ABSTRACT Drawing empirically on the examples of the Church of Latter-Day Saints’ (Mormon) participation in anti-ERA (Equal Rights Amendment) ratification campaign and its latest attempts to influence other polit ical issues of moral consequences (such as same-sex marriage) in the United States, the paper attempts to analyse the dynamics of a contemporary religiously motivated political movement. Despite being, on any of these issues, a part of a wider coalition of political actors, the Mormon church displayed a specifically religious motivation, justification and modus operandi. Owing to strong religious legitimacy of their power – based on the doctrine of continuous revelation and enhanced by a sort of “personality cult” of the Church President-Prophet developing in late 20th century – the leadership of the church has been capable of effective grassroots mobilization, achieved through a disciplined universal priesthood structure. While, from the theoretical point of view, this Mormon political movement is of a traditional, “old” variation (ideological and social cohesion of members, well-defined, stable membership, hierarchical lead ership, etc.) it has nevertheless been relatively successful in modern political environment. The Mormon engagement, at least in the anti-ERA campaign, had made a difference certainly in Utah, and probably elsewhere as well. Key words: Mormonism, political movements, religion and politics, Equal Rights Amendment, same-sex marriage MACIEJ POTZ A Contemporary Religious Political Movement: LDS Church’s Nation-Wide Political Campaigns DOI:10.17951/k.2016.23.1.147 DOI:10.17951/k.2016.23.1.147 INTRODUCTION Ever since Joseph Smith’s bid for presidency in 1844, Mormonism has been engaged in American national politics to some extent. Throughout the 19th century, this engagement centred around gaining statehood for Mormon-governed Utah while 148 Maciej Potz defending the practice of polygamy and the religious-political idea of a theocratic Kingdom of God [Hansen 1967: chapter VII and ff]. When this effort failed and Utah was admitted to the Union as a state in 1896, only after the Church of Jesus Christ of Latter-Day Saints (LDS) gave up polygamy and its political monopoly, Mormonism has moved, in the 20th century, into the mainstream of American political life, usually remaining on its conservative side. In the second half of the 20th century political activity of the LDS Church focused on specific issues deemed to be of importance (usually, though not exclusively, for moral reasons) for Mormons, rather than on constant support for any of the two major political parties. The Church’s political actions on two such issues – the ratification of Equal Rights Amendment (ERA) and the legality of same-sex marriage – will be analysed in the article as examples of its activity as a political movement. Mormonism itself can be interpreted as a religious movement of which the Church of Jesus Christ of Latter-Day Saints (with headquarters in Salt Lake City, Utah) is the largest, though not the only, organization. To qualify as a political move ment, it has to be an intentional mass effort towards transformation of some aspect of social reality by political means. As we shall see, the Mormon campaigns (the anti-ERA campaign will be our special focus) fit this definition sufficiently: they were (the anti-gay marriage campaign still is) well-planned and coordinated actions, involving numerous church members, intending to stop potentially threatening le gal and ultimately social developments, using a whole repertoire of political tools. Unlike modern social movements (i.e. alterglobalism), characterized by informal, horizontal networks of members, who share some collective identity [della Porta, Diani 2006: 20–21] but may not form a distinct and integrated social group, Mormon political activity represents the “old”, traditional type of political movements: based on a distinct, cohesive social group, acting through a formal organization, headed by hierarchical leadership. To describe the LDS Church, in the context of these cam paigns, as a political movement, is not to imply that politics is the defining feature or raison d’etre of the Church. PUTTING THE ANTI-ERA CAMPAIGN IN THE CONTEXT Equal Rights Amendment is a proposed amendment to the United States Consti tution which read, in its 1972 version: “Equality of rights under the law shall not be denied or abridged by the United States or by any State on account of sex” (section 1). Its declared purpose was to extend equality of rights for women beyond voting rights guaranteed to them in the Nineteenth Amendment. ERA failed when intro duced in Congress for the first time in 1923 and on a few subsequent occasions, but passed both houses in 1972. To become law, it needed the ratification of at least three fourths or 38 states. Initially, the ratification process proceeded smoothly, reaching 35 states in 1977. But no other state approved the amendment before the deadline of 1979, later extended by Congress to 1982, and five states rescinded their previous ratification [Mansbridge 1986: 12–13]1. The ultimate failure of ERA, despite the support of the majority of American population, is largely attributable to an intense campaign weighed by a variety of conservative political and religious organizations that managed to counter the pro-ratification campaign of feminist and other left-wing or liberal organizations. It is important to recognize that, when fighting ERA, the Mormon church was a part of a large coalition of conservative forces, both secular and religious. The seventies of the 20th century saw the emergence of a broad political movement of Protestant fundamentalist and evangelical churches known as Religious Right2. The hitherto politically inactive members of these churches have, for a variety of rea sons, engaged themselves into politics through organizations such as Jerry Falwell’s Moral Majority, James Dobson’s Focus on the Family or, later, Pat Robertson’s Christian Coalition [Potz 2008: 159–183]. Whether Mormonism has been a part of Religious Right is disputable (if only because Mormons are not Protestants and many Protestants and Catholics even refuse to recognize them as Christians), but formal identification is of little consequence here. The crux of the matter is that the LDS Church has acted in concert with Religious Right organizations – notably Beverly LaHaye’s Concerned Women of America or the Catholic Phyllis Schlafly’s Eagle Forum [Wilcox 1996: 65–66] – on a limited number of issues important to each of these groups, including ERA and same-sex marriage. 2 For general works on Religious Right see C. Wilcox, Onward Christian Soldiers. The Religious Right in American Politics, Boulder: Westview Press 1996; W. Martin, With God on Our Side. The Rise of the Religious Right in America, New York: Broadway Books 1996. 149 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 149 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 149 1 Whether the rescissions had any legal effect is controversial (Mansbridge thinks not), but in any case they certainly, as she correctly observes, had a political effect on the state which had yet to ratify. INTRODUCTION On the contrary, for most of its leaders and members alike, it is primarily a religious organization, offering supernatural rewards to its members. Such designation is, rather, an analytical device to highlight a particular (political, in this case) aspect of the organization’s activity.i The purpose of the article is, first, to present the Mormon anti-ERA campaign against the background of conservative religious groups’ political activity, looking at strategies and methods used by the Church, and then, second, try to explain the LDS’ extraordinary mobilization capacity and political efficiency. I contend that what made a difference in comparison with other, not only secular but also religious participants of the ERA battle, was exceptionally strong legitimacy of the church leadership power grounded in the doctrine of continuous revelation which elevates the President of the Church to the position of the living Prophet. The article concludes with observations on recent Mormon political activity related to the issue of same-sex marriage, suggesting striking similarities with, but also some differences from, the anti-ERA campaign. 1 Whether the rescissions had any legal effect is controversial (Mansbridge thinks not), but in any case they certainly, as she correctly observes, had a political effect on the state which had yet to ratify. 2 For general works on Religious Right see C. Wilcox, Onward Christian Soldiers. The Religious Right in American Politics, Boulder: Westview Press 1996; W. Martin, With God on Our Side. The Rise of the Religious Right in America, New York: Broadway Books 1996. 3 Up to this point Mormon legislators voted according to their conscience, both locally and nation ally. Some of them supported ERA in the U.S. Congress, without adverse reaction from the Church. Ibid., p. 375, 377. PUTTING THE ANTI-ERA CAMPAIGN IN THE CONTEXT Such interfaith alliances based on moral common ground while temporarily, at least, disregarding theolog ical differences, have been characteristic for, and certainly one of the reasons of political effectiveness of, religiously motivated conservative politics since roughly the seventies. They are also illustrative of a general trend in American religion to align along ideological, not only denominational lines [Williams 2002: 343–345]. 150 Maciej Potz The conservative-liberal divide cuts across religious traditions, leaving Southern Baptists and Missouri Synod Lutherans, for example, on the opposite sides of the cultural battlefront from their liberal coreligionists from American Baptist Churches USA and Evangelical Lutheran Church in America, respectively. Mormonism is not itself utterly free from such divisions (there is, for example, a considerable number of liberally-leaning Mormons), but, compared to other denominations, it has been able to maintain extraordinary unity when acting on a few crucial matters. I will try to explain this phenomenon in a later section, and now let us turn to a description of the Church’s involvement in the anti-ERA campaign. 151 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 151 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 151 Modus operandi of those church-related organizations was typical for grass-roots lobbying, as opposed to traditional, direct lobbying. Thousands of Mormon men, but mostly women (an important public relations aspect to show that women them selves, and not their husbands, naturally interested in keeping them in submission, rejected the amendment) in states where ratification was pending – such as Nevada, Florida or Virginia – were asked to write letters to state legislators; participate in demonstrations and rallies; sign anti-ERA petitions; donate and raise funds; distribute literature to and personally contact their non-Mormon neighbours, especially in cases where, like in Nevada, the decision on ratification was to be made in a referendum; vote in a referendum or critical elections whose results might determine the fate of ERA in state legislatures. A particularly telling example of the LDS Church’s mobilization abilities was Mormon participation in International Women Year’s state conferences. The pur pose of these events, organized under the auspices of United Nations and the U.S. President, was, first, to elect state delegates to the national conference in Houston in November 1977 and, second, to discuss and vote on a number of issues related to women, from social security and medical care to rape laws and pornography to lesbian rights [Martin 1996: 164–165]4. Primarily, however, the conferences were perceived as expressing women’s attitude to Equal Rights Amendment [Young 2007: 632]. The LDS leaders, initially intending to boycott these seemingly feminist events, decided instead to take them over. In Utah, Mormon women, on the call of their priests, packed the Salt Lake City conference centre and, outnumbering non-Mormons approximately thirteen thousand to one thousand, voted down all resolutions, related to the ERA or not, and elected almost exclusively Mormon delegation to the national conference [Young 2007: 635]. Similar efforts in other states proved somewhat less effective, if only for demographical reasons, but Mormon participants could often influence the outcomes due to excellent coordination (e.g. moving en masse from committee to committee whenever a crucial vote was to be held). It is important to note that most of the above described political actions were more or less directly organized by the LDS Church and through its structures. 4 Martin argues that the inclusion and passage of the three “hot-button issues”: abortion, ERA and lesbian rights during the Houston conference spurred the strongest opposition from conservative organi zations. ERA: THE STRATEGY OF ACTION For more than two years after the ERA enactment in Congress, the LDS Church remained virtually silent on the matter, despite over thirty ratifications that occurred during this period. It only launched its campaign in early 1975. This is largely ex plainable by a change in the post of Church President, its highest leader: conservative but non-committal and unwilling to engage politically Harold Lee died in December 1973 and was succeeded by more vigorous and very popular among the rank-and-file Spencer Kimball [Quinn 1997: 376]. In January 1975, the “Church News” section of “Deseret News”, LDS official newspaper, published an editorial expressing the Church stand against ERA [Equal Rights Amendment 1975: 16]. This well-timed move was sufficient to defeat ratification of the amendment in Utah legislature in February, all Mormon members voting against. The impact of the Apostle’s statement is obvious: Mormon legislators who earlier supported the ERA now switched their votes, explicitly invoking the Church position as justification3. Once this position was announced, the phase of direct political action, in and outside of Utah, followed. p p Organizationally, the anti-ERA campaign was coordinated by Special Affairs Committee including high LDS officials, rather than any of the governing bodies of the church (First Presidency, Quorum of the Twelve Apostles, Quorum of the Seventies of Presiding Bishopric). This was probably to conceal the Church’s direct involvement. For the same reason, in several states the Church acted through or ganizations ostensibly not related to it (and bearing no mention of Mormons in their names), but actually established or led by its functionaries. They were not financed directly from LDS Church coffins, but through donations of individual members, strongly recommended by the leaders. These organizations often used church prem ises and, most importantly, Mormon human resources. 5 Technically, franchise for women was introduced first in the territory of Wyoming two months earlier, but Utah held first elections in which women could vote [Bigler 1998: 283]. While this was, admit tedly, a strategic move by Brigham Young to counter some of the bad publicity the Church was receiving in the midst of the anti-polygamy campaign, Mormons have never since had any second thoughts about it. The crux of the issue was how Mormon women vote, not if they should vote at all. Indeed, they almost always voted along with their male coreligionists and, generally, according to the Church positions. 151 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… Mem bers were “asked”, “called” or even “set apart” for the anti-ERA activities [Kendall White 1989: 256], the last phrase in Mormon parlance being roughly equivalent to ordaining to a priestly office, which obviously added religious legitimacy to the function. Bishops were asked by an Apostle coordinating from LDS headquarters to send at least ten women from each ward [Quinn 2000: 379] (the basic ecclesiastical territorial unit in the Church) to the Utah IWY conference. Instructions and calls for action regarding ERA were read during church meetings and church telephone-trees were used to mobilize members. Even Mormon missionaries might occasionally be 152 Maciej Potz employed in anti-ERA activities [Quinn 2000: 386]. Clearly, “The Mormon campaign throughout the nation depended upon the ecclesiastical structure of Mormonism – the social networks, the Relief Society [Mormon women organization – M.P.], the coordination and tacit support of top officials, the local facilities and the local and general church media. It could not have occurred without the support of the Mormon hierarchy” [Kendall White 1989: 257]. employed in anti-ERA activities [Quinn 2000: 386]. Clearly, “The Mormon campaign throughout the nation depended upon the ecclesiastical structure of Mormonism – the social networks, the Relief Society [Mormon women organization – M.P.], the i employed in anti-ERA activities [Quinn 2000: 386]. Clearly, “The Mormon campaign throughout the nation depended upon the ecclesiastical structure of Mormonism EXPLAINING MORMON POLITICAL INVOLVEMENT Having established an intense commitment in anti-ERA political activity on part of both the Latter-Day Saints Church as an institution and many of its members, let us now inquire into the reasons 1) why would Mormons so vehemently oppose Equal Rights Amendment in the first place and 2) why they were so effective in mobilizing against it. i As concerns the first question, many arguments have been advanced against the Amendment. The LDS Church has not opposed basic civil rights for women; in fact, Mormon-governed Utah was, in 1870, the first United States territory or state to allow women to vote5. However, the ERA was feared to accomplish much more than this basic legal equality. From the legal standpoint, it could actually nullify some privileges obtained by women in areas such as child care or social security. Even more troubling for Mormons were its potential social consequences: in refusing to acknowledge natural differences between men and women, it could lead to gender-blind society, succinctly summarized under the slogans of unisex toilets and women on the battlefronts [Kintz 1998: 132]. Still worse, by “stifling many God-given feminine instincts”, an official First Presidency statement warned, “It would strike at the family, humankind’s basic institution” [First Presidency Issues Statement Opposing Equal Rights Amendment 1976]. The ERA threatened to question traditional male and female roles within family, with men failing to provide and women failing to care for their families. While these dire consequences may look a far-fetched misinterpretation of an innocuous proposition of legal equality of sexes, the range of possible interpre tations of such a general clause in American constitutional law is almost infinite. In American political system, it must be remembered, the courts, with federal Su preme Court as the ultimate instance, exercise judicial review, i.e. the right to control the constitutionality of laws. In the process, they are entitled to determine the content 153 CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 153 and meaning of constitutional rules as applied to legal conflicts under scrutiny. This involves them in actual law-making, especially when the interpretation assigned to a constitutional norm is very removed from the literal content of the Constitution. EXPLAINING MORMON POLITICAL INVOLVEMENT It is due in part to the very laconic, general and archaic character of the Constitution (many present-day problems were, naturally, not envisaged in the 18th century) and in part to the activist legal philosophy of some judges, who understand their mission as creative application of these over two-centuries old “majestic generalities and ennobling pronouncements”, as former United States Supreme Court justice William Brennan once called them [Brennan Jr. 1993: 55]. In any case, the defenders of “tra ditional family” were not totally unjustified in fearing unpredictable applications of the Equal Rights Amendment, especially in the wake of a recent 1973 Roe v. Wade [410 U.S. 113] decision, in which, from the conservative perspective, the Supreme Court inferred the right to abortion (non-existent in the Constitution) from the right to privacy (also non-existent in the Constitution)6. What could be interpreted from the new amendment, only time would tell. Crucially, even though all of these arguments against ERA were advanced by other members of the conservative anti-ERA coalition, as well, Mormons had special reasons to fear disintegration of a traditional model of family and the feminine and masculine role definitions within it. These reasons were not merely traditional but the ological and, even more importantly, eschatological. According to Mormon beliefs, an individual’s fate in the afterlife is directly related to his or her marital status. Mormon heaven is divided into three hierarchically arranged spheres, the celestial (highest), terrestial and telestial (lowest) kingdoms. While even all but the worst sinners can get to the telestial sphere, the highest degree of glory is reserved, in principle, only for the righteous Mormon couples [Riess, Kimball Bigelow 2005: 35–37]. Singles are at a distinct disadvantage, as the deceased join or are joined in heaven by their family members and continue eternally with their earthly spouses, having been sealed to each other for eternity in a temple ceremony. Faithful Mormons who are married with many children are more likely to progress to a godly status and inherit their own heavenly kingdoms. Therefore, “Told that the ERA would eradicate the basic distinctions be tween the sexes and loosen men and women from the gender-based obligations of marriage, Mormon men and women opposed ERA because it contradicted their most fundamental beliefs about the nature of both life and the afterlife” [Young 2007: 631]. 6 Further impact of the Roe v. Wade decision lied in the perception of state legislators, who were about to decide on the ERA ratification, that adding an amendment to the Constitution takes significant amount of political power away from them and into the hands of Supreme Court justices, who may then use it the way they did in Roe [Mansbridge 1986: 13]. EXPLAINING MORMON POLITICAL INVOLVEMENT This is certainly the sort of a motivation that the secular antagonists of the ERA and most religious ones, too, are lacking. It also provides a contemporary vindication for Max Weber’s insistence on the relation between a religion’s vision of salvation and social behavior of its members [Weber 1978: chapter 6, section vii and ff]. 154 Maciej Potz This brings us to the other question posed at the beginning of this section. While this doctrinal rationale helps explain the sincerity and intensity with which many devout Mormons felt about the issue, such feelings need not necessarily translate into effective political mobilization. Moreover, not all Mormons perceived the con tent of the Equal Rights Amendment as a threat to those important family values. As a matter of fact, over 50% of them approved (while 38% disapproved) of the statement: “Equality of rights under the law shall not be denied or abridged by the United States or by any State on account of sex” when it was read to them without the mention that it is actually the proposed ERA7. There was even some internal dissent within the Church (pro-ERA Mormon women established MERA – Mormons for ERA and its leader Sonia Johnson, an excommunicated Mormon feminist, gained some publicity), although its significance was marginal, both for the Church and the campaign8. The bulk of the believers, initially, had not have any definite position on or even knowledge of the issue. They followed the Church’s lead despite having been unaware of the Equal Rights Amendment at all prior to the start of the LDS engagement [Young 2007: 624] or, when informed, still failed to grasp the poten tially detrimental consequences of the amendment. So why did they respond to their Church’s call for action in such a disciplined way? Trivial as it may seem, the simplest answer would be: because they were Mor mons. As Campbell and Monson demonstrated, there exists a statistically significant correlation between religious participation and political engagement. The more someone is involved in the life of their church (as measured by church attendance, participation in non-worship church activity, charity giving, etc.), the more this person is likely to be politically involved, too. While this is true for some other religious denominations, as well, the correlation is strongest in the case of Mormons [Camp bell, Monson 2007: 119–120]. 8 It tends to be overestimated by the critics of LDS Church engagement, such as Kendall White [1989: 258–260]. Suffice it to say that MERA had a little over a thousand members nationally at the peak of its popularity, while the church hierarchy managed to mobilize some thirteen thousand Mormon women for a single event (IWY conference in Utah) and many more nationwide [Young 2007: 635]. 7 The survey was conducted among Utah Mormons in 1980. When asked about the passage of the ERA, three-quarters were against [Kendall White 1989: 261]. EXPLAINING MORMON POLITICAL INVOLVEMENT One likely rationale for this is that churches with strict standards of behavior, demanding a lot of commitment from their members – and LDS Church clearly qualifies, with its dietary prohibitions, plenty of time spent in church-related activities, ten percent tithing and a degree of cultural distinctiveness which may result in relative social isolation, especially outside of Utah – are nat urally well-poised to overcome the free-riders problem, a classic dilemma of any collective action. The insiders are inclined to accept the Church’s call for action, just one more among their many commitments. Additionally, everyday involvement within the church helps members develop certain civic skills such as organizing and attending meetings, making presentations, writing letters, taking responsibility for various tasks, however small, etc. These skills, along with social capital built through 155 CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 155 intense contacts within the church environment, are instrumental in effective political activity [Campbell, Monson 2007: 115–116]. These structural conditions, characteristic of a culturally distinct, tightly-knit religious group, provide important background for the LDS success as a political movement. The most compelling single explanation, however, of the rank-and-file obedience to the Church’s directives, including those with political context, lies in the strong legitimacy of the Mormon leaders’ power. The sheer religious character of this legitimacy – the fact that it is grounded in supernatural justifications – puts religious rulers at a certain advantage over secular ones: unlike the latter, the former are brokers of supernatural rewards, often of ultimate value to a true believers, such as blessings, significant rituals, etc., but most of all – eternal salvation. While this may account for the edge the Mormons have in comparison with sec ular organizations whose members neither fight for religious ends nor are answerable to administrators or providers of religious rewards, it does not in itself explain their greater capacity for concerted action than other religious groups have. This concerns not only liberal churches from the other side of the cultural divide, but also their allies from the conservative camp, such as Southern Baptists or Roman Catholics. Here the difference lies, I argue, in the special position of Mormon church leaders grounded in the doctrine of continuous revelation. Nearly all major Christian denominations, including Catholics and most Protestants, consider revelation a finished process which began with God’s theophany in the Garden of Eden and ended with Christ’s sacrificial death. 9 Although the term is not accepted in Mormon theology nor used by the leaders, a de facto infalli bility of the President and, to a lesser extent, the Apostles, is a prominent feature of late 20th and early 21st century LDS church. See W. van Beek [2011: 20–22]. EXPLAINING MORMON POLITICAL INVOLVEMENT The church, in the course of its history, can only strive to better understand and refine this revelation, but cannot receive any new one. By contrast, Mormons, along with some other heterodox Christian groups, believe that God constantly guides His people through history. This is done by revealing His will to the church leaders – to a group containing, at least, the First Presidency and the Apostles, with the President’s prophetic role being the most prominent. Accordingly, the President and the Apostles have been accorded the title of “prophet, seer and revelator”. One of the Church’s scriptures, Doctrines and Covenants, is a series of revelations received mainly by Joseph Smith, and most significant decisions in the LDS Church history (such as introduction and renunciation of polygamy or, as recently as 1978, opening the priesthood to non-white members) [“Official Declaration 2”] were prompted or approved by divine communication [Potz 2013: 425–430]. The view of the Church Presidents as, effectively, infallible prophets9 acting on God’s command has been perpetuated by the hierarchy. Apostle Ezra Taft Benson went as far as saying that “The living prophet is more vital to us than the stand ard works” [Benson 1980], where “the standard works” mean the scriptures of the Church. This seemingly radical statement fits well with the logic of active God 156 Maciej Potz leading His flock through history, constantly adding to and even revising earlier revelations (including Mormonism’s holy books). As proclaimed in the groups’ Ar ticles of Faith, “We believe all that God has revealed, all that He does now reveal, and we believe that He will yet reveal many great and important things pertaining to the Kingdom of God” [Articles of Faith]. The truth of the revelations is not to be doubted, for, as the fourth Church President Wilford Woodruff declared,“The Lord will never permit me or any other man who stands as President of this Church to lead you astray” [Woodruff 1890]. The briefest statement of this doctrine, however, was given, in the time (though not directly in the context) of the ERA campaign, by the Church’s Young Women organization president Elaine Cannon: “When the Prophet speaks, sisters, the debate is over” [Cannon 1978]10. It was later endorsed by First Presidency First Counsellor (second in command after Church President) Eldon Tanner [Tanner 1979]. 10 She also called President Kimball “the mouthpiece of the Lord” (access 12.08.2014). 12 There is also a first amendment free-exercise clause concern of a church’s or religious function ary’s right to refuse to bless a same-sex marriage, but the problem has never been real. Typically, laws legalizing same-sex marriage do not require a minister to perform it, but merely recognize it if a minister chooses to do so (see R. Crapo, Chronology of Mormon/LDS Involvement in Same-Sex Marriage Politics, http://www.mormonsocialscience.org/wp-content/uploads/2009/11/Crapo-R1997-Chronology-of-LDS- Involvement-In-Same-Sex-Marriage-Politics.pdf (access 17.08.2014) for a Hawaii Circuit Court Ruling to that effect). Conversely, these statutes do not confer on same-sex couples a right to have religious mar riage ceremony. 157 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 157 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 11 However, D. M. Quinn documents a pervasive homophobia inside the LDS church [Quinn 2000: 36 and ff]. i EXPLAINING MORMON POLITICAL INVOLVEMENT i What is more, after being downplayed in the first half of the 20th century, this prophetic designation of Church Presidents has again been highlighted since the late 1960’, amounting to what D. Michael Quinn likens to a personality cult [Quinn 1997: 363–366]. Since then the President has routinely been referred to as the Liv ing Prophet in official Church publications. Importantly, however, the continuous revelation doctrine and the resulting position of the Church top official is willingly accepted by the majority of Mormons, too, as a fundamental of their faith, not some thing imposed from above. In the case of the ERA, Mormons who took anti-ERA stance despite having no interest in the matter previously, or even switched their position completely, openly admitted they did it because of the supreme knowledge or ability to predict the future by the Church leaders and especially the President [Gordon, Gillespie 2012: 348; Quinn 1997: 377]. In sum, the 19th century Mormon territorial theocratic political system has turned into the 20th and 21st century internal church theocracy with extremely strongly legitimized power of its rulers. Few other religious leaders, especially of groups of comparable size, enjoy such a legitimacy – not even hierarchs of Roman Catholic church, a structurally similar organization. Effectively, in contemporary Mormonism, “adherence to the prophet’s instructions in all matters is a hallmark of Mormon religious observance, including with regard to political questions” [Campbell, Monson 2007: 121]. Lastly, not only is the LDS Church a hierarchical organization, but it is based on a principle of non-professional universal male priesthood. All worthy boys at the age of 12 are ordained to the lowest priestly function and then progress up the ecclesiastical ladder well into their adult years [O’Dea 1964: 175–176]. As a con sequence, they perform all their assignments within the church in their capacity as priests, which adds religious significance – a sense of mission or calling – to even the down-to-earth, everyday tasks. MORMONS AND THE LEGALIZATION OF SAME-SEX MARRIAGE The Church issued a series of statements, usually as a response to or a warning against some unwelcome developments (enactment of legalizing 158 Maciej Potz legislation, court ruling upholding it, etc.), explaining and justifying its position and urging Church members to act. For instance, First Presidency issued a letter to California Mormons to support with their time and money “Proposition 8”, a pro posed amendment to the state’s constitution outlawing same-sex marriages in 2008 [Moore 2008]. In supporting a similar amendment to the federal Constitution in 2006, the leadership went to the lengths of having its letter read from all Mormon church pulpits [Perry, Cronin 2012: 97]. The members have been mobilized through Church channels, but, again familiarly, the hierarchy has usually preferred to disguise its direct involvement, acting through organizations or political action committees with out the Church name of affiliation13. The Saints have often constituted a significant proportion of activists in such organizations, engaging in door-to-door canvassing, letter-writing, leaflet distribution, demonstrating, fundraising, etc. There are also, however, differences between the two campaigns. In the case of ERA, the goals were fairly well-defined and localized (to stop ratification by the legislature or referendum in respective states). Same-sex marriage campaign required more nationally-oriented efforts, since the federal government is a part to the issue, as well, and it was waged on multiple fronts. For example, since not only legislative, but also judicial decisions matter, the Church has tried to act as amicus curiae or otherwise involve itself in court proceedings. Also, it naturally makes extensive use of the Internet, both for publicity and fundraising [Gordon, Gillespie 2012: 359–360]. p y g [ p ] In June 2015 the divided (5–4) United State Supreme Court issued a landmark decision which declared a right to marry for same-sex couples a constitutional right, thereby making all state legislation banning such marriages unconstitutional [Obergefell v. Hodges (576 US. 14–556), 2015]. From now on, same-sex marriage must be allowed and recognized in the whole territory of the United States. 13 In one case, the same organization – Hawaiian Hana Pono – was used in both campaigns (Crapo, Chronology). As an aside, this practice has been criticized, in both campaigns, as of dubious legality, but to lose its tax-exempt status a religious organization would have to endorse a specific candidate or party, and not just support or oppose a political issue. The LDS Church has consistently avoided such an endorsement. MORMONS AND THE LEGALIZATION OF SAME-SEX MARRIAGE The legalization of same-sex marriage is the other major political issue in which the Church of Jesus Christ of Latter-Day Saints chose to engage. Its involvement dates back to the mid-90s, when controversies over the issue began in several states, notably Hawaii. The LDS campaign against homosexual marriage displays many similarities and even continuities with the anti-ERA campaign, in terms of the Church’s justifi cation and the members’ motivation for, as well as the strategy of, political action. Doctrinally, opposition to same-sex marriage is rooted in the same traditional, eschatologically grounded conception of family we have seen in the case of ERA. In a recent statement of January 2014 the First Presidency reminded: “Marriage between a man and a woman was instituted by God and is central to His plan for His children and for the well-being of society. Strong families, guided by a loving mother and father, serve as the fundamental institution for nurturing children, instilling faith, and transmitting to future generations the moral strengths and values that are impor tant to civilization and crucial to eternal salvation” [italics added – M.P.] [Church Instructs Leaders on Same-Sex Marriage 2014]. While homosexual behaviour (as opposed to homosexual feelings or attitudes) is in itself regarded sinful, at the core of the LDS Church’s opposition to gay marriage lay its eschatological ramifications. Indeed, the Church is motivated by “more than just a belief in the deleterious effects on society, or the sinfulness of the individual behaviour. Gay marriage is seen as the utter destruction of God’s plan for humanity” [Perry, Cronin 2012: 94]. Interestingly, the Church came to grudgingly accept, if only for tactical reasons, legal recognition of various kinds of civil unions or domestic partnerships, not only homosexual, and their rights in spheres like medical care, hospitalization, fair housing [The Divine Institution of Marriage 2008; Moore 2008], etc., reserving its opposition to defining marriage as other than between a man and a woman. It also displays an image of not being anti-gay, but only seeking to protect the traditional model of family11. In moral terms, at stake is the sanctity of the God-given and es chatologically privileged union between husband and wife12. In terms of methods, essentially the same patterns were followed as in the an ti-ERA campaign. CONCLUSIONS CONCLUSIONS The analysis of the Church of Jesus Christ of Latter-Day Saints’ involvement in the issues of Equal Rights Amendment ratification and the legality of same-sex marriage reveals that, in both campaigns, the Church has, functionally, acted as a po litical movement. It consisted of a large group of highly mobilized members under an efficient leadership, determined to achieve certain goals. The mass character of the phenomenon and the genuine motivation of its participants justifies its categorization as a political movement, rather than pressure or interest group. i The strategies used by the Church were, to a significant extent, typical for Amer ican political processes, especially in the context of religiously-motivated conserv ative politics of the last decades of the 20th century. They were based on grass-roots lobbying (letter- and petition-writing, demonstrating, rallying, etc.) aimed at states legislators and the electorate at large (especially when referendums were held). At times the Church was able to use its human resources even more creatively, as in the case of mass participation in International Women Year conferences. An important part of the strategy were tactical alliances with other conservative groups, mainly Protestant churches belonging to the Religious Right and the Roman Catholic church. Although it is impossible to precisely determine the extent to which Mormons influenced the outcomes of the political campaigns they were engaged into, or, spe cifically, whether they “tipped the scale” in favour of the ERA rejection, it is safe to conclude that their influence was disproportionately large compared to their share of population. Consider, for instance, that, in states where Mormons constitute less than 1% of population, some 85% of letters flooding legislators’ offices were written by them [Quinn 2000: 13]; the turnout among the Nevada Mormons for the ERA ratifi cation referendum reached 95% [Young 2007: 637], similarly to other referendums on the issue, while national average was only 53.2% [Quinn 1997: 385–386]; and in California, some organizations fighting legalization of same-sex marriage were staffed predominantly by Mormon volunteers [Gordon, Gillespie 2012: 356]. More over, the final political or legal outcome should not be regarded the sole criterion. In the case of same-sex marriage, although the cause of preventing its legalization was ultimately lost, the campaign consisted of a series of battles (referenda, court cases), which, whether or not successful, clearly demonstrated the mobilization potential and political efficiency of the Mormon Church. 159 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 159 MORMONS AND THE LEGALIZATION OF SAME-SEX MARRIAGE In response, the First Presidency issued a statement, to be read to all the members, affirming the position of the LDS Church on the issue: as a matter of doctrine, all sexual relations outside of a valid marriage between a man and a woman are sinful; as a matter of policy, some benefits can be granted to same-sex unions but without elevating them to the status of marriages. At the same time, the Church asserts its right to refuse to solemnize such unions and publicly express their opinions on the subject [Council of the First Presidency and Quorum of the Twelve 2015]. Politically, the Supreme Court decision very likely closes the political campaign of the Church, unless it chooses to work actively for the reversal of the ruling in some future favourable circumstances. This, however, seems a remote possibility at the moment. CONCLUSIONS We have put forward several explanations of such LDS Church’s political ef fectiveness. They fall into two broad categories that can be shortly described as le gitimation and organization. First, from the point of view of the faithful, the Church made a persuasive case for its defence of traditional family – the bottom line of the two campaigns – that was grounded in Mormon eschatology and offered the ultimate reward for adhering to those sanctified family values and gender roles. Second, by insisting that both the ERA and same-sex marriage are moral rather than merely political issues, it justified its involvement and was able to extract even more com 160 Maciej Potz mitment from the rank-and-file. Third, and perhaps crucially, by perpetuating and reinforcing the status of the Church President as a “prophet, seer and revelator” acting on direct inspiration from God, the leadership could muster the allegiance of even those Mormons who failed to grasp so easily the threatening this- or otherworldly consequences of the two issues. Organizationally, the reasons for the Church’s relative success are also three-fold. First, the hierarchical structure of the group, with well-defined chain of command, makes possible quick mobilization of a large number of activists. Second, due to uni versal (male) priesthood, virtually all male Mormons over the age of 12 are ordained priests of some level and thus perceive their commissions from their superiors as religious callings, not just administrative tasks. Third, as a sociological observation, the membership in LDS Church makes Mormons more committed and, at the same time, equips them with more civic skills than members of comparable organizations, both secular and religious; this, naturally, bears positively on their political effective ness. All these explanations should not be treated as mutually exclusive, but rather as multiple factors enabling the LDS Church to engage in successful political activity. BIBLIOGRAPHY Articles of Faith, art. 9, [in:] Pearl of Great Price, J. Smith, Filiquarian Publishing, 2006. son, E. T. Fourteen Fundamentals in Following the Prophet, speech at Brigham Young University delivered on February 26, 1980, http://speeches.byu.edu/?act=viewitem&id=88 (access 11.08.2014) Bigler, D. L. 1998. Forgotten Kingdom. The Mormon Theocracy in the American West, 1847–1896, U State University Press, Logan.i Brennan, Jr., W. J. 1993. The Constitution of the United States: Contemporary Ratification, [in:] Readings in American Government and Politics, R. B. Ripley, E. E. Slotnick (eds.), Wadsworth, Belmont.i Campbell, D. E., Monson, J. Q. 2007. Dry Kindling: A Political Profile of American Mormons, [in:] From Pews to Polling Places: Faith and Politics in the American Religious Mosaic, J. Wilson (ed.), George town University Press, Washington. Cannon, E. 1978. If We want to Get Up, We Have to Get On, “Ensign”, November, https://www.lds.org/ ensign/1978/11/if-we-want-to-go-up-we-have-to-get-on?lang=eng (access 12.08.2014). Church Instructs Leaders on Same-Sex Marriage, January 10, 2014, http://www.mormonnewsroom.org/ article/church-instructs-leaders-on-same-sex-marriage (access 17.08.2014). Crapo, R. Chronology of Mormon/LDS Involvement in Same-Sex Marriage Politics, http://www.mormon socialscience.org/wp-content/uploads/2009/11/Crapo-R1997-Chronology-of-LDS-Involvement-In- Same-Sex-Marriage-Politics.pdf (access 17.08.2014). g p ( ) Della Porta, D., Diani, M. 2006. Social movements. An introduction, Blackwell, Malden–Oxford–Carlton. l h A d “ ” 11 19 16 h // l / g p ( ) Della Porta, D., Diani, M. 2006. Social movements. An introduction, Blackwell, Malden–Oxford–Carlton. Equal Rights Amendment, “Deseret News”, January 11, 1975, p. 16, http://news.google.com/newspa Della Porta, D., Diani, M. 2006. Social movements. An introduction, Blackwell, Malden–Ox Della Porta, D., Diani, M. 2006. Social movements. An introduction, Blackwell, Malden–Oxford–Carlton. Equal Rights Amendment, “Deseret News”, January 11, 1975, p. 16, http://news.google.com/newspa pers?nid=Aul-kAQHnToC&dat=19750111&printsec=frontpage&hl=en (access 21 08 2014) Equal Rights Amendment, “Deseret News”, January 11, 1975, p. 16, http://news.google.com/news pers?nid=Aul-kAQHnToC&dat=19750111&printsec=frontpage&hl=en (access 21.08.2014). First Presidency Issues Statement Opposing Equal Rights Amendment, 1976, https://www.lds.org sign/1976/12/news-of-the-church?lang=eng (access 9.08.2014). Gordon, E. E., Gillespie, W. 2012. The Culture of Obedience and the Politics of Stealth: Mormon Mobili zation Against ERA and Same-Sex Marriage, “Religion and Politics”, no. 5. Hansen, K. 1967. Quest for Empire. Political Kingdom of God and the Council of Fifty in Mormon History, Nebraska University Press, Lincoln. 161 A CONTEMPORARY RELIGIOUS POLITICAL MOVEMENT: LDS CHURCH’S NATION-WIDE POLITICAL… 161 Kendall White, O. 1989. Mormonism and the Equal Rights Amendment, “Journal of Church and State”, vol. 31, spring. Kintz, L. 1998. Clarity, Mothers, and the Mass-Mediated National Soul: A Defense of Ambiguity, [in:] Media, Culture and the Religious Right, L. Kintz, J. BIBLIOGRAPHY “The Era is a Moral Issue”: The Mormon Church, LDS Women, and the Defeat of the Equal Rights Amendment, “American Quarterly”, vol. 59, no. 3. BIBLIOGRAPHY Lesage (eds.), University of Minnesota Press, Minneapolis–London. Mansbridge, J. 1986. Why We Lost the ERA, University of Chicago Press, Chicago and London. Martin, W. 1996. With God on Our Side. The Rise of the Religious Right in America, Broadway New York. Moore, C. 2008. LDS Church issues statement on same-sex marriage, September 10, http://www.de seretnews.com/article/700257603/LDS-Church-issues-statement-on-same-sex-marriage.html?pg=all (access 16.08.2014). O’Dea, T. 1964. The Mormons, University of Chicago Press, Chicago and London. Obergefell v. Hodges (576 US. 14–556), 2015. “Official Declaration 2”, Doctrine and Covenants, https://www.lds.org/scriptures/dc-testament/ od/2?lang=eng (access 11.08.2014). Perry, L., Cronin, C. 2012. Mormons in American Politics. From Persecution to Power, Praeger, Santa Barbara. Potz, M. 2008. Granice wolności religijnej, Fundacja Nauki Polskiej, Wrocław. Potz, M. 2013. Religious doctrine as a factor of stability of political systems. A study of two North American theocracies, “Politics and Religion Journal”, vol. 7, no. 2. Quinn, D. M. 1997. The Mormon Hierarchy: Extensions of Power, Signature Books, Salt Lake City. Quinn, D. M. 2000. Prelude to the National “Defense of Marriage” Campaign: Civil Discrimination Against Feared or Despised Minorities, “Dialogue: A Journal of Mormon Thought”, vol. 33, no. 3. Q f f g p g Against Feared or Despised Minorities, “Dialogue: A Journal of Mormon Thought”, vol. 33, no. 3. Response to the Supreme Court decision legalizing same-sex marriage in the United States, June 29, 20 Response to the Supreme Court decision legalizing same-sex marriage in the United States, June 29, 2015, LDS Church Newsroom, http://www.mormonnewsroom.org/article/top-church-leaders-counsel-mem bers-after-supreme-court-same-sex-marriage-decision (access 11.08.2015). LDS Church Newsroom, http://www.mormonnewsroom.org/article/top-church-leaders-counsel-mem bers-after-supreme-court-same-sex-marriage-decision (access 11.08.2015). LDS Church Newsroom, http://www.mormonnewsroom.org/article/top-church-leaders-counsel-m b ft t i d i i ( 11 08 2015) LDS Church Newsroom, http://www.mormonnewsroom.org/article/top-church bers-after-supreme-court-same-sex-marriage-decision (access 11.08.2015). bers-after-supreme-court-same-sex-marriage-decision (access 11.08.2015). Riess, R., Kimball Bigelow, C. 2005. Mormonism for Dummies, Wiley, Indianapolis. Tanner, N. E. 1979. The Debate is Over, “Ensign”, August, https://www.lds.org/ensign/1979/08/the-debate- is-over?lang=eng (access 12.08.2014). The Divine Institution of Marriage, August 13, 2008, http://www.mormonnewsroom.org/ldsnewsroom/ eng/commentary/the-divine-institution-of-marriage (access 16.08.2014). van Beek, W. 2011. The Infallibility Trap: the Sacralisation of Religious Authority, “International Journal of Mormon Studies”, vol. 4. Weber, M. 1978. Economy and Society, University of California Press: Berkeley–Los Angeles Wilcox, C. 1996. Onward Christian Soldiers. The Religious Right in American Politics, Westview Press, Boulder.i Williams, P. 2002. America’s Religions. From Their Origins to the Twenty-first Century, University of Illinois Press, Urbana and Chicago. Woodruff, W. 1890. Address at 61 Semiannual General Conference of the Church, October 6, https://www. lds.org/scriptures/dc-testament/od/1?lang=eng (access 30.08.2014). Young, N. J. 2007. BIOGRAPHY Maciej Potz, born 1977 in Łódź, is an Associate Professor of Political Science at the Department of Political Systems, Faculty of International and Political Studies, University of Łódź. He earned his Ph.D. in Political Science in 2006 from the Silesian 162 Maciej Potz University in Katowice. His research focuses on political theory (especially, theory of power and democratic theory) and religion in public life. As a Foundation for Polish Science scholar, he conducted research on Shakers and Mormons in USA. His publications include Granice wolności religijnej, Fundacja Nauki Polskiej, Wrocław 2008; Legitimation mechanisms as third-dimension power practices. The case of the Shakers, “Journal of Political Power” 2012, vol. 5, no 3; Disentangling Democracy, [in:] European Culture in Diversity, eds. K. Kujawińska-Courtney et al., Cambridge: Cambridge Scholars Publishing, 2011; Religious doctrine as a factor of stability of political systems. A study of two North American theocracies, “Politics and Religion Journal” 2013, vol. 7, no. 2. E-mail: maciekpotz@hotmail.com
W4301772741.txt	https://figshare.com/ndownloader/files/2426320	en		Analysis of the Proportional Hazards Model With Sparse Longitudinal Covariates	Carolina Digital Repository (University of North Carolina at Chapel Hill)	2,015	cc-by	5,530	Supplemental material for analysis of the proportional hazards model with sparse longitudinal covariates Hongyuan Cao1 , Mathew M. Churpek2 , Donglin Zeng3 and Jason P. Fine3 Abstract This supplementary material provides details on the proofs of Theorem 1, Theorem 2 and Corollary 1, 2 and 3. 1 Proofs of Lemmas 1.1 Proof of Lemma 1 Lemma 1 Under conditions of Theorem 1, we have Bn (β, τ ) = op (1), where Bn (β, t) = n − 1 2 3 −2 Mi n Z tX X h Khn (u − Rik )I(Rik ≤ u) (β − β0 )T Zi (Rik ) i=1 0 k=1 n S (0) (β, u) oi2 T n log (0) Yi (u)eβ0 Zi (u) λ0 (u)du. Sn (β0 , u) Department of Statistics, University of Missouri-Columbia, Columbia, MO, 65201. Department of Health Studies and Department of Medicine, University of Chicago, Chicago, IL, 60637. Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, NC, 27514. 1 Proof. We can write Bn (β, t) as Bn (β, t) = n −1 t Z n −1 0 n Z X i=1 ∞ Z 0 ∞ {Khn (u − r1 )I(r1 ≤ u)dN ∗ (r1 )Khn (u − r2 )I(r2 ≤ u)dN ∗ (r2 )} 0 n S (0) (β, u) o n (β − β0 ) Zi (r1 )Zi (r2 ) (β − β0 ) − (β − β0 ) Zi (r1 )log (0) Sn (β0 , u) n S (0) (β, u) o h n S (0) (β, u) oi⊗2 T n n − (β − β0 )T Zi (r2 )log (0) + log (0) Yi (u)eβ0 Zi (u) λ0 (u)du. Sn (β0 , u) Sn (β0 , u) T T T This can be expanded further: Bn (β, t) = n −1 t Z n −1 0 n Z X i=1 ∞ Z 0 ∞ Khn (u − r1 )Khn (u − r2 )I(r1 ≤ u)I(r2 ≤ u)dN ∗ (r1 )dN ∗ (r2 ) 0 T (β − β0 )T Zi (r1 )Zi (r2 )T (β − β0 )Yi (u)eβ0 Zi (u) λ0 (u)du Z t n Z ∞Z ∞ X −1 −1 − n n Khn (u − r1 )Khn (u − r2 )I(r1 ≤ u)I(r2 ≤ u)dN ∗ (r1 )dN ∗ (r2 ) 0 i=1 0 0 (0) (β − β0 )T Zi (r1 )log{ − n −1 t Z n −1 0 n Z X i=1 ∞ Sn (β, u) (0) Sn (β0 , u) Z 0 T }Yi (u)eβ0 Zi (u) λ0 (u)du ∞ Khn (u − r1 )Khn (u − r2 )I(r1 ≤ u)I(r2 ≤ u)dN ∗ (r1 )dN ∗ (r2 ) 0 (0) (β − β0 )T Zi (r2 )log{ + n −1 t Z n −1 0 n Z X i=1 ∞ 0 Sn (β, u) (0) Sn (β0 , u) Z T }Yi (u)eβ0 Zi (u) λ0 (u)du ∞ Khn (u − r1 )Khn (u − r2 )I(r1 ≤ u)I(r2 ≤ u)dN ∗ (r1 )dN ∗ (r2 ) 0 (0) [log{ Sn (β, u) (0) T }]⊗2 Yi (u)eβ0 Zi (u) λ0 (u)du Sn (β0 , u) = I + II + III + IV. Assume that for t 6= r, pr(dN ∗ (t) = 1 \| N ∗ (r) − N ∗ (r−) = 1) = g(t, r)dt, where g(t, r) is continuous for t 6= r and g(t±, t±) exists. After taking expectation, we obtain 2 E[I] = n −1 t Z −1 n 0 n hZ X r6=r0 i=1 0 0 0 Khn (u − r)Khn (u − r )I(r ≤ u)I(r ≤ u)E{dN ∗ (r)dN ∗ (r )} n o Z 0 T T β0T Zi (u) E (β − β0 ) Zi (r)Zi (r ) (β − β0 )Yi (u)e + Khn (u − r)2 I(r ≤ u)E{dN ∗ (r)} r oi n T T β0T Zi (u) λ0 (u)du E (β − β0 ) Zi (r)Zi (r) (β − β0 )Yi (u)e Z Z t n h X 0 0 0 0 −1 −1 n Khn (u − r)Khn (u − r )I(r ≤ u)I(r ≤ u)g(r, r )E{dN ∗ (r )}dr = n 0 r6=r0 i=1 n o Z 0 T T β0T Zi (u) E (β − β0 ) Zi (r)Zi (r ) (β − β0 )Yi (u)e + Khn (u − r)2 I(r ≤ u)λ∗ (r)dr r n oi T T β0T Zi (u) E (β − β0 ) Zi (r)Zi (r) (β − β0 )Yi (u)e λ0 (u)du Z tZ n o ∗ −1 T T β0T Z(u) Khn (u − r)I(r ≤ u)λ (u)E (β − β0 ) Z(r)Z(u) (β − β0 )Y (u)e = n r 0 nZ ∞ o −1 2 ∗ {g(r, u+) + g(r, u−)}/4dr + o(1) + hn K(z) λ (u)dz + o(1) 0 n n o X T T −1 [(β − β0 ) n E Zi (u − hn z)Zi (u − hn z)T Yi (u)eβ0 Zi (u) (β − β0 )] λ0 (u)du. i=1 Using change of variables, we have −1 Z th {(β − β0 )T s(2) (β0 , u)(β − β0 )}{g(u+, u+) + g(u+, u−) + g(u−, u+) 0 Z ∞ −1 + g(u−, u−)}/8 + o(1) + 2hn { K(z)2 dz + o(1)}{(β − β0 )T s(2) (β0 , u)(β − β0 ) 0 i + O(hn )} λ0 (u)du E{I} = n = O{(nhn )−1 }. Similar order can be obtained for II, III and IV and we omit the details here. Therefore, Bn (β, τ ) converges in probability to 0. 3 1.2 Proof of Lemma 2 Lemma 2 Under assumptions of Theorem 1, (nhn )1/2 Un (β0 ) → N {0, Σ(β0 )}, where Σ(β0 )) = R∞ 0 R τ n (2) K(z) dz 0 s (β0 , u) − 2 s(1) (β0 ,u)⊗2 s(0) (β0 ,u) o λ0 (u)du Proof. We can decompose (nhn )1/2 Un (β0 ) into two parts: 1/2 (nhn ) Un (β0 ) = −1/2 h1/2 n n n Z X i=1 + hn1/2 n−1/2 n Z X i=1 0 τ Z τ ∞ Z 0 Khn (u − r)I(r ≤ u)dNi∗ (r)[Zi (r) − Z̄(β0 , u)]dMi (u) 0 ∞ T Khn (u − r)I(r ≤ u)dNi∗ (r){Zi (r) − Z̄(β0 , u)}Yi (u)eβ0 Zi (u) λ0 (u)du 0 = I(β0 , τ ) + II(β0 , τ ), where T dMi (u) = dNi (u) − Yi (u)eβ0 Zi (u) λ0 (u)du. Note that II(β0 , τ ) = −1/2 h1/2 n n Z 0 T τ Mi n X X Khn (u − Rij )I(Rij ≤ u){Zi (Rij ) − Z̄(β0 , u)} i=1 j=1 T Yi (u){eβ0 Zi (u) − eβ0 Zi (Rij ) }λ0 (u)du. h i T T Define F (u, s) = E {Z(u−s)− Z̄(β0 , u)}Y (u){eβ0 Z(u) −eβ0 Z(u−s) } . After taking expectation 4 together with Taylor expansion we have n1/2 h1/2 n E{II(β0 , τ )} = τ Z 0 {e β0T Z(u) −e ∞ Z h K(z)E {Z(u − hn z) − Z̄(β0 , u)}Y (u) 0 β0T Z(u−hn z) = n1/2 h1/2 n {F (u, 0) + }dzλ0 (u)du ∂F (u, s) \|s=0 hn z + o(hn )} ∂s = O(n1/2 h3/2 n ) = o(1) by (A5). Therefore II(β0 , τ ) converges to 0 in probability. We now derive the asymptotic normality of the term I(β0 , τ ). By the martingale property, we have < I(β0 ), I(β0 ) > (τ ) = hn n n Z X −1 i=1 0 τ ∞ hZ Khn (u − r)I(r ≤ u)dNi∗ (r){Zi (r) − Z̄(β0 , u)} i2 0 T Yi (u)eβ0 Zi (u) λ0 (u)du. Similar as in the derivation of E{Bn (β, t)}, after taking expectation and change of variables, we have E{I(β0 , τ )2 } = E < I(β0 ), I(β0 ) > (τ ) n Z τ Z ∞ X −1 = n K(z)2 λ∗ (u)dz i=1 0 0 (1) (1) Sn (β0 , u)⊗2 i T E Zi (u)Zi (u) − 2Zi (u) (0) + (0) Yi (u)eβ0 Zi (u) λ0 (u)du + o(1) Sn (β0 , u) Sn (β0 , u)2 Z ∞ Z τh s(1) (β0 , u)⊗2 i 2 (2) = 2 K(z) dz λ0 (u)du + o(1). s (β0 , u) − (0) s (β0 , u) 0 0 h T Sn (β0 , u) 5 Next, we verify that Lindeberg condition holds for I(β0 , u). For ∀ > 0, consider Z τ 0 n h X −1/2 h1/2 n n i2 Khn (u − Rlk )I(Rlk ≤ u){Zl (Rlk ) − Z̄(β0 , u)} k=1 l=1 Yl (u)e Ml X β0T Zl (u) Ml o n X 1/2 −1/2 Khn (u − Rlk )I(Rlk ≤ u){Zl (Rlk ) − Z̄(β0 , u)}\| > du. λ0 (u)I \|hn n k=1 This can be decomposed into two parts. Z τ 0 n h X −1/2 h1/2 n n Ml X i2 Khn (u − Rlk )I(Rlk ≤ u){Zl (Rlk ) − Z̄(β0 , u)} I(Ml ≤ a) k=1 l=1 Ml o n X T T −1/2 Yl (u)eβ0 Zl (u) λ0 (u)I \|h1/2 (u − R )I(R ≤ u){Z (R ) − Z̄(β , u)}\| > du n K lk lk l lk 0 hn n k=1 Z τ + 0 n h X −1/2 h1/2 n n l=1 i2 Khn (u − Rlk )I(Rlk ≤ u){Zl (Rlk ) − Z̄(β0 , u)} I(Ml > a) k=1 β0T Zl (u)T Yl (u)e Ml X Ml n o X 1/2 −1/2 λ0 (u)I \|hn n Khn (u − Rlk )I(Rlk ≤ u){Zl (Rlk ) − Z̄(β0 , u)}\| > du k=1 = I + II. By (A1), Z I = Op (1) τ −1/2 I(Op (h1/2 ) > )λ0 (u)du → 0 as n → ∞ and hn → 0. n n 0 Since pr(Ml > a) → 0 as a → ∞ by (A2), we have II → 0. This shows that (nhn )1/2 Un (β0 , ·) converges weakly to a certain continuous Gaussian process. 6 Since this process evaluated at time t = τ has covariance matrix Σ(β0 ), therefore, we have (nhn )1/2 Un (β0 , u) → N {0, Σ(β0 )}, where Σ(β0 ) = 2 1.3 R∞ 0 R τ n (2) s (β0 , u) − 0 K(z)2 dz s(1) (β0 ,u)⊗2 s(0) (β0 ,u) (1.1) o λ0 (u)du. Proof of Lemma 3 Lemma 3 Under conditions of Theorem 2, we have (nhn )1/2 E "Z τ Khn (t − r) Z(r) − 0 = (nhn ) 1/2 ( ∞ Z 0 (1) S̃n (β, t) ) # dN ∗ (r)dN (t) (0) S̃n (β, t) A(β0 )(β − β0 ) + Dn1/2 hn5/2 + o{(nhn )1/2 \|β − β0 \|}. Proof. After change of variables, we have K ≡ (nhn )1/2 E "Z τ ( Z Khn (t − s) Z(s) − 0 Z τ hZ (1) S̃n (β, t) (0) S̃n (β, t) ) # dN ∗ (s)dN (t) T K(z)E{Z(t − hn z)Y (t)eβ0 Z(t) }λ∗ (t − hn z)dz = (nhn )1/2 z 0 Z i s̃(1) (β, t) T + op (1)}E{Y (t)eβ0 Z(t) }λ∗ (t − hn z)dz µ0 (t)dt K(z){ (0) − s̃ (β, t) z T Denote G(t, s) = E{Z(t − s)Y (t)eβ0 Z(t) }. We then do Taylor expansions 1/2 Z τ K = (nhn ) h i T E{Z(t)Y (t)eβ0 Z(t) λ∗ (t)} µ0 (t)dt 0 − (nhn )1/2 Z 0 τ s̃(1) (β, t) T E{Y (t)eβ0 Z(t) λ∗ (t)}µ0 (t)dt + (nh5n )1/2 D1 + o(1), (0) s̃ (β, t) 7 R∞ where we used the fact that Z D1 = τ Z 2 K(z)z dz h 1 ∂G2 (t, s) ∂s2 2 0 z K(z)dz = 1, −∞ R∞ −∞ \|s=0 λ∗ (t) − zK(z)dz = 0 and i ∂λ∗ (x) ∂G(t, s) \|s=0 \|x=t µ0 (t)dt + o(\|β̂ − β0 \|). ∂s ∂x We do a further Taylor expansion of K around β0 and obtain 1/2 τ Z {s̃(2) (β0 , t) − K = −(nhn ) 0 = (nhn ) 1/2 s̃(1) (β0 , t)⊗2 }µ0 (t)dt(β − β0 ) + Dn1/2 hn5/2 + o{1 + (nhn )1/2 \|β − β0 \|} s̃(0) (β0 , t) A(β0 )(β − β0 ) + Dn1/2 hn5/2 + o{1 + (nhn )1/2 \|β − β0 \|}, where Z D= τ Z 2 K(z)z dz z 0 h 1 ∂G2 (t, s) 2 ∂s2 i ∂λ∗ (x) ∂G(t, s) \|s=0 λ (t) − \|s=0 \|x=t µ0 (t)dt ∂s ∂x ∗ and τ n s̃(1) (β0 , t)⊗2 o s̃(2) (β0 , t) − (0) µ0 (t)dt s̃ (β0 , t) 0 Z τ h i s̃(1) (β0 , t) T s̃(1) (β0 , t) β0T Z(t) }{Z(t) − (0) } Y (t)e µ0 (t)dt. = − E {Z(t) − (0) s̃ (β0 , t) s̃ (β0 , t) 0 Z A(β0 ) = − It is a non-negative definite matrix. From (C5), A(β0 ) is non-singular. 1.4 Proof of Lemma 4 Lemma 4 Under conditions of Theorem 2, var hZ τ Z (1) h1/2 n Khn (t − r){Z(r) − S̃n (β0 , t) (0) i }dN (r)dN (t) ∗ S̃n (β0 , t) Z Z τn s̃(1) (β0 , t)⊗2 o 2 (2) = K(z) dz s̃ (β0 , t) − (0) µ0 (t)dt. s̃ (β0 , t) 0 0 8 Proof. This can be calculated as follows: Σ = var hZ τ Z (1) h1/2 n Khn (t 0 − r){Z(r) − S̃n (β0 , t) (0) i }dN ∗ (r)dN (t) S̃n (β0 , t) (1) ZZ ZZ Khn (t1 − r1 )Khn (t2 − r2 ){Z(r1 ) − = Ehn h Z dN (r1 )dN (r2 )dN (t1 )dN (t2 ) − E ∗ τ ∗ Z S̃n (β0 , t1 ) (0) S̃n (β0 , t1 ) (1) }{Z(r2 ) − S̃n (β0 , t2 ) (0) (1) h1/2 n Khn (t − r){Z(r) − 0 S̃n (β0 , t) (0) }dN ∗ (r)dN (t) S̃n (β0 , t) = I − II. For II, we get h Z II = hn E = hn − = Z τ 0 τ Z (1) Z Khn (t − r){Z(r) − S̃n (β0 , t) (0) S̃n (β0 , t) i2 T }λ∗ (r)dreβ0 Z(t) Y (t)µ0 (t)dt h T K(z) E{Z(t − hz)Y (t)eβ0 Z(t) }λ∗ (t − hz)µ0 (t)dt 0 (1) i 2 S̃n (β0 , t) β0T Z(t) ∗ Y (t)e }λ (t − hn z)µ0 (t)dt dz E{ (0) S̃n (β0 , t) Z τ [s̃(1) (β0 , t) − s̃(1) (β0 , t) + Op {(nhn )−1 }]µ0 (t)dt hn 0 = o(hn ). 9 } S̃n (β0 , t2 ) + O(h2n ) 2 i2 Next we decompose I into four parts. Z (1) Z Kh (t1 − r1 )Kh (t2 t1 6=t2 r1 6=r2 E{dN ∗ (r1 )dN ∗ (r2 )dN (t1 )dN (t2 )} I = hn E Z − r2 ){Z(r1 ) − (0) S̃n (β0 , t1 ) (1) }{Z(r2 ) − (1) Z Kh (t1 − r)Kh (t2 − r){Z(r) − + hn E S̃n (β0 , t1 ) t1 6=t2 ∗ S̃n (β0 , t1 ) (0) S̃n (β0 , t1 ) S̃n (β0 , t2 ) (0) } S̃n (β0 , t2 ) (1) }{Z(r) − S̃n (β0 , t2 ) (0) } S̃n (β0 , t2 ) E{dN (r)dN (t1 )dN (t2 )} Z Z (1) (1) S̃n (β0 , t) S̃n (β0 , t) + hn E Kh (t − r1 )Kh (t − r2 ){Z(r1 ) − (0) }{Z(r2 ) − (0) } S̃n (β0 , t) S̃n (β0 , t) r1 6=r2 E{dN ∗ (r1 )dN ∗ (r2 )dN (t)} Z Z (1) S̃n (β0 , t) ⊗2 } E{dN ∗ (r)dN (t)} + hn E Kh (t − r)2 {Z(r) − (0) S̃n (β0 , t) = I1 + I2 + I3 + I4 . It is easy to see that I1 = O(hn ), I2 = O(hn ) and I3 = O(hn ). Now we look at I4 : Z τ Z I4 = hn E 2 h−2 n K(z) n (1) Z(t − hn z) − 0 S̃n (β0 , t) o⊗2 (0) S̃n (β0 , t) T λ∗ (t − hn z)hn dzY (t)eβ0 Z(t) µ0 (t)dt (1) h S̃n (β0 , t) T K(z)2 E Z(t − hn z)Z(t − hn z)T Y (t)eβ0 Z(t) λ∗ (t − hn z) − 2 (0) S̃n (β0 , t) 0 (1) i S̃n (β0 , t) ⊗2 β0T Z(t) ∗ β0T Z(t) ∗ Z(t − hn z)Y (t)e λ (t − hn z) + { (0) } Y (t)e λ (t − hn z) dzµ0 (t)dt S̃n (β0 , t) Z Z n s̃(1) (β0 , t)⊗2 o 2 (2) = K(z) dz s̃ (β0 , t) − (0) µ0 (t)dt + O(hn ) + O{(nhn )−1 }. s̃ (β0 , t) Z τ Z = Therefore, we have Z Σ̃(β0 ) = 2 Z K(z) dz 0 τ n s̃(1) (β0 , t)⊗2 o µ0 (t)dt. s̃(2) (β0 , t) − (0) s̃ (β0 , t) 10 (1.2) 2 Proofs of Main Results 2.1 Proof of Theorem 1 Our main tool is the martingale central limit theorem (Theorem 5.3.5 in Fleming and Harrington (2005)). First we need the following proposition: Proposition 1 Under (A1), (A2) and (A5), for any compact neighbourhood B of β0 , we have limn→∞ sup0≤t≤τ,β∈B \|\|Sn(k) (β, t) − s(k) (β, t)\|\| = 0 a.s. for k = 0, 1, 2. (2.3) (k) Proof. This follows from Theorem 37 of Polland (1984) and the observation that Sn (β, t) is Lipschitz continuous in β ∈ B. To show the consistency of β̂n , first it follows from the definition of u1 (β) that u1 (β0 ) = 0. Second, it follows from condition (A4) and the fact that v1 (β) is semi-positive definite for any β that β0 is the unique root to the equation u1 (β) = 0. Finally we need to show that Un (β) converges in probability to u1 (β) uniformly in B. Consider the process Fn (β, t) = ln∗ (β, t) − ln∗ (β0 , t) Mi n Z tX X −1 = n Khn (u − Rik )I(Rik ≤ u)(β − β0 )T Zi (Rik )dNi (u) i=1 − n −1 0 k=1 Mi n Z tX X i=1 0 n S (0) (β, u) o n dNi (u), Khn (u − Rik )I(Rik ≤ u)log (0) Sn (β0 , u) k=1 and the process Gn (β, t) = n −1 Mi n Z tX X i=1 − n−1 T Khn (u − Rik )I(Rik ≤ u)(β − β0 )T Zi (Rik )Yi (u)eβ0 Zi (u) λ0 (u)du 0 k=1 Mi n Z tX X i=1 0 n S (0) (β, u) o T n Khn (u − Rik )I(Rik ≤ u)log (0) Yi (u)eβ0 Zi (u) λ0 (u)du. Sn (β0 , u) k=1 11 Then for each β, Fn (β, ·) − Gn (β, ·) is a local square integrable martingale with < Fn (β, ·) − Gn (β, ·), Fn (β, ·) − Gn (β, ·) >= Bn (β, ·), where Bn (β, t) = n − −2 Mi n Z tX X h Khn (u − Rik )I(Rik ≤ u) (β − β0 )T Zi (Rik ) i=1 0 k=1 n S (0) (β, u) oi2 T n log (0) Yi (u)eβ0 Zi (u) λ0 (u)du. Sn (β0 , u) From Lemma 1, we have Bn (β, τ ) converges in probability to 0. Now we look at Gn (β, τ ). After taking expectation and change of variables, we have E{Gn (β, τ )} = n −1 n Z X i=1 τ Z 0 u/hn K(z)I(z ≥ 0)λ∗ (u − hn z)dz 0 β0T Zi (u) T (β − β0 ) E{Zi (u − hn z)Yi (u)e h n S (0) (β, u) oi T n Yi (u)eβ0 Zi (u) λ0 (u)du } − E log (0) Sn (β0 , u) It follows that for each β ∈ B, Z Gn (β, τ ) → 0 τ h n s(0) (β, u) o i (0) (β − β0 ) s (β0 , u) − log (0) s (β0 , u) λ0 (u)du in probability. s (β0 , u) (2.4) T (1) Thus by the inequality of Lenglart (Corollary 3.4.1 in Fleming and Harrington (2005)), Fn (β, τ ) converges in probability to the same limit as Gn (β, τ ) for each β ∈ B. Now by the boundedness condition we may evaluate the first and second derivatives of this limiting function of β by taking partial derivatives inside the integral. These derivatives 12 equal to τ Z n o s(1) (β0 , u) − s(0) (β0 , u)z̄(β, u) λ0 (u)du = u1 (β) 0 and τ n s(0) (β0 , u) s(0) (β0 , u) o s(2) (β, u) (0) − s(1) (β, u)⊗2 (0) λ0 (u)du s (β, u) s (β, u)2 0 Z τ h (2) i s (β, u) ⊗2 (0) = − − z̄(β, u) s (β0 , u)λ0 (u)du = −v1 (β). s(0) (β, u) 0 Z − The first derivative is zero at β = β0 ; the second is minus a positive semi-definite matrix; and at β = β0 is a minus positive definite matrix. Thus for each β ∈ B, Fn (β, τ ) converges in probability to a concave function of β with a unique maximum at β = β0 . Since β̂n maximizes the random concave function Fn (β, τ ), by the fact that pointwise convergence in probability of random concave functions implies uniform convergence on compact subspaces (Andersen and Gill (1982)), it follows that β̂n → β0 in probability. Next we show the asymptotic normality of β̂n . By Taylor expansion of Un (β̂n , τ ), we have 0 = Un (β̂n , τ ) = Un (β0 , τ ) + ∂Un (β, τ ) \|β=β ∗ (β̂n − β0 ), ∂β (2.5) where β ∗ lies in the segment between β̂n and β0 . We have n ∂U (β, τ ) o−1 n (nhn )1/2 (β̂n − β0 ) = − \|β=β ∗ (nhn )1/2 Un (β0 , τ ). ∂β (2.6) Thus, we have two tasks here: first to establish the asymptotic normality of (nhn )1/2 Un (β0 , τ ) and second to find the limiting distribution of ∂Un (β,τ ) \|β=β ∗ ∂β 13 for any β ∗ between β̂ and β0 . The first part follows from Lemma 2. For the second part of the proof, note that − = n ∂Un (β, u) \|β=β ∗ ∂β n Z τ Z u/hn X −1 0 i=1 Khn (u − r)I(r ≤ u)dNi∗ (r) 0 n S (2) (β ∗ , u) n (0) Sn (β ∗ , u) (1) − Sn (β ∗ , u)⊗2 o (0) Sn (β ∗ , u)2 dNi (u) and that Z W (β0 ) = 0 (2) Define Vn (β, t) = Sn (β,t) (0) Sn (β,t) τ n s(1) (β0 , u)⊗2 o s(2) (β0 , u) − λ0 (u)du. s(0) (β0 , u) − Z̄(β, t)⊗2 and v(β, t) = s(2) (β,t) s(0) (β,t) − z̄(β, t)⊗2 . Hence ∂Un (β, τ ) \|β=β ∗ − W (β0 )\|\| ∂β n Z τ Z ∞ X −1 ≤ \|\|n Khn (u − r)I(r ≤ u)dNi∗ (r){Vn (β ∗ , u) − v(β ∗ , u)}dNi (u)\|\| \|\| − + \|\|n−1 + \|\|n−1 i=1 0 n Z τ X i=1 0 n Z τ X i=1 Z + \|\| 0 τ {n−1 0 ∞ Z Khn (u − r)I(r ≤ u)dNi∗ (r){v(β ∗ , u) − v(β0 , u)}dNi (u)\|\| 0 ∞ Z 0 0 n Z ∞ X i=1 T Khn (u − r)I(r ≤ u)dNi∗ (r)v(β0 , u){dNi (u) − Yi (u)eβ0 Zi (u) λ0 (u)du}\|\| T Khn (u − r)I(r ≤ u)dNi∗ (r)Yi (u)eβ0 Zi (u) − s(0) (β0 , u)}v(β0 , u)λ0 (u)du\|\| 0 = I + II + III + IV (2.7) By (A1) and Theorem III.1 in Andersen and Gill (1982), it follows that supt∈[0,τ ],β∈B \|\|Vn (β, t) − v(β, t)\|\| → 0 in probability. 14 (2.8) Hence β ∗ → β0 in probability. By Chebyshev’s inequality, τ Z n pr{ −1 0 Rτ 0 ≤ n Z X ∞ Khn (u − r)I(r ≤ u)dNi∗ (r)dNi (u) > c} 0 i=1 (0) s (β0 , u)λ0 (u)du →0 c (2.9) as c → ∞ by (A2) and (A3). Therefore, I = op (1). Again, (2.8), (2.9) together with the continuity of v(β, t) in β, uniformly for t implies that II is also asymptotically negligible. For III, using Lenglart’s inequality as in Theorem I.1 in Andersen and Gill (1982) and Chebyshev’s inequality. We have nZ pr 0 ≤ ηδ −2 ≤ ηδ −2 τ n −1 n Z X Khn (u − r)I(r ≤ u)dNi∗ (r)v(β0 , u)dMi (u) > δ o 0 i=1 h + pr n−1 ∞ Z τ n −1 0 Z n X { i=1 ∞ T Khn (u − r)I(r ≤ u)dNi∗ (r)}2 v(β0 , u)2 Yi (u)eβ0 Zi (u) λ0 (u)du > η 0 −1 + O{(nhn η) }. Thus, III disappears as n → ∞. (0) Finally, IV = op (1) by (A2) and the uniform convergence of Sn (β0 , u) to s(0) (β0 , u). 2.2 Proof of Corollary 1 We next show the consistency of the variance estimate. It follows from the proof of Theorem 1 that − ∂Un (β, τ ) \|β=β̂n → W (β0 ) in probability. ∂β 15 i On the other hand, by law of large numbers, consistency of β̂n for β0 and the continuous mapping theorem Σ̂ = n −2 n hZ X 0 i=1 τ Z ∞ i⊗2 p Khn (u − r)I(r ≤ u){Zi (r) − Z̄(β̂n , u)}dNi∗ (r)dNi (u) → E{Σ̂(β0 )}. 0 (2.10) Note that −1 E{Σ̂(β0 )} = n E hZ 0 τ Z ∞ i⊗2 Khn (u − r)I(r ≤ u){Z(r) − Z̄(β0 , u)}dN (r)dN (u) ∗ 0 After change of variables, and by (A1), 1 E{Σ̂(β0 )} = nhn Z ∞ 2 Z K(z) dz τ {s(2) (β0 , u) − 0 0 s(1) (β0 , u)⊗2 }du. s(0) (β0 , u) Therefore, p (nhn )Σ̂ → Σ(β0 ) as nhn → ∞. The consistency of variance estimate follows. 2.3 Proof of Theorem 2 Our main tools are empirical processes (van der Vaart and Wellner (1996)). The key idea is to establish the following relationship sup (nhn )1/2 Ũn (β) − (nhn )1/2 [Ũn (β0 ) − E{Ũn (β0 )}] − (nhn )1/2 A(β0 )(β − β0 ) \|β−β0 \|<M (nhn )−1/2 1/2 = Dn1/2 h5/2 \|β − β0 \|}, n + op {1 + (nhn ) (2.11) where A(β0 ) is given in Theorem 2. To obtain (2.11), first, using Pn and P to denote the empirical measure and true probability 16 measure respectively, we obtain (nhn )1/2 Ũn (β) = (nhn )1/2 (Pn − P) "Z + (nhn )1/2 E τ Z Z ( ∞ 0 ( ∞ S̃n (β, t) 0 ) (0) S̃n (β, t) (1) Khn (t − r) Z(r) − 0 (1) Khn (t − r) Z(r) − 0 "Z τ S̃n (β, t) (0) S̃n (β, t) # dN ∗ (r)dN (t) ) # dN ∗ (r)dN (t) = I + II. (2.12) For the second term on the right-hand side of (2.12), from Lemma 3, we have II = (nhn )1/2 A(β0 )(β − β0 ) + Dn1/2 hn5/2 + o{(nhn )1/2 \|β − β0 \|}, (2.13) where Z ∞ Z 2 D= z K(z)dz −∞ τ h i 0 T 0 00 T E{Z(t) Y (t)eβ0 Z(t) }λ∗ (t) + 2−1 E{Z(t) Y (t)eβ0 Z(t) }λ∗ (t) µ0 (t)dt 0 and τ s̃(1) (β0 , t)⊗2 o µ0 (t)dt A(β0 ) = − s̃ (β0 , t) − (0) s̃ (β0 , t) 0 Z τ h i s̃(1) (β0 , t) s̃(1) (β0 , t) T T = − E {Z(t) − (0) }{Z(t) − (0) } Y (t)eβ0 Z(t) µ0 (t)dt. s̃ (β0 , t) s̃ (β0 , t) 0 Z n (2) The matrix A(β0 ) is a non-negative definite and by assumption (C5) non-singular. For the first term on the right-hand side of (2.12), we consider the class of functions Z n 1/2 hn 0 τ (1) Z Khn (t − r){Z(r) − S̃n (β, t) (0) S̃n (β, t) o }dN (r)dN (t) : \|β − β0 \| < ∗ for a given constant . Note that the functions in this class are Lipschitz continuous in β and 17 the Lipschitz constant is uniformly bounded by τ Z M1 Z hn1/2 Khn (t − r)dN ∗ (r)dN (t), 0 which has finite second moment and M1 is the upper bound of (1) s̃(2) (β,t) −{ s̃s̃(0) (β,t) }⊗2 . s̃(0) (β,t) (β,t) Therefore, this class is P-Donsker class by the Jain-Marcus theorem (van der Vaart and Wellner (1996)). As the result, we obtain that the first term in the right-hand side of (2.12) for \|β − β0 \| < M (nhn )−1/2 is equal to (nhn )1/2 (Pn − P) "Z τ ( Z Kh (t − r) Z(r) − 0 (1) S̃n (β0 , t) ) # dN ∗ (r)dN (t) + op (1) (0) S̃n (β0 , t) h i = (nhn )1/2 Ũn (β0 ) − E{Ũn (β0 )} + op (1). (2.14) Combining (2.13) and (2.14), we obtain (2.11). Consequently, 1/2 (nhn )1/2 A(β0 )(β̃n − β0 ) + Op (n1/2 h5/2 \|β̃n − β0 \|} n ) + op {1 + (nhn ) = (nhn )1/2 [Ũn (β0 ) − E{Ũn (β0 )}]. (2.15) On the other hand, from Lemma 4, we obtain Σ̃(β0 ) = var hZ τ Z (1) h1/2 n Kh (t − r){Z(r) − S̃n (β0 , t) (0) i }dN ∗ (r)dN (t) S̃n (β0 , t) Z Z τn (1) s̃ (β0 , t)⊗2 o = K(z)2 dz s̃(2) (β0 , t) − (0) µ0 (t)dt. s̃ (β0 , t) 0 0 To prove the asymptotic normality, we verify that Lyapunov condition holds. Define 1/2 −1 ψi = (nhn ) n (1) Z Z Khn (t − r){Z(r) − 18 S̃n (β0 , t) (0) S̃n (β0 , t) }dN ∗ (r)dN (t). Similar to the calculation of Σ(β0 ), n X −1/2 E \|ψi − Eψi \|3 = nO{(nhn )3/2 n−3 h−2 }. n } = O{(nhn ) i=1 Thus, h i (nhn )1/2 Ũn (β0 ) − E{Ũn (β0 )} → N {0, Σ̃(β0 )}. Combing with (2.15), we finish the proof of Theorem 2. 2.4 Proof of Corollary 3 To begin with, we have n X ∂ Ũn (β, u) − = n−1 ∂β i=1 Z τ Z 0 u/hn 0 n S̃ (2) (β, u) S̃ (1) (β, u)⊗2 o n n Khn (u − r)dNi∗ (r) (0) − (0) dNi (u). S̃n (β, u) S̃n (β, u)2 Using the similar argument to obtain equation (2.14), we show nZ Z o n S̃ (2) (β, u) S̃ (1) (β, u)⊗2 o n n − (0) dN (u) : \|β − β0 \| < Khn (u − r)dN (r) (0) S̃n (β, u) S̃n (β, u)2 ∗ n (β) n (β) is a P-Glivenko-Cantelli class. Therefore, sup\|β−β0 \|< \| ∂ Ũ∂β \|β=β̃n − E{ ∂ Ũ∂β \|β=β̃n }\| → 0 in n (β) \|β=β̃n conprobability. Since β̃n is consistent for β0 , by continuous mapping theorem, ∂ Ũ∂β h R R P τ ∞ ˆ verges in probability to A(β0 ). Similarly, let Σ̃(β) = n−2 ni=1 0 0 Khn (u − r){Zi (r) − i⊗2 ˆ ˆ Z̃(β, u)}dNi∗ (r)dNi (u) , then sup\|β−β0 \|< \|Σ̃(β) − E{Σ̃(β)}\| → 0 in probability. On the other hand, ˆ E{Σ̃(β)} = n−1 E hZ 0 τ Z ∞ i⊗2 Khn (u − r){Zi (r) − Z̃(β, u)}dNi∗ (r)dNi (u) . 0 19 After change of variables, and by (C4), ˆ E{Σ̃(β)} = 1 nhn Z ∞ Z 2 K(z) dz 0 0 τ {s̃(2) (β0 , u) − s̃(1) (β0 , u)⊗2 }du. s̃(0) (β0 , u) Therefore, p ˆ→ (nhn )Σ̃ Σ̃(β0 ) as nhn → ∞. The consistency of variance estimate follows. References Andersen, P. and Gill, R. (1982), “Cox’s regression models for counting processes: a largesample study,” Ann. Statist, 10, 1100-1120. Fleming, T. R. and Harrington, D. P. (2005), “Counting processes and survival analysis,” Wiley series in probability and statistics. Polland, D. (1984), “Convergence of stochastic processes,” Springer, New York. van der Vaart, A. and Wellner, J. (1996), “Weak convergence and empirical processes,” Springer, New York. 20
https://openalex.org/W4297463498	https://dergipark.org.tr/tr/download/article-file/2522612	English	null	Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry	Academic Platform journal of engineering and smart systems	2,022	cc-by	6,490	Research Paper Academic Platform Journal of Engineering and Smart Systems (APJESS) 10(3), 115-123, 2022 https://doi.org/10.21541/apjess.1139862 Received: 02-Jul-2022 Accepted: 23-Jul-2022 ISSN: 2822-2385 Research Paper Academic Platform Journal of Engineering and Smart Systems (APJESS) 10(3), 115-123, 2022 https://doi.org/10.21541/apjess.1139862 Received: 02-Jul-2022 Accepted: 23-Jul-2022 ISSN: 2822-2385 Research Paper Research Paper Academic Platform Journal of Engineering and Smart Systems (APJESS) 10(3), 115-123, 2022 https://doi.org/10.21541/apjess.1139862 Received: 02-Jul-2022 Accepted: 23-Jul-2022 ISSN: 2822-2385 Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry 1Levent ÇALLI, 2Sena KASIM 1 Department of Management and Organization, Sakarya University, Sakarya, Turkey, lcalli@sakarya.edu.tr 2 Department of Information Systems Engineering, Sakarya University, Sakarya, Turkey, senakasim@gmail.com Abstract Companies must retain their customers and maintain long-term relationships in industries with intense competition. Customer churn analysis is defined in the literature as identifying customers who may leave a company to take appropriate marketing precautions. While customer churn research is prevalent in B2C (Business to Customer) business models such as the telecoms and retail sectors, customer churn analysis in B2B (business to business) models is a relatively emerging topic. In this regard, the study carried out a customer churn analysis by considering an ERP (enterprise resource planning) company with a software as a service (SaaS) business model. Different machine learning algorithms analyzed ten features determined by selection methods and expert opinions. According to the analysis results, the random forest algorithm gave the best result. Additionally, it has been observed that the number of products and customer features has a relatively higher weight for the prediction of churner. Keywords: Customer Churn, SaaS, Machine Learning, Random Forest, Data Mining 1. INTRODUCTION customers commonly make larger and more frequent purchases, for instance, compared with the telecommunication or supermarket industries. Hence, customers are more valuable for companies operating the B2B model, and customer retention is central to developing long relationships [18]. In an era where the reduction of costs is an essential factor coupled with intense competitive pressure, organizations must fully maximize their existing customer; therefore, customer retention efforts are implemented to detect decreasing loyalty of the customers with churn analysis [1]. Dissatisfaction, high cost, poor quality, lack of features, privacy concerns, or many different features can be caused the loss of customers. Hence, identifying these features that change according to the industries that decrease loyalty of customers and making reasonable efforts by companies will contribute to a positive change in the situations of customers who are likely to churn [2]. Consequently, loyal customers make more purchases, pay a premium price, and acquire new customers through favorable word-of-mouth, which positively impacts the company’s long-term reputation [3]. Any company that wants to survive in business cannot simply ignore the churner and loyalty concept. In this regard, this study aims to fill the literature gap with a customer churn approach considering the SaaS (software as a service) industry. This research analyzed customer churn using various machine learning algorithms in a cloud ERP company. Furthermore, feature selection techniques were used, and more effective ones for customer churn analysis were identified. The research findings are expected to have academic and practical benefits, particularly in the SaaS field, where few studies exist. Corresponding Author 2.1. Churn Analysis [22] • Logistic Regression • Support Vector Machine • Decision Tree • Random Forest • Chi Square • Anova • 23 Features • 1788 Observations Inventory Management • Random Forest performed best with 92% accuracy • The number of transactions is the feature with the highest weight. [23] • Logistic Regression • Random Forest No • 43 Feature under 4 Categories • 8869 Observations • Random oversampling and undersampling methods were used. Cloud-based business phone system and call center • Successful results with little data could not be obtained for both algorithms considering precision and recall ratios. • The features with the highest weight are the number of users, number of integrations and call quality. The customer churn study conducted for the banking industry shows that customers who use more banking services have become more loyal. Customers who use less than three services are the group that needs attention [13]. In the study conducted by Keramati et al. [12] in the banking industry, variables such as the number of mobile, internet, and telephone bank transactions and demographic variables such as age, gender, and educational differences are influential on customer churn prediction. potential will bring more competition, it is vital to determine the factors clients consider in their SaaS preferences to retain and gain new customers. Allen [28] states that a SaaS firm’s 3% and 8% monthly customer churn is average. In this respect, monitoring customer churn rates through data mining approaches and intervention to prevent loss are essential for this highly competitive market. In the academic literature, churn analysis studies tend to focus on the telecommunications, financial services, and retail sectors as B2C business models, while there are just a few studies in the SaaS industry as a B2B business model. These studies are shown in Table 1. While customer churn analysis-related studies conducted in the field of B2C are common in the literature, there are few studies on companies using the B2B business model, where the SaaS companies are one of them. The following section discusses the concept of SaaS and research done in this area. Ge et al. [20] performed churn analysis using logistic regression, random forests, and XGBoost algorithms in their study conducted with the data of a SaaS company, considering 8256 observations and 21 features. They observed that Logistic regression and XGBoost algorithms made relatively good predictions for churner. 2.1. Churn Analysis [21] • Long short-term memory (LSTM) • Convolutional Neural Network (CNN) • Support Vector Machine • Random Forest No • 5 Feature Categories • No data size information Advertising • Random Forest and Support Vector Machine performed best with 83% and %81 accuracies, respectively. • The LSTM and CNN, as the deep learning methods, performed poorly due to a lack of data. • As platform usage data, the customer’s minutes spent on the platform and the number of active users carries the most weight. [22] • Logistic Regression • Support Vector Machine • Decision Tree • Random Forest • Chi Square • Anova • 23 Features • 1788 Observations Inventory Management • Random Forest performed best with 92% accuracy • The number of transactions is the feature with the highest weight. [23] • Logistic Regression • Random Forest No • 43 Feature under 4 Categories • 8869 Observations • Random oversampling and undersampling methods were used. Cloud-based business phone system and call center • Successful results with little data could not be obtained for both algorithms considering precision and recall ratios. • The features with the highest weight are the number of users, number of integrations and call quality. Table 1. Academic literature review on churn analysis in SaaS Table 1. Academic literature review on churn analysis in SaaS Study Methods Feature Selection Features and Data Size Sector Findings [20] • Logistic Regression • Random Forest • XGBoost No • 21 Features • 8,256 Observations No Information • Logistic Regression and XGboost performed relatively well with %72 and %75 AUC scores, respectively. • The number of Type-A User Login is the feature with the highest weight. [21] • Long short-term memory (LSTM) • Convolutional Neural Network (CNN) • Support Vector Machine • Random Forest No • 5 Feature Categories • No data size information Advertising • Random Forest and Support Vector Machine performed best with 83% and %81 accuracies, respectively. • The LSTM and CNN, as the deep learning methods, performed poorly due to a lack of data. • As platform usage data, the customer’s minutes spent on the platform and the number of active users carries the most weight. 2.1. Churn Analysis Additionally, it has been discovered that online usage behaviors, such as login and project numbers, strongly predict customer churn. 2.1. Churn Analysis Churn analysis has been carried out in different industries in the academic literature. The most intense studies focus on B2C business models such as telecommunication [2], [4]– [10], financial services [11]–[14], and retail [15]–[17] industries. However, few studies have been conducted considering the B2B business model. In B2B markets, fewer Today, our ability to store and analyze all types of data as a result of the information society due to the rapid development of information and communication technologies (ICT) in recent years has resulted in the acquisition of valuable knowledge through data mining * Corresponding Author https://dergipark.org.tr/tr/pub/apjess https://dergipark.org.tr/tr/pub/apjess Attribution (CC BY) license https://creativecommons.org/licenses/by/4.0/ Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry approach. Telecom is the most widely researched industry in this field. For example; Huang et al. [4] indicate that different features and machine learning methods can effectively analyze customer churn. Ahn et al. [7] show that customer transaction and billing data are important factors for customer churn in mobile telecom services. Verbeke et al. [6] present that a small number of datasets can predict churner with high accuracy and Bhattacharyya and Dash [10] discuss the customer churn analysis in the telecom industry from a bibliometric perspective. methods [19]. Customer churn analyses using the data collected from the customer are often used today within the framework of the loyalty concept, which is vital for all businesses to achieve long-term relationships with their customers. In this way, companies can identify customers with the potential to leave them using various data mining and machine learning methods to effort the necessary marketing activities for this group called churners. In the literature, companies using the B2C business model have commonly been analyzed using the customer churn Table 1. Academic literature review on churn analysis in SaaS Study Methods Feature Selection Features and Data Size Sector Findings [20] • Logistic Regression • Random Forest • XGBoost No • 21 Features • 8,256 Observations No Information • Logistic Regression and XGboost performed relatively well with %72 and %75 AUC scores, respectively. • The number of Type-A User Login is the feature with the highest weight. 3.1. Feature Selection In data mining, feature selection is the preferred technique to reduce dataset size to achieve more efficient analysis and adapt the dataset better to match the preferred analysis method [29]. This study used the chi-square, information gain, gain ratio, and Gini index methods to determine the features to be considered in the churn analysis. 𝑥𝑐 2 = ∑(𝑂𝑖−𝐸𝑖)2 𝐸𝑖 (1) (1) In formula (1), c represents the degrees of freedom, O the observed values, and E the expected value [31]. Figure 1. Research process Figure 1. Research process categorical variables [30]. The formula for the Chi-square is as follows. 2.2. Software-as-a-Service (SaaS) SaaS is a business model that offers cloud-based services to clients as an alternative to standalone software that requires installation, maintenance, IT infrastructure, and support services (backup, upgrade, security), especially for B2B [24]. According to Fortune Business Insights (2022)’s report [25] and Jones [27], the global SaaS marketplace is expected to grow from $130 billion in 2021 to $716 billion in 2028, with Microsoft, Salesforce, Adobe, SAP, and Oracle accounting for 51% of the market. Since the market’s growth In another study conducted in the scope of SaaS, Rautio [21] performed churn analysis with different machine learning methods, considering a company operating in the advertising industry. The research’s features mainly focused on client Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 116 Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry business data, such as spending, platform usage, previous customer service interactions, and service-related customer feedback. In the study, where the support vector machine algorithm gave the best results, it was determined that platform usage metrics feature relatively more weight in prediction. The study also found that deep learning methods did not perform well in predicting churner. revealed that neither the random forest approach nor logistic regression could achieve sufficient accuracy rates due to strongly overfitting [23]. The number of users, number of integrations, and call quality features were shown to be more critical in predicting churners with the random forest algorithm when random oversampling and undersampling approaches were utilized. 3. METHOD Amornvetchayakul and Phumchusri [22] used four machine learning algorithms in their study with 1778 samples for churn analysis, considering a SaaS company that provides inventory management services (inventory levels, purchase orders, delivery, etc.) for SMEs (Small and Medium Enterprises). The number of transactions in the current month and the number of transactions in the previous month features were discovered as having considerably higher weights than other features in their research, which considered 23 features. The prediction of the random forest algorithm has stated that it has higher accuracy than the decision tree, logistic regression, and support vector machine algorithms. This research focused on a software company based in Germany and Turkey that specialized mainly in the ERP (enterprise resource planning) industry. A total of 1951 observations were analyzed. Initially, the weight of sixteen features was evaluated, and then the prediction phase that considered churn and non-churn customers was performed using various machine learning methods. The feature selection stage used the Chi-square, information gain, gain ratio, and Gini index methods. Decision Tree, Logistic Regression, Naive Bayes, K-NN, Random Forrest, and Neural Networks algorithms were utilized as classification methods. This process is shown in Figure 1. Lastly, a study focused on churn analysis of a firm in the cloud-based business phone system and call center industry Figure 1. Research process Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 3.1.2. Information Gain and Gain Ratio The Chi-square test, a non-parametric method, is used to examine whether there is a relationship between two Information gain is used to identify the best features that provide the most information about a class and uses the idea Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 117 Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry of entropy which is defined as a measure of purity or the degree of uncertainty of a random variable [32]. The information gain calculates the entropy difference before and after the division and determines the purity of the in-class elements. The entropy is calculated using the following formula (2) [33]; 3.1.3. Gini Index As an impurity splitting method, Gini Index is appropriate for binary and continuous numeric values for calculating feature weight. Assume that S is the collection of s samples with m different classes (Ci ,i=1,…m). S can be divided into m subsets based on class differences (Si ,i=1,…m). If Si is the sample set for class Ci, and Si is the number of samples in Si, then the Gini index of S is calculated with the following formula (4) [35]; 𝐻(𝑆) = −∑𝑝𝑖 𝑙𝑜𝑔2 𝑝𝑖 𝑁 𝑖=1 (2) (2) 𝐺𝑖𝑛𝑖 𝐼𝑛𝑑𝑒𝑥 (𝑆) = 1 −∑𝑃𝑖 2 𝑚 𝑖=1 (4) S: Set of all examples in the dataset N: Number of distinct class values pi: Event probability (4) In this formula, Pi refers to the probability that each given sample belongs to Ci and is measured with si/s. When Gini Index has a minimum value of 0, it indicates that all members of the set fall under the same class, indicating that it can gather the most valuable information [35]. IBM SPSS and Orange Data Mining software were utilized for feature selection algorithms [36], [37]. The information gain is calculated with the formula (3) shown below [33]; The information gain is calculated with the formula (3) shown below [33]; 𝐼𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝐺𝑎𝑖𝑛(𝐴, 𝑆) = 𝐻(𝑆) −∑ \|𝑆𝑗\| \|𝑆\| 𝑣 𝑗=1 . 𝐻(𝑆𝑗) = 𝐻(𝑆) −𝐻(𝐴, 𝑆) (3) (3) H(S): Entropy of the whole dataset S H(S): Entropy of the whole dataset S \|Sj\|: Number of the instance with j value of an attribute A \|S\|: Total number of instances in dataset S v: Set of distinct values of an attribute A H(Sj): Entropy of subset of instances for attribute A H (A,S): Entropy of an attribute A H(S): Entropy of the whole dataset S H(S): Entropy of the whole dataset S \|Sj\|: Number of the instance with j value of an attribute A \|S\|: Total number of instances in dataset S v: Set of distinct values of an attribute A H(Sj): Entropy of subset of instances for attribute A H (A,S): Entropy of an attribute A Table 2 shows the ranking of features according to the four feature selection methods. Ten key features were selected for churn analysis within the scope of analysis results and expert opinion. 3.2.2. Logistic Regression The random forest algorithm is applied to a wide range of prediction problems as a well-known approach in the literature due to its capacity to handle small sample sizes and as well as high-dimensional feature spaces, with a few parameters for tuning [42]. The random forest comprises many independent decision trees that act as an ensemble method, and each tree in the random forest generates a class prediction. The model’s prediction is based on the class with the most votes, as seen in Figure 4 [43]. Logistic regression is a commonly used statistical model for predicting event probability. In the Logistic regression model seen in detail in formula 5, the dependent variable y is a binary in the formula used to determine whether an event occurred. 𝑝𝑟𝑜𝑏(𝑦= 1) = 𝑒𝛽0+∑ 𝛽𝑘𝑥𝑘 𝐾 𝑘=1 1 −𝑒𝛽0+∑ 𝛽𝑘𝑥𝑘 𝐾 𝑘=1 (5) (5) The independent inputs are x1, x2,....,xk. The maximum likelihood approach can estimate β1, β2,…., βk as the regression coefficients based on the available training data [4]. Figure 4. Random Forest 3.2. Predictive Analysis Algorithms The gain ratio method is estimated by dividing the information gain value by the entropy value to get accurate results due to the asymmetrical nature of the information gain results [34]. Churn analysis was performed with Decision Tree, Logistic Regression, Naive Bayes, K-NN, Random Forest, and Neural Networks algorithms which are common approaches in the literature for this study. Each algorithm shortly explains in this section. Table 2. Feature weight ranking by selection methods Table 2. Feature weight ranking by selection methods Ranking Chi-square Gini Index Information Gain Gain Ratio 1 Number of Invoices The number of customers The number of customers Number of Cash Register Connections 2 Number of Offers Number of Invoices Number of Invoices The number of customers 3 Number of Support Number of products Number of Offers Custom Report Usage 4 The number of customers Number of Offers Number of Support Mail Connection 5 Number of products Number of Support Number of products Number of Offers 6 Number of Cash Register Connections Number of Users Number of Users Number of Users 7 Cargo Usage Cargo Usage Number of Cash Register Connections Number of Invoices 8 Number of Payment Documents Number of Cash Register Connections Cargo Usage Number of Support 9 Custom Report Usage Custom Report Usage Custom Report Usage Cargo Usage 10 Mail Connection Number of Payment Documents Number of Payment Documents Number of Payment Documents 11 Number of Cash Register Receipts Number of Orders Mail Connection Number of products 12 Number of Users Number of Cash Register Receipts Number of Orders Number of Cash Register Receipts 13 Number of Orders Mail Connection Number of Cash Register Receipts Number of Production Orders 14 Customer Group Customer Group Customer Group Number of Orders 15 Number of Production Orders Number of Production Orders Number of Production Orders Customer Group 16 Number of Marketplaces Number of Marketplaces Number of Marketplaces Number of Marketplaces Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 118 Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry 3.2.1. Decision Tree The K-Nearest Neighbors algorithm is an easy-to- implement, simple with few hyperparameters required, supervised learning algorithm that produces classifications or predictions for clustering of a single data point using proximity techniques such as Euclid, Manhattan, Minkowski, and Hamming [40]. The decision tree is one of the most popular data categorization techniques in literature, which is based on the information gain theory explained in the previous section. The decision tree is one of the most popular data categorization techniques in literature, which is based on the information gain theory explained in the previous section. Nodes and branches create the decision tree structure. Nodes represent tests on certain features, while branches indicate test results. Nodes and branches create the decision tree structure. Nodes represent tests on certain features, while branches indicate test results. Figure 2. Decision Tree model Figure 3. The K-Nearest Neighbors algorithm Figure 3. The K-Nearest Neighbors algorithm Figure 2. Decision Tree model Figure 2. Decision Tree model Figure 2. Decision Tree model Figure 3. The K-Nearest Neighbors algorithm Although the decision tree technique has several alternative approaches, the most well-known C5 method splits the sample using the feature with the maximum information gain. It then repeats this process until the subset can no longer be divided [17]. An example decision tree model is shown in Figure 2. K-NN uses most k nearest neighbors for a new data point whose class is being looked for to assign [41]. This situation is seen in Figure 3. For instance, the red shape containing an unclassified data point may belong to either class A or B according to the alternative k values. 3.2.3. Naïve Bayes The Naive Bayes algorithm is a well-known classification method used in the machine learning literature. It has attracted much interest due to its ease of use and good performance [38]. The formula (6) is shown below. 𝑃(𝑐\|𝑥) = 𝑃(𝑐\|𝑥)𝑃(𝑐) 𝑃(𝑥) (6) Figure 4. Random Forest Figure 4. Random Forest (6) 3.2.6. Neural Networks Bayes theorem estimates the posterior probability of class given predictor, P(c\|x), from P(c) (the prior probability of a class), P(x) (the prior probability of predictor), and P(x\|c) (represents the likelihood, which is the probability of predictor class given) [39]. Bayes theorem estimates the posterior probability of class given predictor, P(c\|x), from P(c) (the prior probability of a class), P(x) (the prior probability of predictor), and P(x\|c) (represents the likelihood, which is the probability of predictor class given) [39]. Neural networks are a method of explaining cognitive, decision-making, and other intelligent control behaviors by using the way the human brain operates as a kind of data processing and analysis [17]. A classic neural network consists of three layers: an input layer, a hidden layer, and an output layer, all connected by neurons, as seen in Figure 5. Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 119 Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry customers (1). In this regard, it is thought that the dependent variable has a balanced distribution. Figure 5. Neural Networks Correlation analysis, an essential part of descriptive statistics, was done in Python to reveal the relationships between variables. Figure 6 shows the correlation of features with each other and the customer churn as the dependent variable. A strong positive correlation is shown by dark colors, whereas light tones indicate a weak positive correlation. Figure 6. Correlation heatmap Figure 5. Neural Networks A key advantage of utilizing neural networks for data modeling is that they can fit complicated nonlinear models without needing to be specified in advance, unlike other nonlinear estimation methods [44]. 4. RESULTS As a result of the feature selection analysis stage and considering the opinions of industry experts, it was determined that ten features would be appropriate for churn analysis. Table 3 shows these features and their descriptions. Table 3. Selected features Features Description Number of products The number of products registered in the customer’s account. Number of Customers The total number of customers and suppliers listed on the customer’s account. Number of Offers The number of offers that the customer creates in the account. Number of Orders The number of orders created by the customer. Number of Invoices The number of invoices created by the customer. Cargo Usage The number of cargo companies the customer has used in the account. Number of Users The number of users the customer has in the ERP software that can access the system at the same time. Custom Report Usage The reports are specially made for the customer, excluding the standard reports in the ERP software. Number of Cash Register Receipts The number of cash register receipts created by the customer in the ERP software. Email Connection Email connection status used for the proposal side of the customer’s ERP software. Table 3. Selected features Table 3. Selected features Figure 6. Correlation heatmap Figure 6. Correlation heatmap The normalization process was conducted in the next step by assigning values between 0 and 1 due to the features’ value variations. This method allows using a single scale while keeping the distinctions in the value ranges and avoiding information loss of each feature. Table 4. Number of training and test data Training Dataset Test Dataset %75-%25 1463 488 %70-%30 1365 586 Table 4. Number of training and test data In the next stage, the dataset was divided into test and training data. Then, the model was tested with different algorithms. This study used two different ratios of training data, 70% and 75%. The number of customers that create the training and test data sets is shown in Table 4. Table 5. Churn Analysis results Predictive Algorithms %75-%25 %70-%30 Accuracy Accuracy Decision Tree 74.59% 73.54% Logistic Regression 57.58% 57.84% Naive Bayes 47.95% 48.12% K-NN 65.36% 74.23% Random Forest 78.27% 77.47% Neural Network 57.99% 58.19% Table 5. Churn Analysis results The churn status of customers was considered in this study based on their active usage of the software. 4. RESULTS 836 customers in the data set are coded as churn (0), and 1115 are active As seen in Table 5, the churn analysis results with the training and test data in both ratios show that the random forest algorithm gives the relatively best results with higher accuracy rates. The results also show that the decision tree Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 120 Levent Çallı, Sena Kasım Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry Phumchusri [22] and Sergue [23]. For example, Amornvetchayakul and Phumchusri [22] indicate that the importance of the number of transactions in the inventory management sector is relatively higher than in others. Additionally, Sergue [23] highlights that the number of users in the Cloud-based business phone system and call center industry is relatively higher weight than other features in predicting churner. In terms of prediction accuracy, the random forest algorithm achieves the best result in this study, which is similar to the results observed by Rautio [21] and Amornvetchayakul and Phumchusri [22]. algorithm produces better results than other algorithms. Additionally, the K-NN method provides much better accuracy when the number of training data sets is decreased. The weights of the features used in the predictive algorithms for the random forest and decision tree are shown in Table 6. Table 6. Features weight Decision Tree Random Forest Feature Weight Feature Weight Number of Customer 45,0% Number of Customer 42,4% Number of Products 29,6% Number of Product 33,3% Number of Invoice 15,4% Number of Invoice 9,4% Number of Orders 4,0% Number of Orders 5,4% Number of Users 2,2% Number of Offer 3,1% Cargo Usage 1,7% Cargo Usage 2,8% Number of Offer 1,3% Number of Users 2,8% Custom Report Usage 0,5% Number of Cash Register Receipts 0,3% Mail connection 0,2% Mail Connection 0,3% Number of Cash Register Receipts 0,0001% Custom Report Usage 0,3% Table 6. Features weight Table 6. Features weight As a result, the churn analysis processes of B2B business models, which have less data for many businesses compared to B2C business models, undoubtedly involve many difficulties. In this study, the customer churn analysis process was carried out for a cloud ERP company and contributed to the academic literature and the practical field. 4. RESULTS Although it is an important gap that there are very few studies in this literature, it is thought that new studies will emerge in the light of the findings of this study. Using different features and testing different predictive algorithms will enrich the academic literature and be guiding for managers. Author Contributions: Concept – L.C., S.K.; Data Collection &/or Processing – S.K; Literature Search – L.C., S.K.; Writing – L.C. In this respect, it is seen that the number of customers and products has a high weight in both algorithms predicting the customer’s churn status. Mail connection, custom report usage, number of cash register receipts, cargo usage, number of the offer, number of users, and number of orders have low weight, while the number of invoice feature has a relatively moderate weight in both algorithms. Conflict of Interest: This study was produced from the master's thesis entitled "Customer Churn Analysis with Data Mining Methods: Software Industry," conducted at Sakarya University, Institute of Natural Sciences, with the supervisor Levent ÇALLI. Financial Disclosure: The authors declared that this study has received no financial support. Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 10.1109/ICCCBDA49378.2020.9095611. [6] W. Verbeke, K. Dejaeger, D. Martens, J. Hur, and B. Baesens, “New insights into churn prediction in the telecommunication sector: A profit driven data mining approach,” Eur. J. Oper. Res., vol. 218, no. 1, pp. 211– 229, 2012, doi: 10.1016/j.ejor.2011.09.031. [18] B. Janssens, M. Bogaert, A. Bagué, and D. Van den Poel, “B2Boost: instance-dependent profit-driven modelling of B2B churn,” Ann. Oper. Res., 2022, doi: 10.1007/s10479-022-04631-5. [7] J. H. Ahn, S. P. Han, and Y. S. Lee, “Customer churn analysis: Churn determinants and mediation effects of partial defection in the Korean mobile telecommunications service industry,” Telecomm. Policy, vol. 30, no. 10–11, pp. 552–568, 2006, doi: 10.1016/j.telpol.2006.09.006. [19] W. Verbeke, D. Martens, C. Mues, and B. Baesens, “Building comprehensible customer churn prediction models with advanced rule induction techniques,” Expert Syst. Appl., vol. 38, no. 3, pp. 2354–2364, 2011, doi: 10.1016/j.eswa.2010.08.023. [20] Y. Ge, S. He, J. Xiong, and D. E. Brown, “Customer churn analysis for a software-as-a-service company,” in 2017 Systems and Information Engineering Design Symposium, SIEDS 2017, 2017, pp. 106–111, doi: 10.1109/SIEDS.2017.7937698. [8] C. F. Tsai and Y. H. Lu, “Customer churn prediction by hybrid neural networks,” Expert Syst. Appl., vol. 36, no. 10, pp. 12547–12553, 2009, doi: 10.1016/j.eswa.2009.05.032. [9] T. Vafeiadis, K. I. Diamantaras, G. Sarigiannidis, and K. C. Chatzisavvas, “A comparison of machine learning techniques for customer churn prediction,” Simul. Model. Pract. Theory, vol. 55, pp. 1–9, 2015, doi: 10.1016/j.simpat.2015.03.003. [21] A. Rautio, “Churn rediction in SaaS using Machine Learning,” 2019. [22] P. Amornvetchayakul and N. Phumchusri, “Customer Churn Prediction for a Software-as-a-Service Inventory Management Software Company: A Case Study in Thailand,” in 2020 IEEE 7th International Conference on Industrial Engineering and Applications, ICIEA 2020, 2020, pp. 514–518, doi: 10.1109/ICIEA49774.2020.9102099. [10] J. Bhattacharyya and M. K. Dash, “What Do We Know About Customer Churn Behaviour in the Telecommunication Industry? A Bibliometric Analysis of Research Trends, 1985–2019,” FIIB Bus. Rev., 2021, doi: 10.1177/23197145211062687. [23] M. Sergue, “Customer Churn Analysis and Prediction using Machine Learning for a B2B SaaS company,” 2020, [Online]. Available: www.kth.se/sci. [11] Y. Xie, X. Li, E. W. T. Ngai, and W. Ying, “Customer churn prediction using improved balanced random forests,” Expert Syst. Appl., vol. 36, no. 3 PART 1, pp. 5445–5449, 2009, doi: 10.1016/j.eswa.2008.06.121. [24] D. Ma, “The Business Model of ‘Software-As-A- Service,’” in IEEE International Conference on Services Computing (SCC 2007), 2007, no. July, pp. 701–702, doi: 10.1109/SCC.2007.118. [12] A. Keramati, H. Ghaneei, and S. REFERENCES Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 121 Levent Çallı, Sena Kasım Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry REFERENCES Customer churn is a severe challenge to any business, and one way to deal with it is to predict which customers are most likely to leave the company and then target those segments with marketing efforts to encourage them to be loyal [45]. Customer loyalty is valuable since acquiring new customers is more costly than keeping existing customers, and a loyal customer serves as an honorary lawyer for the company, becoming a positive reference for potential customers and, as a result, generating more profit [19]. While academic studies on customer churn analysis are popular in a few industries, especially in B2C fields such as telecommunications, grocery retail, or bank, research on software as a service (SaaS) as a B2B industry is very limited. Thus, the study aimed to fill a gap in the academic literature by focusing on a cloud ERP business. [1] N. Glady, B. Baesens, and C. Croux, “Modeling churn using customer lifetime value,” Eur. J. Oper. Res., vol. 197, no. 1, pp. 402–411, 2009, doi: 10.1016/j.ejor.2008.06.027. [2] A. Amin, F. Al-Obeidat, B. Shah, A. Adnan, J. Loo, and S. Anwar, “Customer churn prediction in telecommunication industry using data certainty,” J. Bus. Res., vol. 94, no. February 2018, pp. 290–301, 2019, doi: 10.1016/j.jbusres.2018.03.003. [3] J. Ganesh, M. J. Arnold, and K. E. Reynolds, “Understanding the customer base of service providers: An examination of the differences between switchers and stayers,” J. Mark., vol. 64, no. 3, pp. 65–87, 2000, doi: 10.1509/jmkg.64.3.65.18028. [4] B. Huang, M. T. Kechadi, and B. Buckley, “Customer churn prediction in telecommunications,” Expert Syst. Appl., vol. 39, no. 1, pp. 1414–1425, 2012, doi: 10.1016/j.eswa.2011.08.024. According to the research findings, it is seen that the number of customers and the number of products are the most important features in customer churn analysis. The number of invoices and orders features also shows a relatively higher weight than others. Thus, research findings reveal that features of the fundamental business process have a meaningful relationship with customer churn and are factors that need more attention for managers. In this sense, the results show some parallelism with Amornvetchayakul and [5] K. Kim, C. H. Jun, and J. Lee, “Improved churn prediction in telecommunication industry by analyzing a large network,” Expert Syst. Appl., vol. 41, no. 15, pp. 6575–6584, 2014, doi: 10.1016/j.eswa.2014.05.014. Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 10.1109/ICCCBDA49378.2020.9095611. M. Mirmohammadi, “Developing a prediction model for customer churn from electronic banking services using data mining,” Financ. Innov., vol. 2, no. 1, 2016, doi: 10.1186/s40854-016-0029-6. [25] Fortunebusinessinsights, “The software as a service market Size,” 2022. https://www.fortunebusinessinsights.com/software-as- a-service-saas-market-102222. [13] A. Bilal Zoric, “Predicting Customer Churn in Banking Industry using Neural Networks,” Interdiscip. Descr. Complex Syst., vol. 14, no. 2, pp. 116–124, 2016, doi: 10.7906/indecs.14.2.1. [26] E. Jones, “Cloud Market Share – a Look at the Cloud Ecosystem in 2022,” KINSTA BLOG, 2022. https://kinsta.com/blog/cloud-market- share/#:~:text=The SaaS market is dominated,impressive annual growth of 34%25. [14] J. Burez and D. Van den Poel, “Handling class imbalance in customer churn prediction,” Expert Syst. Appl., vol. 36, no. 3 PART 1, pp. 4626–4636, 2009, doi: 10.1016/j.eswa.2008.05.027. [27] E. Jones, “Cloud Market Share – a Look at the Cloud Ecosystem in 2022,” KINSTA BLOG, 2022. . [15] M. A. de la Llave Montiel and F. López, “Spatial models for online retail churn: Evidence from an online grocery delivery service in Madrid,” Pap. Reg. Sci., vol. 99, no. 6, pp. 1643–1665, 2020, doi: 10.1111/pirs.12552. [28] K. Allen, “Churn Rate vs Retention Rate: How to Calculate These SaaS KPIs,” woopra.com, 2022. https://www.woopra.com/blog/churn-rate-vs- retention-rate. [16] W. Buckinx and D. Van Den Poel, “Customer base analysis: Partial defection of behaviourally loyal clients in a non-contractual FMCG retail setting,” Eur. J. Oper. Res., vol. 164, no. 1, pp. 252–268, 2005, doi: 10.1016/j.ejor.2003.12.010. [29] A. Jović, K. Brkić, and N. Bogunović, “A review of feature selection methods with applications,” 2015 38th Int. Conv. Inf. Commun. Technol. Electron. Microelectron. MIPRO 2015 - Proc., no. May, pp. 1200–1205, 2015, doi: 10.1109/MIPRO.2015.7160458. [17] X. Hu, Y. Yang, L. Chen, and S. Zhu, “Research on a Customer Churn Combination Prediction Model Based on Decision Tree and Neural Network,” 2020 IEEE 5th Int. Conf. Cloud Comput. Big Data Anal. ICCCBDA 2020, pp. 129–132, 2020, doi: [30] A. Field, Discovering Statistics Using IBM SPSS Statistics. SAGE Publications Ltd, 2018. [31] sampath kumar Gajawada, “Chi-Square Test for Feature Selection in Machine learning,” Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 122 Levent Çallı, Sena Kasım Levent Çallı, Sena Kasım Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry Using Machine Learning Algorithms to Analyze Customer Churn in the Software as a Service (SaaS) Industry https://towardsdatascience.com/, 2019. https://towardsdatascience.com/chi-square-test-for- feature-selection-in-machine-learning-206b1f0b8223. Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 2019. [32] N. Tyagi, “What is Information Gain and Gini Index in Decision Trees?,” https://www.analyticssteps.com/, 2021. https://www.analyticssteps.com/blogs/what- gini-index-and-information-gain-decision-trees. [33] U. Krčadinac, “Classification – Decision Trees,” 2015. http://ai.fon.bg.ac.rs/wp- content/uploads/2015/04/Classification-Decision- Trees-2015.pdf. [34] S. K. Trivedi, “A study on credit scoring modeling with different feature selection and machine learning approaches,” Technol. Soc., vol. 63, no. September, p. 101413, 2020, doi: 10.1016/j.techsoc.2020.101413. [35] W. Shang, H. Huang, H. Zhu, Y. Lin, Y. Qu, and Z. Wang, “A novel feature selection algorithm for text categorization,” Expert Syst. Appl., vol. 33, no. 1, pp. 1–5, 2007, doi: 10.1016/j.eswa.2006.04.001. [36] J. Demšar et al., “Orange: Data Mining Toolbox in Python,” J. Mach. Learn. Res., vol. 14, pp. 2349–2353, 2013, [Online]. Available: http://jmlr.org/papers/v14/demsar13a.html. [37] IBM Corp., “IBM SPSS Statistics for Windows, Version 26.0,” 2019. 2019. [38] V. V. Saradhi and G. K. Palshikar, “Employee churn prediction,” Expert Syst. Appl., vol. 38, no. 3, pp. 1999–2006, 2011, doi: 10.1016/j.eswa.2010.07.134. [39] S. Sayad, “An Introduction to Data Science,” 2022. saedsayad.com/data_mining_map.htm. [40] IBM, “What is the k-nearest neighbors algorithm?,” 2022. https://www.ibm.com/topics/knn. [41] O. Kramer, Dimensionality Reduction with Unsupervised Nearest Neighbors, vol. 51. 2013. [42] G. Biau and E. Scornet, “A random forest guided tour,” Test, vol. 25, no. 2, pp. 197–227, 2016, doi: 10.1007/s11749-016-0481-7. [43] T. Yiu, “Understanding Random Forest,” https://towardsdatascience.com/, 2019. https://towardsdatascience.com/understanding- random-forest-58381e0602d2. [44] D. J. Livingstone, D. T. Manallack, and I. V. Tetko, “Data modelling with neural networks: Advantages and limitations,” J. Comput. Aided. Mol. Des., vol. 11, no. 2, pp. 135–142, 1997, doi: 10.1023/A:1008074223811. [45] S. A. Neslin, S. Gupta, W. Kamakura, L. U. Junxiang, and C. H. Mason, “Defection detection: Measuring and understanding the predictive accuracy of customer churn models,” J. Mark. Res., vol. 43, no. 2, pp. 204– 211, 2006, doi: 10.1509/jmkr.43.2.204. Academic Platform Journal of Engineering and Smart Systems 10(3), 115-123, 2022 123
https://openalex.org/W3190829675	http://dspace.stir.ac.uk/bitstream/1893/33072/1/s12868-021-00653-0.pdf	English	null	Stressor controllability modulates the stress response in fish	BMC neuroscience	2,021	cc-by	10,180	Abstract Background: In humans the stress response is known to be modulated to a great extent by psychological factors, particularly by the predictability and the perceived control that the subject has of the stressor. This psychological dimension of the stress response has also been demonstrated in animals phylogenetically closer to humans (i.e. mam- mals). However, its occurrence in fish, which represent a divergent vertebrate evolutionary lineage from that of mam- mals, has not been established yet, and, if present, would indicate a deep evolutionary origin of these mechanisms across vertebrates. Moreover, the fact that psychological modulation of stress is implemented in mammals by a brain cortical top-down inhibitory control over subcortical stress-responsive structures, and the absence of a brain cortex in fish, has been used as an argument against the possibility of psychological stress in fish, with implications for the assessment of fish sentience and welfare. Here, we have investigated the occurrence of psychological stress in fish by assessing how stressor controllability modulates the stress response in European seabass (Dicentrarchus labrax). Results: Fish were exposed to either a controllable or an uncontrollable stressor (i.e. possibility or impossibility to escape a signaled stressor). The effect of loss of control (possibility to escape followed by impossibility to escape) was also assessed. Both behavioral and circulating cortisol data indicates that the perception of control reduces the response to the stressor, when compared to the uncontrollable situation. Losing control had the most detrimental effect. The brain activity of the teleost homologues to the sensory cortex (Dld) and hippocampus (Dlv) parallels the uncontrolled and loss of control stressors, respectively, whereas the activity of the lateral septum (Vv) homologue responds in different ways depending on the gene marker of brain activity used. Conclusions: These results suggest the psychological modulation of the stress response to be evolutionary con- served across vertebrates, despite being implemented by different brain circuits in mammals (pre-frontal cortex) and fish (Dld-Dlv). Keywords: Stress, Controllability, Cortisol, Immediate early genes, Dorsolateral pallium, Fish welfare was first investigated in laboratory animals, in particular in monkeys, dogs and rodents (e.g. [2–7]) and only sub- sequently extended to humans (e.g. [8–10]). The semi- nal studies of stressor controllability in animal models have compared the stress response in animals that can escape the stressor (i.e. that has control over the stressor, originally a tail-shock) to animals exposed to the same stressor but that have no control over it. © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Cerqueira et al. BMC Neurosci (2021) 22:48 https://doi.org/10.1186/s12868-021-00653-0 Cerqueira et al. BMC Neurosci (2021) 22:48 https://doi.org/10.1186/s12868-021-00653-0 BMC Neuroscience Stressor controllability modulates the stress response in fish Marco Cerqueira1,2, Sandie Millot3, Tomé Silva4, Ana S. Félix5,6, Maria Filipa Castanheira1, Sonia Rey7, Simon MacKenzie7, Gonçalo A. Oliveira5, Catarina C. V. Oliveira1 and Rui F. Oliveira5,6,8* Background It has long been established that psychological fac- tors, such as controllability and predictability, can be as important as intrinsic characteristics of the stressor (e.g. intensity) on the effects of stress in humans [1]. Interest- ingly, this psychological dimension of the stress response Correspondence: ruiol@ispa.pt 5 ISPA—Instituto Universitário, 1149‑041 Lisbon, Portugal Full list of author information is available at the end of the article Abstract The effects of lack of stressor control (aka learned helplessness), include increased anxiety, decreased social exploration, and © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Cerqueira et al. BMC Neurosci (2021) 22:48 Cerqueira et al. BMC Neurosci (2021) 22:48 Page 2 of 12 Page 2 of 12 telencephalic area (Dld), which has been shown to inte- grate multimodal sensory information and is the puta- tive homologue of the sensory cortex [28, 29]; and (4) the ventral nucleus of the ventral telencephalic area (Vv), which has been shown to respond to stimulus salience [33] and is a putative homologue of the lateral septum and of the nucleus accumbens [34, 35]. The role of psychological factors in the regulation of the stress response in fish has been seldom addressed [36], and when it has, the focus has predominantly been on the effects of stressor predictability [37–40], in which species-specific neural responses have been reported [[e.g. increased activity in Vv in seabream, [33]; and increased activity in Dm and decreased activity in Dlv in Sea Bass, in which Vv was not sampled, [41]]. To the best of our knowledge, the effect of controllability in fish has been only addressed in a study in which rain- bow trout (Oncorhynchus mykiss) had the chance to actively avoid being defeated by a larger conspecific in a conditioning paradigm. In this study, the trout that could escape the social stressor exhibited a lower cor- tisol response to the conditioned stimulus than those that cannot escape social defeat [42]. However, this evi- dence has not yet been investigated as supporting the occurrence of psychological stress in fish, and its neural bases have not been investigated yet in fish. heightened fear conditioning and delayed fear extinction [1]. These effects are not observed in animals that have controllability over the stressor. Similarly, in humans, laboratory tasks that induce uncontrollable (but not controllable) stress impair fear extinction and executive functioning in a Stroop task [11, 12], and increase per- ceived helplessness, depression and anger [13, 14]. Research on the neural mechanisms of stressor con- trollability suggests that the blunted stress response induced by perceived stressor control is due to a corti- costriatal circuit involving the ventral medial pre-frontal cortex (mPFC) and the posterior dorsomedial striatum that exerts top-down inhibitory control over subcorti- cal stress-responsive structures, namely the serotoner- gic system of the dorsal raphe nucleus and the amygdala [15]. At the neuroendocrine level, the hypothalamic– pituitary–adrenal (HPA) axis response to stress has been shown to be influenced by controllability in humans, where in response to laboratory tasks, cortisol levels are lower in groups where stressor control is perceived than in groups where stressor control is not perceived [14] and perceived control is negatively correlated with cortisol [16, 17]. In contrast, in rodents the HPA stress response seems insensitive to stressor controllability, but repeated exposure to controllable stressors leads to a lower response [18–20]. Given the central role of the mPFC on stressor controllability, it could be questioned whether this psychological moderation of the stress response is also present in non-mammals that lack a cerebral cor- tex. In this respect, querying its occurrence in fish, which represent a phylogenetically divergent vertebrate lineage from that of mammals [21], whose brains lack a mPFC homologue [22], is of major relevance. Its presence in fish would indicate that stressor controllability is an ancestral psychological trait that can be implemented by different neural substrates in fish and mammals. In fact, the com- putations needed for similar cognitive abilities can be implemented by similar neuronal circuitry irrespective of its organization in a layered or nuclear architecture. This has recently been shown by the involvement of the meso- pallium and nidopallium in the complex cognitive abili- ties of birds [23–25]. g yi Here, we investigated the occurrence of stressor controllability in the European seabass (Dicentrarchus labrax), which is a key species for European aquacul- ture. Effect of stressor controllability on patterns of brain activation and neurogenomic statesf The expression of all genes in the Dlv was positively cor- related both with cortisol and with freezing behavior, and negatively correlated with exploratory behaviour (Table 2). A negative correlation was observed between cortisol and the expression of bdnf in the Vv. Finally, there were positive correlations between cortisol and freezing behavior and escape attempts (Table 2). Stressor controllability elicited different patterns of brain activation towards the stressor, as measured by the expression of immediate early genes, used as indicators of brain activity (Table 1; Fig. 2). In the Dlv, the loss of stressor control induced higher activity than either expo- sure to a controlled stressor (as indicated by the expres- sion of all IEGs), or exposure to an uncontrolled stressor (as indicated by the expression of c-fos, bdnf and npas1). In the Dld, the loss of stressor control induced higher activity than exposure to an uncontrolled, but not to a controlled stressor (as indicated by the expression of all Effect of stressor controllability on cortisolif Stressor controllability significantly affected cortisol lev- els (Table 1): control over the stressor decreased cortisol levels, whereas loss of controllability induced higher cor- tisol levels than uncontrolled stress itself (Fig. 1f). Effect of stressor controllability on behavior Effect of stressor controllability on behavior Analysis of the behavior from the test session showed a significant effect of controllability on all behavioral vari- ables measured (Table 1; Fig. 1a–d). Freezing was higher in the loss of control treatment, than in the uncontrolled or controlled treatments, and the latter two treatments also differed significantly (Fig. 1a). The number of escape attempts and shoal cohesion showed a similar pattern with both the uncontrolled and loss of control treat- ments, showing higher levels that the control treatment (Fig. 1b, c). Conversely, exploration was significantly lower in the loss of control treatment than in any of the other two treatments (Fig. 1d). The neurogenomic states, as characterized by the pat- tern of co-expression of the immediate early genes on each brain region, are specific for each experimental treatment (i.e. CTR vs. UnCTR vs. CTRUn) in Dm, Dld and Dlv (Fig. 3). The neurogenomic state of Vv is similar between controllable and uncontrolled treatments (see Additional file 1: Table S1 for detailed information on QAP correlations, used to infer significance differences between pairs of co-expression matrices). Thus, the results of this study will not only allow testing the hypothesis of the deep evolutionary origin of stressor controllability effects on the stress response and its brain substrates, but also has direct implica- tions for the welfare of farmed fish. For this purpose, we used a conditioning protocol to associate a condi- tioned stimulus (CS = light) with a stressor (US = con- finement) under different conditions: (1) Controllable stressor (CTR)—fish had the choice to escape from US by a door; (2) Uncontrollable stressor (UnCTR)— fish had no choice to escape from US; and (3) Loss of stressor control (CTRUn)—fish were subjected during 5 conditioning sessions to the same conditions as CTR followed by 2 sessions under UnCTR conditions (Addi- tional file 1: Table S1). In the test session, all experi- mental groups were exposed to the CS in the absence of the US. We have characterized the effects of controlla- bility on the stress response at multiple levels: behavior, circulating cortisol and activation of the 4 candidate teleost homologues of tetrapod subcortical regions involved in cognitive appraisal of stressors discussed above (i.e. Dm, Dld, Dlv, Vv), using immediate early genes (IEG: egr-1, c-fos, bdnf and npas4) as markers of neural activity. In teleost fish three pallial areas and one sub-pallial area have been suggested to play a role in cognitive and affective components of behavior: (1) the medial zone of the dorsal telencephalic area (Dm), which has been shown to be involved in emotional learning [26, 27] and is the putative homologue of the mammalian pallial amygdala [28, 29]; (2) the ventral division of the lateral telencephalic area (Dlv), which has been shown to be involved in spatial and time-related learning [30–32] and is the putative homologue of the mammalian hip- pocampus [28, 29]; (3) the dorsal division of the lateral Cerqueira et al. BMC Neurosci (2021) 22:48 Cerqueira et al. BMC Neurosci (2021) 22:48 Page 3 of 12 Page 3 of 12 Time in freezing, frequency of escape events, shoal cohesion, and exploratory behavior expressed during the test session and cortisol levels and immediate early gene (IEGs) mRNA expression measured 30 min after the test session. IEGs mRNA from each of the candidate brain nuclei in seabass (i.e. Dm, Dld, Dlv and Vv) are indicated. Significant differences are expressed with (p < 0.05) Resultsf IEGs), with similar levels of IEG expression in controlled and loss of control treatments. In Vv loss of control is associated with lower egr-1 expression than exposure to an uncontrolled stressor, and the expression of bdnf is higher when exposed to a controllable stressor than when exposed to either an uncontrolled stressor or when there is a loss of control of the stressor. Finally, for Dm there were no significant differences in IEG expression between treatments. Gene expression and physiological state predict stress coping state A linear discriminant function analysis, combining IEGs expression on the target brain nuclei and cortisol levels, was used to assess if the perception of controllability of Table 1 Linear Mixed Model main effects of the behavioral, physiological and of IEGs mRNA responses expressed between experimental conditions (i.e. CTR = stressor controllability; UnCTR = stressor uncontrollability and CTRUn = loss of stressor controllability) controllability) Time in freezing, frequency of escape events, shoal cohesion, and exploratory behavior expressed during the test session and cortisol levels and immediate early gene (IEGs) mRNA expression measured 30 min after the test session. IEGs mRNA from each of the candidate brain nuclei in seabass (i.e. Dm, Dld, Dlv and Vv) are indicated. if Behaviors df Time freezing Escape events Shoal cohesion Exploratory behavior Cortisol F p F p F p F p F p 2.15 17.55 < 0.001 2.88 0.087 3.74 0.048 13.9 < 0.001 45.1 < 0.001 Seabass nuclei Dm Dld Dlv Vv IEGs F p F p F p F p egr-1 2.4 1.07 0.425 1.66 0.297 2.53 0.194 2.66 0.184 c-fos 2.4 0.93 0.465 2.00 0.250 4.653 0.090 0.157 0.859 Bdnf 2.4 1.41 0.344 2.56 0.192 3.74 0.121 6.03 0.062 npas4 2.4 0.93 0.464 2.16 0.230 6.92 0.050* 0.43 0.675 Time in freezing, frequency of escape events, shoal cohesion, and exploratory behavior expressed during the test session and cortisol levels and immediate early gene (IEGs) mRNA expression measured 30 min after the test session. IEGs mRNA from each of the candidate brain nuclei in seabass (i.e. Dm, Dld, Dlv and Vv) are indicated. Significant differences are expressed with * (p < 0.05) Cerqueira et al. BMC Neurosci (2021) 22:48 Cerqueira et al. BMC Neurosci Page 4 of 12 the same stimuli elicits different internal states. Two discriminant functions were significantly loaded that explained 84% and 16% of variation (function 1: Wilk’s lambda = 0.171, chi-square = 18.57, p = 0.001; function 2: Wilk’s lambda = 0.651, chi-square = 4.50, p = 0.034; Fig. 4). Function 1 was significantly loaded by bdnf expression in Dld (0.691), whereas function 2 was sig- nificantly loaded by cortisol (0.786). These discriminant Fig. Gene expression and physiological state predict stress coping state 1 Behavioral and physiological responses of seabass expressed by fish during the test session (experimental treatments: CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability): a Time in freezing (%), b frequency of escape attempts, c shoal cohesion, d Exploratory behavior, f plasma cortisol concentrations measured 30 min after the test session. Linear mixed models with planned comparisons are indicated; * p < .05; p < .01; * p < .001; ns—non-significant. All descriptive statistics are mean ± SEM Fig. 1 Behavioral and physiological responses of seabass expressed by fish during the test session (experimental treatments: CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability): a Time in freezing (%), b frequency of escape attempts, c shoal cohesion, d Exploratory behavior, f plasma cortisol concentrations measured 30 min after the test session. Linear mixed models with planned comparisons are indicated; * p < .05; p < .01; * p < .001; ns—non-significant. All descriptive statistics are mean ± SEM Fig. 1 Behavioral and physiological responses of seabass expressed by fish during the test session (experimental treatments: CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability): a Time in freezing (%), b frequency of escape attempts, c shoal cohesion, d Exploratory behavior, f plasma cortisol concentrations measured 30 min after the test session. Linear mixed models with planned comparisons are indicated; * p < .05; p < .01; * p < .001; ns—non-significant. All descriptive statistics are mean ± SEM the same stimuli elicits different internal states. Two discriminant functions were significantly loaded that explained 84% and 16% of variation (function 1: Wilk’s lambda = 0.171, chi-square = 18.57, p = 0.001; function 2: Wilk’s lambda = 0.651, chi-square = 4.50, p = 0.034; Fig. 4). Function 1 was significantly loaded by bdnf expression in Dld (0.691), whereas function 2 was sig- nificantly loaded by cortisol (0.786). These discriminant Cerqueira et al. BMC Neurosci (2021) 22:48 Page 5 of 12 Fig. 2 Expression (mean ± SEM) of the immediate early genes egr-1, c-fos, bdnf and npas4 in the Dm, Dld, Dlv and Vv regions of Sea bass brain under different experimental conditions (CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability). Significant differences (planned comparisons) in expression levels between experimental conditions (i.e. CTR vs. UnCTR; CTR vs. Gene expression and physiological state predict stress coping state CTRUn and UnCTR vs. CTRUn) are indicated by asterisks: * p < .05; p < .01; * p < .001; ns—non-significant Fig. 2 Expression (mean ± SEM) of the immediate early genes egr-1, c-fos, bdnf and npas4 in the Dm, Dld, Dlv and Vv regions of Sea bass brain under different experimental conditions (CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability). Significant differences (planned comparisons) in expression levels between experimental conditions (i.e. CTR vs. UnCTR; CTR vs. CTRUn and UnCTR vs. CTRUn) are indicated by asterisks: * p < .05; p < .01; * p < .001; ns—non-significant Fig. 2 Expression (mean ± SEM) of the immediate early genes egr-1, c-fos, bdnf and npas4 in the Dm, Dld, Dlv and Vv regions of Sea bass brain under different experimental conditions (CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability). Significant differences (planned comparisons) in expression levels between experimental conditions (i.e. CTR vs. UnCTR; CTR vs. CTRUn and UnCTR vs. CTRUn) are indicated by asterisks: * p < .05; p < .01; * p < .001; ns—non-significant defensive behaviors, namely freezing, escape attempts and shoal cohesion, and increases exploratory behavior. Furthermore, freezing, which is a universal fear response, characterized by a state of attentive immobility [45, 46], is higher in the group who lost control of the stressor than in the uncontrolled stressor treatment. These results are consistent with the finding in rodents that freezing is a hallmark of learned helplessness, lasting longer in situa- tions where no escape from a stressor is available than when escaping is possible [1, 47]. Thus, it is plausible that sea bass have learned that outcomes in the loss of control treatment, came independent of their response, and hence further increased their freezing behavior. The higher shoal cohesion, observed in the uncontrollable and lack of control of stressor treatments, is in agree- ment with the use of shoal cohesion as a measure of anxi- ety and fear in fish [48, 49]. Thus, stressor controllability seems to reduce anxiety/fear in zebrafish facing an aver- sive stimulus. functions allowed the correct classification of 87.5% of either CTR or UnCTR and 75% of CTRUn (overall 83.3%). These 2 functions allowed the correct classifi- cation of all individuals according to their stress coping state. Discussion 3 Neurogenomic states of seabass, as described by correlation (r) matrices of immediate early genes expression in the different brain nuclei (Dm, medial zone of the dorsal telencephalic area; Dl, lateral zone of the dorsal telencephalic area; Dld, dorsal lateral zone of the dorsal telencephalic area; Dlv, ventral lateral zone of the dorsal telencephalic area; Vv, ventral nucleus of the ventral telencephalic area) for each experimental treatment (CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability); Color scheme represents r-values from − 1 (blue) to 1 (red); Asterisks indicate significant correlations after p-value adjustment for multiple correlations: p < 0.05; p < 0.01; *p < 0.001; different capital letters indicate significantly different co-expression patterns among experimental treatments, and different small letters indicate significantly different co-expression patterns among brain nuclei, using the QAP correlation test Table 2 Pearson correlations between controllability-driven behavioral, physiological and of IEGs mRNA responses expressed between experimental conditions (i.e. CTR = stressor controllability; UnCTR = stressor uncontrollability and CTRUn = loss of stressor controllability) Table 2 Pearson correlations between controllability-driven behavioral, physiological and of IEGs mRNA responses expressed between experimental conditions (i.e. CTR = stressor controllability; UnCTR = stressor uncontrollability and CTRUn = loss of stressor controllability) Table 2 Pearson correlations between controllability-driven behavioral, physiological and of IEGs mRNA responses expressed between experimental conditions (i.e. CTR = stressor controllability; UnCTR = stressor uncontrollability and CTRUn = loss of stressor controllability) Time in freezing, frequency of escape events, and exploratory behavior expressed during the test session and cortisol levels and immediate early gene (IEGs) mRNA expression measured 30 min after the test session. IEGs mRNA from each of the candidate brain nuclei in seabass that presented significant correlations (i.e. Dlv and Vv) are indicated. Discussion In this study, we have shown that stressor controllabil- ity modulates the stress response of sea bass, as meas- ured by behavioral, physiological and neural indicators. It is important to highlight that a conditioning protocol was used in this study and that in the test phase fish were responding to the CS alone and not to the stressor itself. Hence, the results presented herein confirm the findings of previous studies [43, 44], that during the test phase of the experiment sea bass have the ability to recall mem- ories of previous aversive events that they have been exposed to in the training period. Given that the training period allowed individuals to assess their ability to con- trol the stressor, their response in the test phase should reflect their stress-coping ability given the available infor- mation on stressor controllability. At the physiological level, the response of circulating cortisol levels to the stressor was higher in the uncon- trollable than in the controllable treatment, and the lack of control elicited even higher levels. Cortisol release is known to be involved in the modulation of arousal, At the behavioral level, the perception of control over the stressor significantly reduces the expression of Cerqueira et al. BMC Neurosci (2021) 22:48 Page 6 of 12 Fig. 3 Neurogenomic states of seabass, as described by correlation (r) matrices of immediate early genes expression in the different brain nuclei (Dm, medial zone of the dorsal telencephalic area; Dl, lateral zone of the dorsal telencephalic area; Dld, dorsal lateral zone of the dorsal telencephalic area; Dlv, ventral lateral zone of the dorsal telencephalic area; Vv, ventral nucleus of the ventral telencephalic area) for each experimental treatment (CTR = stressor controllability; UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability); Color scheme represents r-values from − 1 (blue) to 1 (red); Asterisks indicate significant correlations after p-value adjustment for multiple correlations: p < 0.05; p < 0.01; p < 0.001; different capital letters indicate significantly different co-expression patterns among experimental treatments, and different small letters indicate significantly different co-expression patterns among brain nuclei, using the QAP correlation test Fig. Discussion Significant differences are expressed with (p < 0.05) Behaviours Freezing behaviour Escape Attempts Exploratory Behaviour Cortisol Brain nuclei genes n Rp p n Rp p n Rp p n Rp p DLV egr1 24 0.483 0.017 24 − 0.465 0.022 23 0.444 0.034 cfos 24 0.533 0.007 24 − 0.7 < 0.001 23 0.526 0.01 bdnf 24 0.466 0.022 24 − 0.624 0.001 23 0.537 0.008 npas4 24 0.492 0.015 24 − 0.498 0.013 23 0.535 0.008 VV Bdnf 24 − 0.594 0.003 Cortisol 69 0.473 < 0.001 69 0.426 < 0.001 Time in freezing, frequency of escape events, and exploratory behavior expressed during the test session and cortisol levels and immediate early gene (IEGs) mRNA expression measured 30 min after the test session. IEGs mRNA from each of the candidate brain nuclei in seabass that presented significant correlations (i.e. Dlv and Vv) are indicated. Significant differences are expressed with * (p < 0.05) Cerqueira et al. BMC Neurosci (2021) 22:48 Page 7 of 12 Page 7 of 12 Fig. 4 Linear discriminant analysis of cortisol and egr-1, c-fos, bdnf and npas4 expressed in the candidate brain nuclei from seabass. The significant functions 1 and 2 highlight the three coping responses of fish under three experimental conditions: CTR = stressor controllability (circles); UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability (triangles). Discriminant scores for each individual are plotted and stars represent the centroid of each classified group the perception of control, the lack of control and the loss of control of a stressor, are each represented at the brain level by a specific pattern of functional connectiv- ity (i.e. pattern of co-activation) between these four brain regions. Finally, we have used linear discriminant func- tion analyses to check if the measured behavioral, physi- ological and neuromolecular variables could efficiently discriminate the three experimental treatments. Two dis- criminant function correctly classified 83.3% of individu- als according to the stressor coping state that they have been exposed to. Thus, physiological and neurogenomic state may well discriminate between three distinct stress coping states. Our data can be interpreted as evidence for the occurrence of cognitive abilities in fish corresponding to different appraisal of their environmental stressors. f Moreover, when brain regions are analyzed individu- ally, they show different patterns. Discussion y yf p There are no differences in the expression of any of the measured genes in Dm, the putative teleost homologue of the tetrapod pallial amygdala [28, 29], across the three experimental treatments, suggesting that this area is not involved in the processing of stress controllability. This is a surprising result, given the established role of Dm in emotional learning in teleost fish [27, 45] and the previ- ous results in sea bass showing a differential expression of egr1 in response to stressor predictability [41]. It is thus, possible that despite the fact that both predictability and controllability are cognitive appraisal variables that mod- ulate the stress response they might be processed by dif- ferent neural networks. In fact, while the former depends on stored information in memory about relations between stimuli (i.e. stimulus–stimulus learning or clas- sical conditioning), the later depends on stored informa- tion about relations between responses and stimuli (i.e. stimulus–response learning or instrumental condition- ing) [52]. Therefore, the effect of predictability is based on perceived contingencies between stimuli (i.e. stimulus expectancies), hence relying on the perceived probability of occurrence of the anticipated event; whereas control- lability is based on contingencies between stimulus and response (i.e. response expectancies), hence relying on the perceived probability of response outcomes [33, 52]. Thus, whereas stimulus-stimulus contingencies seem to be associated with the differential activation of Dm, as measured by egr1 expression, stimulus–response contin- gencies do not seem to require differential activation of this region. Fig. 4 Linear discriminant analysis of cortisol and egr-1, c-fos, bdnf and npas4 expressed in the candidate brain nuclei from seabass. The significant functions 1 and 2 highlight the three coping responses of fish under three experimental conditions: CTR = stressor controllability (circles); UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability (triangles). Discriminant scores for each individual are plotted and stars represent the centroid of each classified group Fig. 4 Linear discriminant analysis of cortisol and egr-1, c-fos, bdnf and npas4 expressed in the candidate brain nuclei from seabass. The significant functions 1 and 2 highlight the three coping responses of fish under three experimental conditions: CTR = stressor controllability (circles); UnCTR = stressor uncontrollability; CTRUn = loss of stressor controllability (triangles). Discussion Discriminant scores for each individual are plotted and stars represent the centroid of each classified group vigilance, attention and memory formation [50], facilitat- ing the encoding of emotion-related memory in fish [33]. Indeed, individuals in the controlled stressor treatment displayed the lowest cortisol concentrations, suggesting that the perception of control, rather than lack or loss of control over a stressful event, is effectively appraised as less harmful. Thus, and in accordance to previous research [37, 42, 51], psychological stress increases the synthesis and release of cortisol, in association with fear and anxiety states as suggested by the behavioral data. In fact, the cortisol results are in agreement with the behav- ioral data supporting a buffering role of controllability in the physiological stress response, and the occurrence of a helplessness state in lack or loss of control situations. Finally, it is worth noting that the cortisol levels meas- ured in this study are within those previously reported for this species, even with different protocols [33, 43, 44]. At the neuromolecular level, the pattern of co-expres- sion of the immediate early genes (egr-1, c-fos, and npas4) and the neural plasticity gene bdnf, used in this study to capture the neurogenomic state of target brain regions, was specific for each experimental treatment in each of the sampled brain regions (Dm, Dld, Dlv, Vv). Taking the expression of the immediate early genes as a proxy for neuronal activity, these results indicate that vigilance, attention and memory formation [50], facilitat- ing the encoding of emotion-related memory in fish [33]. Indeed, individuals in the controlled stressor treatment displayed the lowest cortisol concentrations, suggesting that the perception of control, rather than lack or loss of control over a stressful event, is effectively appraised as less harmful. Thus, and in accordance to previous research [37, 42, 51], psychological stress increases the synthesis and release of cortisol, in association with fear and anxiety states as suggested by the behavioral data. In fact, the cortisol results are in agreement with the behav- ioral data supporting a buffering role of controllability in the physiological stress response, and the occurrence of a helplessness state in lack or loss of control situations. Finally, it is worth noting that the cortisol levels meas- ured in this study are within those previously reported for this species, even with different protocols [33, 43, 44]. Discussion Although this region has been mainly associated with spatial learning, increasing evidence has linked its activity with time-related learning [30, 31], which might be the basis for its involvement in the here reported vio- lation of response expectancies. The occurrence of stressor controllability in fish that lack the cortical areas (e.g. mPFC) known to regulate the stress response as a function of perceived stressor con- trol, through a top-down inhibitory control over sub- cortical stress-responsive structures [15], illustrates how similar cognitive abilities (i.e. learning response contin- gencies) may be implemented by different neural circuits that have evolved separately in divergent vertebrate taxa (i.e. teleost vs. mammals). In recent years, research on the neuronal basis of the well develop cognitive capacities of birds, which also lack a layered cerebral cortex, reached similar conclusions. Irrespective of layered (in mammals) or nuclear (in birds) telencephalic neuronal architectures, cognitive ability seems to rely on associative capacities implemented by interneurons placed between sensory input and motor output neurons, which integrate sensory input and information stored in memory [23–25]. Thus, the view that fish are not able of complex cognitive abili- ties due to their relatively smaller brains that lack a cor- tical area homologous to that of mammals (e.g. [57, 58], which has direct implications for fish sentience and wel- fare, should be revised. p p The Vv, which is the putative teleost homologue of the lateral septum and of the nucleus accumbens [34, 35], is the only brain region where the expression of the four genes was not similar, and only two genes showed differ- ent levels of expression between the experimental treat- ments, egr-1 and bdnf. The expression of egr-1 was higher in the uncontrollable stressor treatment than either in the controlled or loss of control treatment, which did not differ between them. Hence, this pattern of expres- sion seems to be associated with the training for per- ceiving the stressor as uncontrollable (i.e. CS signals a lack of a coping response). It is important to note that c-fos and npas1 have similar patterns, but the compari- sons between treatments were not significant. Interest- ingly, these patterns of response are exactly the inverse of the patterns of expression of all four genes in the Dld reported above, which seem to be associated with the response expectancies (i.e. CS signals a coping escape response) in the controlled and loss of stressor control treatments. Discussion On the other hand, bdnf presents a specific pattern of expression in the Vv, with higher levels in con- trollable than in either uncontrollable or loss of stressor Conclusions Here we present behavioral, physiological and neuromo- lecular evidence for the modulation of the stress response by the perception of stressor controllability in fish, sug- gesting a deep evolutionary origin of this cognitive ability in vertebrate evolution, which is implemented by differ- ent neural substrates in divergent vertebrate taxa. These results also have implications for fish welfare opening the way for using psychological factors in the management of farmed fish, namely in handling and husbandry practices. Discussion At the neuromolecular level, the pattern of co-expres- sion of the immediate early genes (egr-1, c-fos, and npas4) and the neural plasticity gene bdnf, used in this study to capture the neurogenomic state of target brain regions, was specific for each experimental treatment in each of the sampled brain regions (Dm, Dld, Dlv, Vv). Taking the expression of the immediate early genes as a proxy for neuronal activity, these results indicate that In contrast, the four measured genes show similar dif- ferential patterns of expression in both in Dld and Dlv. In Dld, the teleost putative homologue of the sensory cor- tex [28, 29], the expression of all four genes was higher in the loss of control treatment than in the uncontrolled stressor treatment. Interestingly, the controlled stressor Cerqueira et al. BMC Neurosci (2021) 22:48 Cerqueira et al. BMC Neurosci (2021) 22:48 Page 8 of 12 treatment elicited a pattern of gene expression for the four genes that is not significantly different from any of the other two treatments, but is much closer to the loss of control treatment. This suggests that the pattern of acti- vation of this region reflects the commonalities between the controlled stressor and the loss of control treatment, which reside on the training during the first 5 train- ing sessions in which similar contingencies between the CS stimulus and the escape response were established. Therefore, the activation of this area seems to be related to response expectancies. The fact that Dld is a sensory higher processing area that receives and integrates mas- sive visual inputs [28, 29, 53], is congruent with the fact that the training used a visual CS. Also, the involvement of Dld in short-term memory and in the performance of activity under stressful situations in goldfish, suggests the potential involvement of this area in memory processes in fish [54]. In the Dlv, which is the putative teleost hom- ologue of the hippocampus [28, 29], the expression of all four genes was significantly higher in the loss of control treatment than either in the controlled or uncontrolled stressor treatments, which have similar levels of expres- sion. Therefore, the activation of this area seems to reflect the violation of expectancies in the test phase experi- enced by the individuals of the loss of control group, which may act as an error signal in the response expec- tancies. Discussion Although this region has been mainly associated with spatial learning, increasing evidence has linked its activity with time-related learning [30, 31], which might be the basis for its involvement in the here reported vio- lation of response expectancies. control treatments. This result is in line with the inhibi- tion of bdnf mRNA expression in whole telencephalon samples in Atlantic salmon, in response to the omission of an expected reward [55]. Thus, the higher expression of bdnf in the Vv seems to be related to perceived stress coping which is probably associated with lower anxiety and fear. This interpretation is supported by the estab- lished role of bdnf in the nucleus accumbens on reward and motivational state [56].hi treatment elicited a pattern of gene expression for the four genes that is not significantly different from any of the other two treatments, but is much closer to the loss of control treatment. This suggests that the pattern of acti- vation of this region reflects the commonalities between the controlled stressor and the loss of control treatment, which reside on the training during the first 5 train- ing sessions in which similar contingencies between the CS stimulus and the escape response were established. Therefore, the activation of this area seems to be related to response expectancies. The fact that Dld is a sensory higher processing area that receives and integrates mas- sive visual inputs [28, 29, 53], is congruent with the fact that the training used a visual CS. Also, the involvement of Dld in short-term memory and in the performance of activity under stressful situations in goldfish, suggests the potential involvement of this area in memory processes in fish [54]. In the Dlv, which is the putative teleost hom- ologue of the hippocampus [28, 29], the expression of all four genes was significantly higher in the loss of control treatment than either in the controlled or uncontrolled stressor treatments, which have similar levels of expres- sion. Therefore, the activation of this area seems to reflect the violation of expectancies in the test phase experi- enced by the individuals of the loss of control group, which may act as an error signal in the response expec- tancies. Blood collection and Hormone Analysishi Thirty min after the test session, fish were caught (n = 72) and euthanized with an overdose of 2-phenoxyethanol (1 ‰, Sigma-Aldrich). Blood was immediately collected from the caudal vein and centrifuged at 2000g for 25 min. Plasma was frozen in dry ice and stored at − 80 °C until further processing. Plasma cortisol were then determined using a commercial ELISA kit RE52061 (IBL Hamburg, Germany), with a sensitivity of 2.5 ng ml−1, and intra and inter-assay coefficients of variation (CV) of 2.9 and 3.5%, respectively. Samples were identified according to the videos labelling. y During the training sessions fish were trained to learn an association between a light (CS; on for 2 min) and a subsequent 5 min confinement (unconditioned stimulus: US). The light (12 V, 10 W) was hung in the top of a lat- eral wall of the aquarium, on the opposite side of the con- finement net. Confinement consisted in moving the net into the light wall direction until reach 15% of the aquar- ium volume. Fish were tested under different conditions: i) Controllable situation (CTR)—fish had the choice to escape from US (confinement) by a door of 10 cm2; ii) Uncontrollable situation (UnCTR)—fish had no choice to escape from US; iii) Loss of Controllability (CTRUn)— fish were subjected during 5 conditioning sessions to the same conditions as CTR followed by 2 sessions under UnCTR conditions (Additional file 1: Table S2). In all cases, the event was created by signalling (CS) the aver- sive stimulus. In the test session all experimental groups were exposed to the CS in the absence of the US. Each aquarium was labelled with a different number to permit further blind analysis of the different samples taken. Experimental fish Seabass were obtained from the experimental research station of IFREMER in Palavas-les-Flots (France) and housed in fibre glass tanks (500 L) during 8 months at Ramalhete Research Station from Centre of Marine Sci- ences (Faro, Portugal), until the start of the experiment in May of 2013. At the start of the experiments the body mass of the fish was 47.12 ± 6.80 g (mean ± SEM). Cerqueira et al. BMC Neurosci (2021) 22:48 Cerqueira et al. BMC Neurosci (2021) 22:48 Page 9 of 12 A/tmaximum where A is the arithmetic mean of the time fish spent in each one of the areas of the tank, and t the maximum time found for any of the areas tested. When exploratory behavior is high, this ratio should be close to 1, while it should be close to 0 when exploratory behavior is low. Brain microdissection and immediate early genes (IEGs) expression Eight individuals from each experimental treatment (n = 24) were randomly selected for the assessment of immediate early gene mRNA expression. After sacrifice (see above) the skull, with the brain inside, was removed from the fish, embedded in Tissue-Tek®, and kept at − 80 °C until further processing. Labelling of the samples were performed as for the blood samples. Thick brain telencephalon coronal Sects. (150 µm) were obtained using a cryostat (Leica, CM 3050S). Brain nuclei of inter- est (Dm, Dld, Dlv, and Vv), identified according to the available brain atlas [59, 60], were microdissected (see Additional file 1: Fig. S2 and Additional file 2 for detailed description). Tissue was collected directly into lysis buffer from Qiagen Lipid Tissue Mini Kit (#74804; Valen- cia, CA), total RNA extracted from the samples, reverse transcribed to cDNA (BioRad iScript cDNA Synthesis Kit; Valencia, CA), and used as a template for quantita- tive polymerase chain reactions (qPCR) of the following IEGs: egr-1, c-fos, bdnf and npas4. The geometric mean of Experimental procedures Netherlands). The conditioned response was assessed by: time in freezing (total time without any movement), escape behavior (i.e. fish swimming strongly, going close to the tank walls or moving in a way consistent with escape attempts), shoal cohesion (1—low cohesion i.e., fish were spread in the tank; 2—medium cohesion i.e., 2 fish were together; 3—high cohesion i.e., three or more fish together) and exploratory behavior (measured by the time fish spent in different areas of the tank: 1—light side area; 2—centre of the tank; 3—confinement net area) summarized by the formula: Seventy-two individuals, randomly selected from the stock tank were measured, weighted and tagged with a floy tag (Floy Tag Manufacturing Inc, Seattle, USA) with a multicolour pearl attached behind the dorsal fin. Groups of 4 fish were randomly distributed by twelve experimen- tal aquaria of approximately 80 l (70 × 40 × 30 cm depth) in an open water under standard housing conditions until the start of the experiments (temperature 18 ± 3 °C, salinity 35 ± 2 ‰, dissolved oxygen above 75%, and a 12 L: 12 D photoperiod). The experiment was developed in three runs, using 24 fish/run. All aquaria walls were covered with opaque partitions in order to prevent con- tact between focal animals and the experimenter. Each aquarium was provided with a confinement net attached to a white rigid structure, with the same size as the lat- eral walls of the tank (Additional file 1: Fig. S1). Each run lasted 12 days, including 8 days of acclimation to the experimental tanks and the last 4 to the experimental period (2 training sessions per day at 10:00 and 15:00 h; the last session of the fourth day was the test session). During this period fish were hand-fed twice a day at 08:00 h and 18:00 h at 3% of biomass day−1. (1) A/tmaximum (1) Acknowledgements h h f The authors are grateful to João Reis, Miguel Viegas and Rui Gonçalves for technical assistance. Availability of data and materials All data analyzed during this study are included in this published article and its supplementary information files. Behavioral analysis Behavior was video recorded (top view) during the CS of the test session. Video recordings were labelled with a number by a researcher different from the one involved in the different analysis to permit a blind analysis of the behaviour. Behavioral responses of 72 fish to the CS was analysed with computerized multi-event recorder software (Observer XT®, from Noldus, Wageningen, Cerqueira et al. BMC Neurosci (2021) 22:48 Page 10 of 12 Cerqueira et al. BMC Neurosci (2021) 22:48 experimental condition, and between experimental conditions within each brain area. The null hypothesis of the QAP test is that when p > 0.05 there is no associa- tion between matrices, hence a non-significant p-value indicates that the correlation matrices are different. LMM, LDA, neurogenomic states and QAP correla- tions were performed using R® (R Development Core Team). Statistical significance was taken at p < 0.05. the expression of two previously established housekeep- ing genes, eef1a and 18S, was used as an internal control (see additional file for qPCR conditions and Additional file 1: Table S3 for primer sequences). Authors’ contributions RFO, SMa, SR and MC designed research; MC, SMi, MFC and ASF performed research; MC, TS and GAO analyzed data; RFO, CCO and MC wrote the paper. All authors read and approved the final manuscript. Funding This research project has been supported by the European Commis- sion under the 7th Framework Programme FP7–KBBE-2010-4 Contract n°: 265957 Copewell. Marco Cerqueira and Sandie Millot were supported by FCT-Fundação para a Ciência e a Tecnologia, Portugal, fellowships (SFRH/ BD/80029/2011and SFRH/BPD/72952/2010, respectively). The funding bodies had no role in the design of the study and collection, analysis, and interpreta- tion of data nor in writing the manuscript. All LMM were estimated using the restricted maxi- mum likelihood method. A priori planned comparisons were used to test for specific differences between experi- mental conditions, namely: CTR vs. UnCTR; CTR vs. CTRUn; UnCTR vs. CTRUn. Pearson correlations were used to depict the association between behavioral vari- ables, between behavior and cortisol, and between those and gene expression. Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s12868-021-00653-0. Additional file 1. Supplementary figures. Additional file 2. Raw data. Additional file 2. Raw data. A linear mixed model (LMM) was used to assess the effect of each experimental condition (i.e. CTR vs. UnCTR; CTR vs. CTRUn; UnCTR vs. CTRUn) on the behavioral variables assessed in the test session, on cor- tisol levels and on IEGs mRNA expression (egr-1, c-fos, bdnf and npas4) in each brain region of seabass (Dm, Dld, Dlv and Vv). Given that we have used more than one individual from the same tank in each condition, we accounted for sampling dependence by adding a random effect for the "tank" factor in each LMM. In general, we did not find an effect of the "tank" variable on the meas- ured responses. Statistical analysis Label scheme was deciphered before the statistical analy- sis to allocate each number to the respective treatment group. All samples were included in the analysis. The assumptions of normality and homoscedasticity were confirmed by analysis of the residuals whereas homoge- neity of variance was checked by Levene’s test. Descrip- tive statistics are expressed as mean ± standard error of the mean (SEM) and log or arcsine transformation was used to achieve homogeneity when required. Consent for publicati Not applicable. Not applicable. Ethics approval and consent to participate Separate stepwise linear discriminant analyses (LDA) was used to define which brain nuclei state (i.e. imme- diate early genes expression in different brain nuclei) and cortisol expression better predict the coping responses to aversive events. The F test statistic was used as a measure of the contribution of each variable (cortisol concentration and IEGs expression in each brain region) to the discriminant functions. An F-value above 3.84 was used as the selection criteria for predic- tors to enter the model and predictors were removed when the F-value dropped below 2.71 (e.g. Maruska et al. [61]). Heatmaps of Pearson correlations matri- ces, with adjusted p-values [62] were used to assess the neurogenomic states of fish, where the quadratic assignment procedure (QAP) correlation test with 5000 permutations [63] test the differences in gene co- expression patterns between brain areas within each Experimental procedures were conducted in accordance with European guidelines on the protection of animals used for scientific purposes (Directive 2010/63/EU) and the Portuguese legislation for the use of laboratory animals, under a “Group- 1” license (permit number 0420/000/000-n.99–09/11/2009) from Direção Geral de Alimentação e Veterinária (DGAV, Portugal), which is the competent Portuguese authority for the protection of animals used in experimental research. References Stress in Atlantic salmon: response to unpredictable chronic stress. J Exp Biol. 2015;218:2538. 11. Hartley CA, Gorun A, Reddan MC, Ramirez F, Phelps EA. Stressor control- lability modulates fear extinction in humans. Neurobiol Learn Mem. 2014;113:149–56. 40. Vindas MA, et al. Frustrative reward omission increases aggressive behav- ior of inferior fighters. Proc R Soc B. 2014;281:20140300. 12. Henderson RK, Snyder HR, Gupta T, Banich MT. When does stress help or harm? The effects of stress controllability and subjective stress response on stroop performance. Front Psychol. 2012;3:179–179. 41. Cerqueira M, et al. Cognitive appraisal in fish: stressor predictability modulates the physiological and neurobehavioral stress response in sea bass. Proc R Soc B. 2020;287:20192922. 13. Markus R, Panhuysen G, Tuiten A, Koppeschaar H. Effects of food on cor- tisol and mood in vulnerable subjects under controllable and uncontrol- lable stress. Physiol Behav. 2000;70:333–42. 42. Carpenter RE, Summers CH. Learning strategies during fear conditioning Neurobiol Learn Mem. 2009;91:415–23. 43. Millot S, et al. Use of conditioned place preference/avoidance tests to assess affective states in fish. Appl Anim Beh Sci. 2014;154:104–11. y 14. Müller MJ. Helplessness and perceived pain intensity: relations to cortisol concentrations after electrocutaneous stimulation in healthy young men. BioPsychoSocial medicine. 2011;5:8. 44. Millot S, et al. Behavioral stress responses predict environmental perception in european sea bass dicentrarchus labrax. PLoS ONE. 2014;9:e108800. y 15. Maier SF. Behavioral control blunts reactions to contemporaneous and future adverse events: medial prefrontal cortex plasticity and a corticos- triatal network. Neurobiol Stress. 2015;1:12–22. 45. Lang PJ, Davis M. Emotion, motivation, and the brain: reflex foundations in animal and human research. Prog Brain Res. 2006;156:3–29. 16. Kern S, et al. Glucose metabolic changes in the prefrontal cortex are associated with HPA axis response to a psychosocial stressor. Psychoneu- roendocrinology. 2008;33:517–29. 46. Lojowska M, Gladwin TE, Hermans EJ, Roelofs K. Freezing promotes perception of coarse visual features. J Exp Psychol Gen. 2015;144:1080–8 47. Blanchard DC, Griebel G, Pobbe R, Blanchard RJ. Risk assessment as an evolved threat detection and analysis process. Neurosci Biobehav Rev. 2011;35:991–8. 17. Sugaya N, et al. Adrenal hormone response and psychophysiological correlates under psychosocial stress in individuals with irritable bowel syndrome. Int J Psychophysiol. 2012;84:39–44. 48. Gerlai R. Zebrafish antipredatory responses: a future for translational research? Behav Brain Res. 2010;207:223–31. 18. Maier SF, Ryan SM, Barksdale CM, Kalin NH. Stressor controllability and the pituitary-adrenal system. Behav Neurosci. 1986;100:669–74. 49. Stewart A, et al. Received: 27 March 2021 Accepted: 28 July 2021 28. Demski LS. The pallium and mind/behavior relationships in teleost fishes. Brain Behav Evol. 2013;82:31–44. 29. Ganz J, et al. Subdivisions of the adult zebrafish pallium based on molecular marker analysis. F1000 Research. 2014;3:308. i molecular marker analysis. F1000 Research. 2014;3:308 30. Ocaña FM, Uceda S, Arias JL, Salas C, Rodríguez F. Dynamics of goldfish subregional hippocampal pallium activity throughout spatial memory formation. Brain Behav Evol. 2017;90:154–70. References Maier SF, Seligman MEP. Learned helplessness at fifty: Insights from 1. Maier SF, Seligman MEP. Learned helplessness at fifty: Insights from 1. Maier SF, Seligman MEP. Learned helplessness at fifty: Insights from neuroscience. Psychol Rev. 2016;123:349–67. 31. Rodrı́guez F, et al. Conservation of spatial memory function in the pallial forebrain of reptiles and ray-finned fishes. J Neurosci. 2002;22: 2894. 2. Brady JV. Ulcers in executive monkeys. Sci Am. 1958;199:95–8. 2. Brady JV. Ulcers in executive monkeys. Sci Am. 1958;199:95–8 32. Uceda S, et al. Spatial learning-related changes in metabolic brain activity contribute to the delimitation of the hippocampal pallium in goldfish. Behav Brain Res. 2015;292:403–8. 3. Seligman ME, Maier SF. Failure to escape traumatic shock. J Exp Psychol. 1967;74:1–9. 4. Maier SF, Seligman ME. Learned helplessness: theory and evidence. J Exp Psychol Gen. 1976;105:3–46. 33. Cerqueira M, et al. Cognitive appraisal of environmental stimuli induces emotion-like states in fish. Sci Rep. 2017;7:13181. 33. Cerqueira M, et al. Cognitive appraisal of environm emotion-like states in fish. Sci Rep. 2017;7:13181. 5. Weiss JM. Effects of coping responses on stress. J Comp Physiol Psychol. 1968;65:251–60. 34. Ganz J, et al. Subdivisions of the adult zebrafish subpallium by molecular marker analysis. J Comp Neurol. 2012;520:633–55. 6. Weiss JM. Somatic effects of predictable and unpredictable shock. Psy- chosom Med. 1970;32:397–408. 35. Goodson JL, Kingsbury MA. What’s in a name? Considerations of homolo- gies and nomenclature for vertebrate social behavior networks. Horm Behav. 2013;64:103–12. 7. Weiss JM. Effects of coping behavior in different warning signal condi- tions on stress pathology in rats. J Comp Physiol Psychol. 1971;77:1–13. 36. Galhardo L, Oliveira RF. Psychological Stress and Welfare in Fish. ARBS Ann Rev Biomed Sci. 2009;11:1–20. 8. Hiroto DS, Seligman ME. Generality of learned helplessness in man. J Pers Soc Psychol. 1975;31:311–27. 37. Galhardo L, Vital J, Oliveira RF. The role of predictability in the stress response of a cichlid fish. Physiol Behav. 2011;102:367–72. 9. Abramson LY, Seligman ME, Teasdale JD. Learned helplessness in humans: Critique and reformulation. J Abnorm Psychol. 1978;87:49–74. response of a cichlid fish. Physiol Behav. 2011;102:367–72. 38. Madaro A, et al. Effect of predictability on the stress response to chasing in Atlantic salmon (Salmo salar L.) parr. Physiol Beh. 2016;153:1–6. 10. Alloy LB, Peterson C, Abramson LY, Seligman ME. Attributional style and the generality of learned helplessness. J Pers Soc Psychol. 1984;46:681–7. 39. Madaro A, et al. Author details 1 Centro de Ciências Do Mar (CCMAR), Universidade Do Algarve, 8005‑139 Faro, Portugal. 2 Fish Welfare Initiative, Normal, IL 61761, USA. 3 Laboratoire Ressources Halieutiques, Ifremer, 17137 L’Houmeau, France. 4 SPAROS, 8700‑221 Olhão, Portugal. 5 ISPA—Instituto Universitário, 1149‑041 Lisbon, Por- tugal. 6 Instituto Gulbenkian de Ciência, 2780‑156 Oeiras, Portugal. 7 Institute of Aquaculture, University of Stirling, Stirling FK9 4LA, UK. 8 Champalimaud Research, 1400‑038 Lisbon, Portugal. Page 11 of 12 Page 11 of 12 Cerqueira et al. BMC Neurosci (2021) 22:48 Cerqueira et al. BMC Neurosci (2021) 22:48 Received: 27 March 2021 Accepted: 28 July 2021 References BMC Neurosci (2021) 22:48 Cerqueira et al. BMC Neurosci (2021) 22:48 Page 12 of 12 Page 12 of 12 Page 12 of 12 62. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc. 1995;57:289–300. 63. Borgatti SP, Everett MG, Johnson JC. Analyzing social networks. New York: SAGE Publications Limited; 2013. 59. Cerda-Reverter JM, Zanuy S, Munoz-Cueto JA. Cytoarchitectonic study of the brain of a perciform species, the sea bass (Dicentrarchus labrax) I The telencephalon. J Morphol. 2001;247:217–28. p pp p g ; 63. Borgatti SP, Everett MG, Johnson JC. Analyzing social networks. New York: SAGE Publications Limited; 2013. 60. Munoz-Cueto JA, Sarasquete C, Zohar AH, Kah O. An atlas of the brain of the gilthead seabream (Sparus aurata). College Park: Maryland Sea Grant; 2001. Cerqueira et al. BMC Neurosci (2021) 22:48 62. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc. 1995;57:289–300. 63. Borgatti SP, Everett MG, Johnson JC. Analyzing social networks. New York: SAGE Publications Limited; 2013. References Modeling anxiety using adult zebrafish: a conceptual review. Neuropharmacology. 2012;62:135–43. 19. Helmreich DL, et al. The effect of stressor controllability on stress-induced neuropeptide mRNA expression within the paraventricular nucleus of the hypothalamus. J Neuroendocrinol. 1999;11:121–8. 50. van Ast VA, et al. Modulatory mechanisms of cortisol effects on emotional learning and memory: Novel perspectives. Psychoneuroendocrinology. 2013;38:1874–82. 20. Sanchís-Ollé M, et al. Controllability affects endocrine response of ado- lescent male rats to stress as well as impulsivity and behavioral flexibility during adulthood. Sci Rep. 2019;9:3180. 51. Øverli Ø, et al. Behavioral and neuroendocrine correlates of displaced aggression in trout. Horm Behav. 2004;45:324–9. 21. Venkatesh B, Erdmann MV, Brenner S. Molecular synapomorphies resolve evolutionary relationships of extant jawed vertebrates. Proc Natl Acad Sci. 2001;98:11382. 52. Ursin H, Eriksen HR. The cognitive activation theory of stress. Psychoneu- roendocrinology. 2004;29:567–92. 53. Wullimann MF, Mueller T. Teleostean and mammalian forebrains contrasted: evidence from genes to behavior. J Comp Neurol. 2004;475:143–62. 22. Ito H, Yamamoto N. Non-laminar cerebral cortex in teleost fishes? Biol Lett. 2009;5:117–21. 23. Briscoe SD, Ragsdale CW. Evolution of the Chordate Telencephalon. Curr Biol. 2019;29:R647–62. 54. Vargas JP, López JC, Portavella M. What are the functions of fish brain pal- lium? Brain Res Bull. 2009;79:436–40. 24. Mehlhorn J, Hunt GR, Gray RD, Rehkämper G, Güntürkün O. Tool-making new caledonian crows have large associative brain areas. Brain Behav Evol. 2010;75:63–70. 55. Vindas MA, et al. Coping with Unpredictability: Dopaminergic and Neu- rotrophic Responses to Omission of Expected Reward in Atlantic Salmon (Salmo salar L.). PLoS ONE. 2014;9:e85543. 25. Olkowicz S, et al. Birds have primate-like numbers of neurons in the forebrain. Proc Natl Acad Sci. 2016;113:7255. 56. Koo JW, Chaudhury D, Han M-H, Nestler EJ. Role of Mesolimbic Brain- Derived Neurotrophic Factor in Depression. Biol Psychiat. 2019;86:738–48 26. Lal P, et al. Identification of a neuronal population in the telencephalon essential for fear conditioning in zebrafish. BMC Biol. 2018;16:45. 57. Rose JD. The neurobehavioral nature of fishes and the question of aware- ness and pain. Rev Fish Sci. 2002;10:1–38. 27. Portavella M, Torres B, Salas C. Avoidance response in goldfish: emotional and temporal involvement of medial and lateral telencephalic pallium. J Neurosci. 2004;24:2335. 27. Portavella M, Torres B, Salas C. Avoidance response in goldfish: emotional and temporal involvement of medial and lateral telencephalic pallium. J Neurosci. 2004;24:2335. 58. Rose JD, et al. Can fish really feel pain? Fish Fish. 2014;15:97–133. Cerqueira et al. g 63. Borgatti SP, Everett MG, Johnson JC. Analyzing social networks. New York: SAGE Publications Limited; 2013. 62. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc. 1995;57:289–300. Publisher’s Note 61. Maruska KP, Zhang A, Neboori A, Fernald RD. Social opportunity causes rapid transcriptional changes in the social behavior network of the brain in an African cichlid fish. J Neuroendocrinol. 2013;25:145–57. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from:
W2005502799.txt	https://downloads.hindawi.com/journals/jps/2014/403764.pdf	en		A General Result on the Mean Integrated Squared Error of the Hard Thresholding Wavelet Estimator under<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1"><mml:mrow><mml:mi>α</mml:mi></mml:mrow></mml:math>-Mixing Dependence	Journal of probability and statistics	2,014	cc-by	8,886	Hindawi Publishing Corporation Journal of Probability and Statistics Volume 2014, Article ID 403764, 12 pages http://dx.doi.org/10.1155/2014/403764 Research Article A General Result on the Mean Integrated Squared Error of the Hard Thresholding Wavelet Estimator under 𝛼-Mixing Dependence Christophe Chesneau Laboratoire de Mathéematiques Nicolas Oresme, Université de Caen, BP 5186, 14032 Caen Cedex, France Correspondence should be addressed to Christophe Chesneau; christophe.chesneau@gmail.com Received 10 April 2013; Accepted 22 October 2013; Published 19 January 2014 Academic Editor: Zhidong Bai Copyright © 2014 Christophe Chesneau. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. We consider the estimation of an unknown function 𝑓 for weakly dependent data (𝛼-mixing) in a general setting. Our contribution is theoretical: we prove that a hard thresholding wavelet estimator attains a sharp rate of convergence under the mean integrated squared error (MISE) over Besov balls without imposing too restrictive assumptions on the model. Applications are given for two types of inverse problems: the deconvolution density estimation and the density estimation in a GARCH-type model, both improve existing results in this dependent context. Another application concerns the regression model with random design. 1. Introduction A general nonparametric problem is adopted: we aim to estimate an unknown function 𝑓 via 𝑛 random variables 𝑉1 , . . . , 𝑉𝑛 from a strictly stationary stochastic process (𝑉𝑡 )𝑡∈Z . We suppose that (𝑉𝑡 )𝑡∈Z has a weak dependence structure; the 𝛼-mixing case is considered. This kind of dependence naturally appears in numerous models as Markov chains, GARCH-type models, and discretely observed diffusions (see, e.g., [1–3]). The problems where 𝑓 is the density of 𝑉1 or a regression function have received a lot of attention. A partial list of related works includes Robinson [4], Roussas [5, 6], Truong and Stone [7], Tran [8], Masry [9, 10], Masry and Fan [11], Bosq [12], and Liebscher [13]. For an efficient estimation of 𝑓, many methods can be considered. The most popular of them are based on kernels, splines and wavelets. In this note we deal with wavelet methods that have been introduced in i.i.d.setting by Donoho and Johnstone [14, 15] and Donoho et al. [16, 17]. These methods enjoy remarkable local adaptivity against discontinuities and spatially varying degree of oscillations. Complete reviews and discussions on wavelets in statistics can be found in, for example, Antoniadis [18] and Härdle et al. [19]. In the context of 𝛼-mixing dependence, various wavelet methods have been elaborated for a wide variety of nonparametric problems. Recent developments can be found in, for example, Leblanc [20], Tribouley and Viennet [21], Masry [22], Patil and Truong [23], Doosti et al. [24], Doosti and Niroumand [25], Doosti et al. [26], Cai and Liang [27], Niu and Liang [28], Benatia and Yahia [29], Chesneau [30– 32], Chaubey and Shirazi [33], and Abbaszadeh and Emadi [34]. In the general dependent setting described above, we provide a theoretical contribution to the performance of a wavelet estimator based on a hard thresholding. This nonlinear wavelet procedure has the features to be fully adaptive and efficient over a large class of functions 𝑓 (see, e.g., [14–17, 35]). Following the spirit of Kerkyacharian and Picard [36], we determine necessary assumptions on (𝑉𝑡 )𝑡∈Z and the wavelet basis to ensure that the considered estimator attains a fast rate of convergence under the MISE over Besov balls. The obtained rate of convergence often corresponds to the near optimal one in the minimax sense for the standard i.i.d. case. The originality of our result is to be general and sharp; it can be applied for nonparametric models of different natures and improves some existing results. This fact is illustrated by the consideration of the density deconvolution estimation problem and the density estimation problem in a 2 Journal of Probability and Statistics GARCH-type model, improving ([30], Proposition 5.1) and ([31], Theorem 2), respectively. A last part is devoted to the regression model with random design. The obtained result completes the one of Patil and Truong [23]. The organization of this note is as follows. In the next section we describe the considered wavelet setting. The hard thresholding estimator and its rate of convergence under the MISE over Besov balls are presented in Section 3. Applications of our general result are given in Section 4. The proofs are carried out in Section 5. 𝑠 (𝑀) with 𝑠 > 0, 𝑝, 𝑟 ≥ 1 2.2. Besov Balls. We say that 𝑓 ∈ 𝐵𝑝,𝑟 and 𝑀 > 0 if and only if there exists a constant 𝐶 > 0 such that the wavelet coefficients of 𝑓 given by (4) satisfy 𝜏(1/2−1/𝑝) 2 󵄨 󵄨𝑝 ( ∑ 󵄨󵄨󵄨𝑐𝜏,𝑘 󵄨󵄨󵄨 ) 1/𝑝 𝑘∈Λ 𝜏 ∞ 𝑗(𝑠+1/2−1/𝑝) + ( ∑ (2 𝑗=𝜏 󵄨 󵄨𝑝 ( ∑ 󵄨󵄨󵄨󵄨𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 ) 1/𝑝 𝑟 )) 1/𝑟 ≤ 𝐶, 𝑘∈Λ 𝑗 2. Wavelets and Besov Balls (5) In this section we introduce some notations corresponding to wavelets and Besov balls. 2.1. Wavelet Basis. We consider the wavelet basis on [0, 1] constructs from the Daubechies wavelets db2N with 𝑁 ≥ 1 (see, e.g., [37]). A brief description of this basis is given below. Let 𝜙 and 𝜓 be the initial wavelet functions of the family db2N. These functions have the particularity to be compactly supported and to belong to the class C𝑎 for 𝑁 > 5𝑎. For any 𝑗 ≥ 0, we set Λ 𝑗 = {0, . . . , 2𝑗 − 1} and, for 𝑘 ∈ Λ 𝑗 , 𝜙𝑗,𝑘 (𝑥) = 2𝑗/2 𝜙 (2𝑗 𝑥 − 𝑘) , 𝜓𝑗,𝑘 (𝑥) = 2𝑗/2 𝜓 (2𝑗 𝑥 − 𝑘) . (1) With appropriated treatments at the boundaries, there exists an integer 𝜏 such that, for any integer ℓ ≥ 𝜏, B = {𝜙ℓ,𝑘 , 𝑘 ∈ Λ ℓ ; 𝜓𝑗,𝑘 ; 𝑗 ∈ N − {0, . . . , ℓ − 1}, 𝑘 ∈ Λ 𝑗 } is an orthonormal basis of L2 ([0, 1]), where L2 ([0, 1]) 1 󵄨2 󵄨 󵄩 󵄩 = {𝑓 : [0, 1] 󳨀→ R; 󵄩󵄩󵄩𝑓󵄩󵄩󵄩2 = (∫ 󵄨󵄨󵄨𝑓 (𝑥)󵄨󵄨󵄨 𝑑𝑥) 0 ∞ 𝑓 (𝑥) = ∑ 𝑐ℓ,𝑘 𝜙ℓ,𝑘 (𝑥) + ∑ ∑ 𝑑𝑗,𝑘 𝜓𝑗,𝑘 (𝑥) , 𝑗=ℓ 𝑘∈Λ 𝑗 (3) 𝑥 ∈ [0, 1] , where 𝑐𝑗,𝑘 and 𝑑𝑗,𝑘 denote the wavelet coefficients of 𝑓 defined by 0 3.1. Statistical Framework. As mentioned in Section 1, a nonparametric estimation setting as general as possible is adopted: we aim to estimate an unknown function 𝑓 ∈ L2 ([0, 1]) via 𝑛 random variables (or vectors) 𝑉1 , . . . , 𝑉𝑛 from a strictly stationary stochastic process (𝑉𝑡 )𝑡∈Z defined on a probability space (Ω, A, P). We suppose that (𝑉𝑡 )𝑡∈Z has a 𝛼-mixing dependence structure with exponential decay rate; that is, there exist two constants 𝛾 > 0 and 𝜃 > 0 such that 𝑚≥1 < ∞} . For any integer ℓ ≥ 𝜏 and 𝑓 ∈ L ([0, 1]), we have the following wavelet expansion: 1 3. Statistical Framework, Estimator and Result sup (𝑒𝜃𝑚 𝛼𝑚 ) ≤ 𝛾, 2 𝑐𝑗,𝑘 = ∫ 𝑓 (𝑥) 𝜙𝑗,𝑘 (𝑥) 𝑑𝑥, Remark 1. We have chosen a wavelet basis on [0, 1] to fix the notations; wavelet basis on another interval can be considered in the rest of the study without affecting the results. 1/2 (2) 𝑘∈Λ ℓ with the usual modifications if 𝑝 = ∞ or 𝑟 = ∞. Note that, for 𝑠 particular choices of 𝑠, 𝑝, and 𝑟, 𝐵𝑝,𝑟 (𝑀) contains the classical Hölder and Sobolev balls (see, e.g., [40] and [19]). 1 where 𝛼𝑚 = sup(𝐴,𝐵)∈F𝑉−∞,0 ×F𝑉𝑚,∞ \|P(𝐴 ∩ 𝐵) − P(𝐴)P(𝐵)\|, F𝑉 −∞,0 is the 𝜎-algebra generated by the random variables (or vectors) . . . , 𝑉−1 , 𝑉0 and F𝑉 𝑚,∞ is the 𝜎-algebra generated by the random variables (or vectors) 𝑉𝑚 , 𝑉𝑚+1 , . . .. The 𝛼-mixing dependence is reasonably weak; it is satisfied by a wide variety of models including Markov chains, GARCH-type models, and discretely observed diffusions (see, for instance, [1–3, 41]). The considered estimator for 𝑓 is presented below. 3.2. Estimator. We define the hard thresholding wavelet estimator 𝑓̂ by 𝑑𝑗,𝑘 = ∫ 𝑓 (𝑥) 𝜓𝑗,𝑘 (𝑥) 𝑑𝑥. 0 (4) Technical details can be found in, for example, Cohen et al. [38] and Mallat [39]. In the main result of this paper, we will investigate the MISE rate of the proposed estimator by assuming that the unknown function of interest 𝑓 belongs to a wide class of functions: the Besov class. Its definition in terms of wavelet coefficients is presented in the following. (6) 𝑗1 𝑓̂ (𝑥) = ∑ 𝑐̂𝑗0 ,𝑘 𝜙𝑗0 ,𝑘 (𝑥) + ∑ ∑ 𝑑̂𝑗,𝑘 1{\|𝑑̂𝑗,𝑘 \|≥𝜅𝜆 𝑗 } 𝜓𝑗,𝑘 (𝑥) , 𝑘∈Λ 𝑗0 𝑗=𝑗0 𝑘∈Λ 𝑗 (7) where 𝑐̂𝑗,𝑘 = 1 𝑛 ∑𝑞 (𝜙𝑗,𝑘 , 𝑉𝑖 ) , 𝑛 𝑖=1 1 𝑛 𝑑̂𝑗,𝑘 = ∑𝑞 (𝜓𝑗,𝑘 , 𝑉𝑖 ) , 𝑛 𝑖=1 (8) Journal of Probability and Statistics 3 1 is the indicator function, 𝜅 > 0 is a large enough constant, 𝑗0 is the integer satisfying 𝑗0 2 = [𝜏 ln 𝑛] , (9) where [𝑎] denotes the integer part of 𝑎 and 𝑗1 is the integer satisfying 2𝑗1 = [( 1/(2𝜌+1) 𝑛 ) ], (ln 𝑛)3 𝜆 𝑗 = 2𝜌𝑗 √ ln 𝑛 . 𝑛 (10) (11) Here it is supposed that there exists a function 𝑞 : L2 ([0, 1]) × 𝑉1 (Ω) → C such that (H1) for 𝛾 ∈ {𝜙, 𝜓}, any integer 𝑗 ≥ 𝑗0 and 𝑘 ∈ Λ 𝑗 , 1 E (𝑞 (𝛾𝑗,𝑘 , 𝑉1 )) = ∫ 𝑓 (𝑥) 𝛾𝑗,𝑘 (𝑥) 𝑑𝑥, 0 Lemma 2. We make the following assumptions. (F1) Let 𝑢 be the density of 𝑉1 and let 𝑢(𝑉1 ,𝑉𝑚+1 ) be the density of (𝑉1 , 𝑉𝑚+1 ) for any 𝑚 ∈ Z. We suppose that there exists a constant 𝐶 > 0 such that 󵄨󵄨 󵄨󵄨𝑢(𝑉1 ,𝑉𝑚+1 ) (𝑥, 𝑦) sup sup 󵄨 𝑚∈{1,...,𝑛−1}Z (𝑥,𝑦)∈𝑉1 (Ω)×𝑉𝑚+1 (Ω) (14) 󵄨󵄨 −𝑢 (𝑥) 𝑢 (𝑦)󵄨󵄨 ≤ 𝐶. (F2) There exist two constants, 𝐶 > 0 and 𝜌 ≥ 0, satisfying, for 𝛾 ∈ {𝜙, 𝜓}, for any integer 𝑗 ≥ 𝑗0 and 𝑘 ∈ Λ 𝑗 , 󵄨󵄨 󵄨 󵄨󵄨𝑞 (𝛾𝑗,𝑘 , 𝑥)󵄨󵄨󵄨 𝑑𝑥 ≤ 𝐶2𝜌𝑗 2−𝑗/2 . 󵄨 󵄨 ∫ 𝑉1 (Ω) (15) Then, under (F1) and (F2), (H2)-(iii) is satisfied. (12) where E denotes the expectation, (H2) there exist two constants, 𝐶 > 0 and 𝜌 ≥ 0, satisfying, for 𝛾 ∈ {𝜙, 𝜓}, for any integer 𝑗 ≥ 𝑗0 and 𝑘 ∈ Λ 𝑗 , (i) sup𝑥∈𝑉1 (Ω) \|𝑞(𝛾𝑗,𝑘 , 𝑥)\| ≤ 𝐶2𝜌𝑗 2𝑗/2 , (ii) E(\|𝑞(𝛾𝑗,𝑘 , 𝑉1 )\|2 ) ≤ 𝐶22𝜌𝑗 , (iii) for any 𝑚 ∈ {1, . . . , 𝑛 − 1} ≥ 1, 󵄨󵄨󵄨C (𝑞 (𝛾 , 𝑉 ) , 𝑞 (𝛾 , 𝑉 ))󵄨󵄨󵄨 ≤ 𝐶22𝜌𝑗 2−𝑗 , (13) 󵄨󵄨 𝑜V 𝑗,𝑘 𝑚+1 𝑗,𝑘 1 󵄨󵄨 where C𝑜V denotes the covariance; that is, C𝑜V (𝑋, 𝑌) = E(𝑋𝑌) − E(𝑋)E(𝑌), 𝑌 denotes the complex conjugate of 𝑌. For well-known nonparametric models in the i.i.d. setting, hard thresholding wavelet estimators and important results can be found in, for example, Donoho and Johnstone [14, 15], Donoho et al. [16, 17], Delyon and Juditsky [35], Kerkyacharian and Picard [36], and Fan and Koo [42]. In the 𝛼-mixing context, 𝑓̂ defined by (7) is a general and improved version of the estimator considered in Chesneau [30, 31]. The main differences are the presence of the tuning parameter 𝜌 and the global definition of the function 𝑞 offering numerous possibilities of applications. Three of them are explored in Section 4. Comments on the Assumptions. The assumption (H1) ensures that (8) are unbiased estimators for 𝑐𝑗,𝑘 and 𝑑𝑗,𝑘 given by (4), whereas (H2) is related to their good performance. See Proposition 10. These assumptions are not too restrictive. For instance, if we consider the standard density estimation problem where (𝑉𝑡 )𝑡∈Z are i.i.d. random variables with bounded density 𝑓, the function 𝑞(𝛾, 𝑥) = 𝛾(𝑥) satisfies (H1) and (H2) with 𝜌 = 0 (note that, thanks to the independence of (𝑉𝑡 )𝑡∈Z , the covariance term in (H2)-(iii) is zero). The technical details are given in Donoho et al. [17]. Lemma 2 describes a simple situation in which assumption (H2)-(iii) is satisfied. 3.3. Result. Theorem 3 determines the rate of convergence attained by 𝑓̂ under the MISE over Besov balls. Theorem 3. We consider the general statistical setting described in Section 3.1. Let 𝑓̂ be (7) under (H1) and (H2). 𝑠 Suppose that 𝑓 ∈ 𝐵𝑝,𝑟 (𝑀) with 𝑟 ≥ 1, {𝑝 ≥ 2 and 𝑠 ∈ (0, 𝑁)}, or {𝑝 ∈ [1, 2) and 𝑠 ∈ ((2𝜌 + 1)/𝑝, 𝑁)}. Then there exists a constant 𝐶 > 0 such that ln 𝑛 2𝑠/(2𝑠+2𝜌+1) 󵄩2 󵄩 . ) E (󵄩󵄩󵄩󵄩𝑓̂ − 𝑓󵄩󵄩󵄩󵄩2 ) ≤ 𝐶( 𝑛 (16) The rate of convergence “((ln 𝑛)/𝑛)2𝑠/(2𝑠+2𝜌+1) ” is often the near optimal one in the minimax sense for numerous statistical problems in a i.i.d. setting (see, e.g., [19, 43]). Moreover, note that Theorem 3 is flexible; the assumptions on (𝑉𝑡 )𝑡∈Z , related to the definition of 𝑞 in (H1) and (H2), are mild. In the next section, this flexibility is illustrated for three sophisticated nonparametric estimation problems: the density deconvolution estimation problem, the density estimation problem in a GARCH-type model, and the regression function estimation in the regression model with random design. 4. Applications 4.1. Density Deconvolution. Let (𝑉𝑡 )𝑡∈Z be a strictly stationary stochastic process such that 𝑉𝑡 = 𝑋𝑡 + 𝜖𝑡 , 𝑡 ∈ Z, (17) where (𝑋𝑡 )𝑡∈Z is a strictly stationary stochastic process with unknown density 𝑓 and (𝜖𝑡 )𝑡∈Z is a strictly stationary stochastic process with known density 𝑔. It is supposed that 𝜖𝑡 and 𝑋𝑡 are independent for any 𝑡 ∈ Z and (𝑉𝑡 )𝑡∈Z is a 𝛼-mixing process with exponential decay rate (see Section 3.1 for a precise definition). Our aim is to estimate 𝑓 via 𝑉1 , . . . , 𝑉𝑛 from (𝑉𝑡 )𝑡∈Z . Some related works are Masry [44], Kulik [45], Comte et al. [46], and Van Zanten and Zareba [47]. We formulate the following assumptions. 4 Journal of Probability and Statistics (G1) The support of 𝑓 is [0, 1]. (18) (i) exactly the rate of convergence attained by the hard thresholding wavelet estimator, (ii) the near optimal rate of convergence in the minimax sense. (G3) Let 𝑢 be the density of 𝑉1 . We suppose that there exists a constant 𝐶 > 0 such that The details can be found in Fan and Koo [42]. Thus, Theorem 4 can be viewed as an extension of this existing result to the weak dependent case. (G2) There exists a constant 𝐶 > 0 such that sup𝑓 (𝑥) ≤ 𝐶 < ∞. 𝑥∈R sup 𝑢 (𝑥) ≤ 𝐶. (19) 𝑥∈R (G4) For any 𝑚 ∈ Z, let 𝑢(𝑉1 ,𝑉𝑚+1 ) be the density of (𝑉1 , 𝑉𝑚+1 ). We suppose that there exists a constant 𝐶 > 0 such that sup sup 𝑢(𝑉1 ,𝑉𝑚+1 ) (𝑥, 𝑦) ≤ 𝐶. (20) 𝑚∈Z (𝑥,𝑦)∈R2 (G5) For any integrable function 𝛾, we define its Fourier transform by F (𝛾) (𝑥) = ∫ ∞ −∞ 𝛾 (𝑦) 𝑒−𝑖𝑥𝑦 𝑑𝑦, 𝑥 ∈ R. (21) We suppose that there exist three known constants 𝐶 > 0, 𝑐 > 0, and 𝛿 > 1 such that, for any 𝑥 ∈ R, (i) the Fourier transform of 𝑔 satisfies 󵄨 󵄨󵄨 󵄨󵄨F (𝑔) (𝑥)󵄨󵄨󵄨 ≥ 𝑐 𝛿/2 (1 + 𝑥2 ) , (22) (ii) for any ℓ ∈ {0, 1, 2}, the ℓth derivative of the Fourier transform of 𝑔 satisfies (23) Theorem 4. We consider the model (17). Suppose that (G1)– (G5) are satisfied. Let 𝑓̂ be defined as in (7) with 1 ∞ F (𝛾) (𝑦) −𝑖𝑦𝑥 𝑒 𝑑𝑦, ∫ 2𝜋 −∞ F (𝑔) (𝑦) (24) where F(𝛾)(𝑦) denotes the complex conjugate of F(𝛾)(𝑦) and 𝜌 = 𝛿 (appearing in(G5)). 𝑠 (𝑀) with 𝑟 ≥ 1, {𝑝 ≥ 2, and 𝑠 ∈ Suppose that 𝑓 ∈ 𝐵𝑝,𝑟 (0, 𝑁)} or {𝑝 ∈ [1, 2) and 𝑠 ∈ ((2𝛿 + 1)/𝑝, 𝑁)}. Then there exists a constant 𝐶 > 0 such that 2𝑠/(2𝑠+2𝛿+1) (26) (𝜎𝑡2 )𝑡∈Z is a strictly stationary stochastic process with unknown density 𝑓, and (𝑍𝑡 )𝑡∈Z is a strictly stationary stochastic process with known density 𝑔. It is supposed that 𝜎𝑡2 and 𝑍𝑡 are independent for any 𝑡 ∈ Z and (𝑉𝑡 )𝑡∈Z is a 𝛼mixing process with exponential decay rate (see Section 3.1 for a precise definition). Our aim is to estimate 𝑓 via 𝑉1 , . . . , 𝑉𝑛 from (𝑉𝑡 )𝑡∈Z . Some related works are Comte et al. [46] and Chesneau [31]. We formulate the following assumptions. (J1) There exists a positive integer 𝛿 such that 𝑔 (𝑥) = 1 (− ln 𝑥)𝛿−1 , (𝛿 − 1)! 𝑥 ∈ [0, 1] . (27) Let us remark that 𝑔 is the density of ∏𝛿𝑖=1 𝑈𝑖 , where 𝑈1 , . . . , 𝑈𝛿 are 𝛿 i.i.d. random variables having the common distribution U([0, 1]). (J2) The support of 𝑓 is [0, 1] and 𝑓 ∈ L2 ([0, 1]). (J3) Let 𝑢 be the density of 𝑉1 . We suppose that there exists a constant 𝐶 > 0 such that . (28) 𝑥∈R (J4) For any 𝑚 ∈ Z, let 𝑢(𝑉1 ,𝑉𝑚+1 ) be the density of (𝑉1 , 𝑉𝑚+1 ). We suppose that there exists a constant 𝐶 > 0 such that We are now in the position to present the result. ln 𝑛 󵄩2 󵄩 E (󵄩󵄩󵄩󵄩𝑓̂ − 𝑓󵄩󵄩󵄩󵄩2 ) ≤ 𝐶( ) 𝑛 𝑉𝑡 = 𝜎𝑡2 𝑍𝑡 , sup 𝑢 (𝑥) ≤ 𝐶. 󵄨 󵄨󵄨 𝐶 󵄨󵄨(F (𝑔) (𝑥))(ℓ) 󵄨󵄨󵄨 ≤ . 󵄨󵄨 󵄨󵄨 (1 + \|𝑥\|)𝛿+ℓ 𝑞 (𝛾, 𝑥) = 4.2. GARCH-Type Model. We consider the strictly stationary stochastic process (𝑉𝑡 )𝑡∈Z where, for any 𝑡 ∈ Z, (25) Theorem 4 improves ([30], Proposition 5.1) in terms of rate of convergence; we gain a logarithmic term. Moreover, it is established that, in the i.i.d. setting, “((ln 𝑛)/𝑛)2𝑠/(2𝑠+2𝛿+1) ” is sup sup 𝑢(𝑉1 ,𝑉𝑚+1 ) (𝑥, 𝑦) ≤ 𝐶. (29) 𝑚∈Z (𝑥,𝑦)∈R2 We are now in the position to present the result. Theorem 5. We consider model (26). Suppose that (J1)–(J4) are satisfied. Let 𝑓̂ be defined as in (7) with 𝑞 (𝛾, 𝑥) = 𝑇𝛿 (𝛾) (𝑥) , (30) 󸀠 where, for any positive integer ℓ, 𝑇(𝛾)(𝑥) = (𝑥𝛾(𝑥)) and 𝑇ℓ (𝛾)(𝑥) = 𝑇(𝑇ℓ−1 (𝛾))(𝑥) and 𝜌 = 𝛿 (appearing in (J1)). 𝑠 (𝑀) with 𝑟 ≥ 1, {𝑝 ≥ 2 and 𝑠 ∈ Suppose that 𝑓 ∈ 𝐵𝑝,𝑟 (0, 𝑁)}, or {𝑝 ∈ [1, 2) and 𝑠 ∈ ((2𝛿 + 1)/𝑝, 𝑁)}. Then there exists a constant 𝐶 > 0 such that ln 𝑛 2𝑠/(2𝑠+2𝛿+1) 󵄩2 󵄩 . E (󵄩󵄩󵄩󵄩𝑓̂ − 𝑓󵄩󵄩󵄩󵄩2 ) ≤ 𝐶( ) 𝑛 (31) Theorem 5 significantly improves ([31], Theorem 2) in terms of rate of convergence; we gain an exponent 1/2. Journal of Probability and Statistics 5 4.3. Nonparametric Regression Model. We consider the strictly stationary stochastic process (𝑉𝑡 )𝑡∈Z where, for any 𝑡 ∈ Z, 𝑉𝑡 = (𝑌𝑡 , 𝑋𝑡 ), 𝑌𝑡 = 𝑓 (𝑋𝑡 ) + 𝜉𝑡 , (32) (𝑋𝑡 )𝑡∈Z is a strictly stationary stochastic process with unknown density 𝑔, (𝜉𝑡 )𝑡∈Z is a strictly stationary centered stochastic process, and 𝑓 is the unknown regression function. It is supposed that 𝑋𝑡 and 𝜉𝑡 are independent for any 𝑡 ∈ Z and (𝑉𝑡 )𝑡∈Z is a 𝛼-mixing process with exponential decay rate (see Section 3.1 for a precise definition). Our aim is to estimate 𝑓 via 𝑉1 , . . . , 𝑉𝑛 from (𝑉𝑡 )𝑡∈Z . Applications of this problem can be found in Härdle [48]. Wavelet methods can be found in Patil and Truong [23], Doosti et al. [24], Doosti et al. [26], and Doosti and Niroumand [25]. We formulate the following assumptions. (K1) The support of 𝑓 and 𝑔 is [0, 1] and 𝑓 and 𝑔 ∈ L2 ([0, 1]). (K2) 𝜉1 (Ω) is bounded. 󵄨 󵄨 sup 󵄨󵄨󵄨𝑓 (𝑥)󵄨󵄨󵄨 ≤ 𝐶. (33) (K4) There exist two constants 𝑐∗ > 0 and 𝐶 > 0 such that 𝑐∗ ≤ inf 𝑔 (𝑥) , sup 𝑔 (𝑥) ≤ 𝐶. 𝑥∈[0,1] 𝑥∈[0,1] (34) (K5) Let 𝑢 be the density of 𝑉1 . We suppose that there exists a constant 𝐶 > 0 such that sup 𝑢 (𝑥) ≤ 𝐶. 𝑥∈R×[0,1] (35) (K6) For any 𝑚 ∈ Z, let 𝑢(𝑉1 ,𝑉𝑚+1 ) be the density of (𝑉1 , 𝑉𝑚+1 ). We suppose that there exists a constant 𝐶 > 0 such that sup sup 𝑢(𝑉1 ,𝑉𝑚+1 ) (𝑥, 𝑦) ≤ 𝐶. 𝑚∈Z (𝑥,𝑦)∈(R×[0,1])×(R×[0,1]) Theorem 6. We consider the model (32). Suppose that (K1)– (K6) are satisfied. Let 𝑓̂ be the truncated ratio estimator. Consider (37) where (i) ̂V is defined as in (7) with 𝑞 (𝛾, (𝑥, 𝑥∗ )) = 𝑥𝛾 (𝑥∗ ) and 𝜌 = 0, (39) and 𝜌 = 0, (iii) 𝑐∗ is the constant defined in (K4). 𝑠 𝑠 Suppose that 𝑓𝑔 ∈ 𝐵𝑝,𝑟 (𝑀) and 𝑔 ∈ 𝐵𝑝,𝑟 (𝑀) with 𝑟 ≥ 1, {𝑝 ≥ 2 and 𝑠 ∈ (0, 𝑁)} or {𝑝 ∈ [1, 2), and 𝑠 ∈ (1/𝑝, 𝑁)}. Then there exists a constant 𝐶 > 0 such that ln 𝑛 2𝑠/(2𝑠+1) 󵄩2 󵄩 . ) E (󵄩󵄩󵄩󵄩𝑓̂ − 𝑓󵄩󵄩󵄩󵄩2 ) ≤ 𝐶( 𝑛 (40) The estimator (37) is derived by combining the procedure of Patil and Truong [23] with the truncated approach of Vasiliev [49]. Theorem 6 completes Patil and Truong [23] in terms of rates of convergence under the MISE over Besov balls. Conclusion. Considering the weak dependent case on the observations, we prove a general result on the rate of convergence attains by a hard wavelet thresholding estimator under the MISE over Besov balls. This result is flexible; it can be applied for a wide class of statistical models. Moreover, the obtained rate of convergence is sharp; it can correspond to the near optimal one in the minimax sense for the standard i.i.d. case. Some recent results on sophisticated statistical problems are improved. Thanks to its flexibility, the perspectives of applications of our theoretical result in other contexts are numerous. 5. Proofs In this section, 𝐶 denotes any constant that does not depend on 𝑗, 𝑘, and 𝑛. Its value may change from one term to another and may depend on 𝜙 or 𝜓. (36) We are now in the position to present the result. ̂V (𝑥) 𝑓̂ (𝑥) = , 1 ̂ ∗ /2} 𝑔̂ (𝑥) {\|𝑔(𝑥)\|≥𝑐 𝑞 (𝛾, 𝑥) = 𝛾 (𝑥) Remark 7. The assumption(K2) can be relaxed with another strategy to the one developed in Theorem 6. Some technical elements are given in Chesneau [32]. (K3) There exists a constant 𝐶 > 0 such that 𝑥∈[0,1] (ii) 𝑔̂ is defined as in (7) with 𝑋𝑡 instead of 𝑉𝑡 , (38) 5.1. Key Lemmas. Let us present two lemmas which will be used in the proofs. Lemma 8 shows a sharp covariance inequality under the 𝛼-mixing condition. Lemma 8 (see [50]). Let (𝑊𝑡 )𝑡∈Z be a strictly stationary 𝛼mixing process with mixing coefficient 𝛼𝑚 , 𝑚 ≥ 0, and let ℎ and 𝑘 be two measurable functions. Let 𝑝 > 0 and 𝑞 > 0 satisfying 1/𝑝 + 1/𝑞 < 1 such that E(\|ℎ(𝑊1 )\|𝑝 ) and E(\|𝑘(𝑊1 )\|𝑞 ) exist. Then there exists a constant 𝐶 > 0 such that 󵄨󵄨󵄨C𝑜V (ℎ (𝑊1 ) , 𝑘 (𝑊𝑚+1 ))󵄨󵄨󵄨 󵄨 󵄨 (41) 󵄨 󵄨𝑝 1/𝑝 󵄨 󵄨𝑞 1/𝑞 1−1/𝑝−1/𝑞 ≤ 𝐶𝛼𝑚 (E (󵄨󵄨󵄨ℎ (𝑊1 )󵄨󵄨󵄨 )) (E(󵄨󵄨󵄨𝑘 (𝑊1 )󵄨󵄨󵄨 )) . Lemma 9 below presents a concentration inequality for 𝛼-mixing processes. 6 Journal of Probability and Statistics Lemma 9 (see [13]). Let (𝑊𝑡 )𝑡∈Z be a strictly stationary process with the 𝑚th strongly mixing coefficient 𝛼𝑚 , 𝑚 ≥ 0, let 𝑛 be a positive integer, let ℎ : R → C be a measurable function, and, for any 𝑡 ∈ Z, 𝑈𝑡 = ℎ(𝑊𝑡 ). We assume that E(𝑈1 ) = 0 and there exists a constant 𝑀 > 0 satisfying \|𝑈1 \| ≤ 𝑀. Then, for any 𝑚 ∈ {1, . . . , [𝑛/2]} and 𝜆 > 0, we have 󵄨 󵄨 1 󵄨󵄨 𝑛 󵄨󵄨 P ( 󵄨󵄨󵄨󵄨 ∑ 𝑈𝑖 󵄨󵄨󵄨󵄨 ≥ 𝜆) 𝑛 󵄨󵄨𝑖=1 󵄨󵄨 ≤ 4 exp (− 𝜆2 𝑛 𝑀 ) + 32 𝑛𝛼𝑚 , 𝜆 16 (𝐷𝑚 /𝑚 + 𝜆𝑀𝑚/3) (42) (c) Let 𝜆 𝑗 be defined as in (11). There exists a constant 𝐶 > 0 such that, for any 𝜅 large enough, 𝑗 ∈ {𝑗0 , . . . , 𝑗1 } and 𝑘 ∈ Λ 𝑗 , we have 1 󵄨 󵄨 𝜅𝜆 𝑗 P (󵄨󵄨󵄨󵄨𝑑̂𝑗,𝑘 − 𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 ≥ ) ≤ 𝐶 4. 2 𝑛 Proof of Proposition 10. (a) Using (H1) and the stationarity of (𝑉𝑡 )𝑡∈Z , we obtain 󵄨 󵄨2 E (󵄨󵄨󵄨󵄨𝑐̂𝑗,𝑘 − 𝑐𝑗,𝑘 󵄨󵄨󵄨󵄨 ) = V (̂𝑐𝑗,𝑘 ) where 𝐷𝑚 = 𝑙∈{1,...,2𝑚} 1 𝑛 2 ∑V (𝑞 (𝜙𝑗,𝑘 , 𝑉𝑖 )) + 2 𝑛2 𝑖=1 𝑛 = 𝑙 max V (∑𝑈𝑖 ) . (43) 𝑛 V−1 𝑖=1 × ∑ ∑ Re (C𝑜V (𝑞 (𝜙𝑗,𝑘 , 𝑉V ) , 𝑞 (𝜙𝑗,𝑘 , 𝑉ℓ ))) V=2ℓ=1 5.2. Intermediary Results Proof of Lemma 2. Using a standard expression of the covariance, and (F1) as well as(F2), we obtain 󵄨󵄨 󵄨 󵄨󵄨C𝑜V (𝑞 (𝛾𝑗,𝑘 , 𝑉𝑚+1 ) , 𝑞 (𝛾𝑗,𝑘 , 𝑉1 ))󵄨󵄨󵄨 󵄨 󵄨 󵄨󵄨 󵄨 = 󵄨󵄨󵄨󵄨∫ 𝑞 (𝛾𝑗,𝑘 , 𝑥) 𝑞 (𝛾𝑗,𝑘 , 𝑦) ∫ 󵄨󵄨 𝑉1 (Ω) 𝑉1 (Ω) 󵄨󵄨 󵄨 × (𝑢(𝑉1 ,𝑉𝑚+1 ) (𝑥, 𝑦) − 𝑢 (𝑥) 𝑢 (𝑦)) 𝑑𝑥𝑑𝑦󵄨󵄨󵄨󵄨 󵄨󵄨 1 𝑛 2 ∑V (𝑞 (𝜙𝑗,k , 𝑉𝑖 )) + 2 2 𝑛 𝑖=1 𝑛 = 𝑛−1 × ∑ (𝑛 − 𝑚) Re (C𝑜V (𝑞 (𝜙𝑗,𝑘 , 𝑉𝑚+1 ) , 𝑞 (𝜙𝑗,𝑘 , 𝑉1 ))) 𝑚=1 1 󵄨2 󵄨 (E (󵄨󵄨󵄨󵄨𝑞 (𝜙𝑗,𝑘 , 𝑉1 )󵄨󵄨󵄨󵄨 ) 𝑛 ≤ 𝑛−1 󵄨 󵄨 +2 ∑ 󵄨󵄨󵄨󵄨C𝑜V (𝑞 (𝜙𝑗,𝑘 , 𝑉𝑚+1 ) , 𝑞 (𝜙𝑗,𝑘 , 𝑉1 ))󵄨󵄨󵄨󵄨) . 󵄨󵄨 󵄨󵄨 󵄨 󵄨󵄨𝑞 (𝛾𝑗,𝑘 , 𝑥)󵄨󵄨󵄨 󵄨󵄨󵄨𝑞 (𝛾𝑗,𝑘 , 𝑦)󵄨󵄨󵄨 ≤∫ ∫ 󵄨 󵄨 󵄨 󵄨 𝑉 (Ω) 𝑉 (Ω) 1 𝑚=1 (49) 1 󵄨 󵄨 × 󵄨󵄨󵄨󵄨𝑢(𝑉1 ,𝑉𝑚+1 ) (𝑥, 𝑦) − 𝑢 (𝑥) 𝑢 (𝑦)󵄨󵄨󵄨󵄨 𝑑𝑥 𝑑𝑦 ≤ 𝐶(∫ 𝑉1 (Ω) (48) By (H2)-(ii) we get 2 󵄨󵄨 󵄨 󵄨󵄨𝑞 (𝛾𝑗,𝑘 , 𝑥)󵄨󵄨󵄨 𝑑𝑥) ≤ 𝐶22𝜌𝑗 2−𝑗 . 󵄨 󵄨 󵄨 󵄨2 E (󵄨󵄨󵄨󵄨𝑞(𝜙𝑗,𝑘 , 𝑉1 )󵄨󵄨󵄨󵄨 ) ≤ 𝐶22𝜌𝑗 . (44) For the covariance term, note that This ends the proof of Lemma 2. Proposition 10 proves probability and moments inequalities satisfied by the estimators (8). Proposition 10. Let 𝛼̂𝑗,𝑘 and 𝛽̂𝑗,𝑘 be defined as in (8) under (H1) and (H2), let 𝑗0 be (9) and let 𝑗1 be (10). (a) There exists a constant 𝐶 > 0 such that, for any 𝑗 ∈ {𝑗0 , . . . , 𝑗1 } and 𝑘 ∈ Λ 𝑗 , 1 󵄨 󵄨2 E (󵄨󵄨󵄨󵄨𝑐̂𝑗,𝑘 − 𝑐𝑗,𝑘 󵄨󵄨󵄨󵄨 ) ≤ 𝐶22𝜌𝑗 , 𝑛 (45) 1 󵄨2 󵄨 E (󵄨󵄨󵄨󵄨𝑑̂𝑗,𝑘 − 𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 ) ≤ 𝐶22𝜌𝑗 . 𝑛 (46) (b) There exists a constant 𝐶 > 0 such that, for any 𝑗 ∈ {𝑗0 , . . . , 𝑗1 } and 𝑘 ∈ Λ 𝑗 , 󵄨4 󵄨 E (󵄨󵄨󵄨󵄨𝑑̂𝑗,𝑘 − 𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 ) ≤ 𝐶24𝜌𝑗 . (50) (47) 𝑛−1 󵄨 󵄨 ∑ 󵄨󵄨󵄨󵄨C𝑜V (𝑞 (𝜙𝑗,𝑘 , 𝑉𝑚+1 ) , 𝑞 (𝜙𝑗,𝑘 , 𝑉1 ))󵄨󵄨󵄨󵄨 = 𝐴 + 𝐵, (51) 𝑚=1 where 𝐴= [ln 𝑛/𝜃]−1 ∑ 𝑚=1 𝐵= 󵄨󵄨 󵄨 󵄨󵄨C𝑜V (𝑞 (𝜙𝑗,𝑘 , 𝑉𝑚+1 ) , 𝑞 (𝜙𝑗,𝑘 , 𝑉1 ))󵄨󵄨󵄨 , 󵄨 󵄨 𝑛−1 ∑ 𝑚=[(ln 𝑛)/𝜃] 󵄨󵄨 󵄨 󵄨󵄨C𝑜V (𝑞 (𝜙𝑗,𝑘 , 𝑉𝑚+1 ) , 𝑞 (𝜙𝑗,𝑘 , 𝑉1 ))󵄨󵄨󵄨 . 󵄨 󵄨 (52) It follows from (H2)-(iii) and 2−𝑗 ≤ 2−𝑗0 < 2(ln 𝑛)−1 that 𝐴 ≤ 𝐶22𝜌𝑗 2−𝑗 [ ln 𝑛 ] ≤ 𝐶22𝜌𝑗 . 𝜃 (53) Journal of Probability and Statistics 7 The Davydov inequality described in Lemma 8 with 𝑝 = 𝑞 = 4, (H2)-(i)-(ii), and 2𝑗 ≤ 2𝑗1 ≤ 𝑛 give 󵄨 󵄨4 𝐵 ≤ 𝐶√ E (󵄨󵄨󵄨󵄨𝑞(𝜙𝑗,𝑘 , 𝑉1 )󵄨󵄨󵄨󵄨 ) 𝑛−1 ∑ 𝑚=[(ln 𝑛)/𝜃] 󵄨 󵄨2 ≤ 𝐶2𝜌𝑗 2𝑗/2 √ E (󵄨󵄨󵄨󵄨𝑞(𝜙𝑗,𝑘 , 𝑉1 )󵄨󵄨󵄨󵄨 ) √𝛼𝑚 ∞ 𝑒−𝜃𝑚/2 ∑ (54) Owing to Lemma 9 applied with 𝑈1 , . . . , 𝑈𝑛 , 𝜆 = 𝜅𝜆 𝑗 /2, 𝑚 = [√𝜅 ln 𝑛], 𝑀 = 𝐶2𝜌𝑗 √𝑛/(ln 𝑛)3 , and the bound (61), we obtain 󵄨 𝜅𝜆 𝑗 󵄨 ) P (󵄨󵄨󵄨󵄨𝑑̂𝑗,𝑘 − 𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 ≥ 2 𝜅2 𝜆2𝑗 𝑛 ≤ 𝐶 (exp (−𝐶 𝑚=[(ln 𝑛)/𝜃] 𝐷𝑚 /𝑚 + 𝜅𝜆 𝑗 𝑚𝑀 = 𝐶22𝜌𝑗 √𝑛𝑒−(ln 𝑛)/2 ≤ 𝐶22𝜌𝑗 . )+ 𝑀 −𝜃𝑚 𝑛𝑒 ) 𝜆𝑗 ≤ 𝐶 ( exp (−𝐶 ((𝜅2 22𝜌𝑗 ln 𝑛) Thus 𝑛−1 󵄨 󵄨 ∑ 󵄨󵄨󵄨󵄨C𝑜V (𝑞 (𝜙𝑗,𝑘 , 𝑉𝑚+1 ) , 𝑞 (𝜙𝑗,𝑘 , 𝑉1 ))󵄨󵄨󵄨󵄨 ≤ 𝐶22𝜌𝑗 . (55) 𝑚=1 × (22𝜌𝑗 + 𝜅2𝜌𝑗 √ Putting (49), (50), and (55) together, the first point in (a) is proved. The proof of the second point is identical with 𝜓 instead of 𝜙. ≤ (b) Thanks to (H2)-(i), we have \|𝑑̂𝑗,𝑘 \| sup𝑥∈𝑉1 (Ω) \|𝑞(𝜓𝑗,𝑘 , 𝑥)\| ≤ 𝐶2𝜌𝑗 2𝑗/2 . It follows from the triangular inequality and \|𝑑𝑗,𝑘 \| ≤ ‖𝑓‖2 ≤ 𝐶 that 󵄨 󵄨 󵄨 󵄨 󵄨 󵄨󵄨 ̂ 󵄨󵄨𝑑𝑗,𝑘 − 𝑑𝑗,𝑘 󵄨󵄨󵄨 ≤ 󵄨󵄨󵄨𝑑̂𝑗,𝑘 󵄨󵄨󵄨 + 󵄨󵄨󵄨𝑑𝑗,𝑘 󵄨󵄨󵄨 ≤ 𝐶2𝜌𝑗 2𝑗/2 . 󵄨 󵄨 󵄨 󵄨 󵄨 󵄨 3 𝑛/(ln 𝑛) −𝜃[√𝜅 ln 𝑛] +√ ) 𝑛𝑒 (ln 𝑛) /𝑛 2 3/2 ≤ 𝐶 (𝑛−𝐶𝜅 /(1+𝜅 1 󵄨 󵄨4 󵄨2 󵄨 E (󵄨󵄨󵄨󵄨𝑑̂𝑗,𝑘 − 𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 ) ≤ 𝐶22𝜌𝑗 2𝑗 E (󵄨󵄨󵄨󵄨𝑑̂𝑗,𝑘 − 𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 ) ≤ 𝐶24𝜌𝑗 2𝑗 . 𝑛 (57) Using 2𝑗 ≤ 2𝑗1 ≤ 𝑛, the proof of (b) is completed. (c) We will use the Liebscher inequality described in Lemma 9. Let us set 𝑈𝑖 = 𝑞 (𝜓𝑗,𝑘 , 𝑉𝑖 ) − E (𝑞 (𝜓𝑗,𝑘 , 𝑉1 )) . −1 𝑛 × [√𝜅 ln 𝑛] 2 √ ) )) (ln 𝑛)3 𝜌𝑗 (56) This inequality and the second result of (a) yield (58) We have E(𝑈1 ) = 0 and, by(H2)-(i) and 2𝑗 ≤ 2𝑗1 ≤ 𝑛/(ln 𝑛)3 , 𝑛 󵄨 󵄨 󵄨󵄨 󵄨󵄨 𝜌𝑗 𝑗/2 𝜌𝑗 , 󵄨󵄨𝑈𝑖 󵄨󵄨 ≤ 2 sup 󵄨󵄨󵄨󵄨𝑞 (𝜓𝑗,𝑘 , 𝑥)󵄨󵄨󵄨󵄨 ≤ 𝐶2 2 ≤ 𝐶2 √ 𝑛)3 (ln 𝑥∈𝑉 (Ω) 1 (59) (so 𝑀 = 𝐶2𝜌𝑗 √𝑛/(ln 𝑛)3 ). Proceeding as for the proofs of the bounds in (a), for any integer 𝑙 ≤ 𝐶 ln 𝑛, since 2−𝑗 ≤ 2−𝑗0 ≤ 2(ln 𝑛)−1 , we show that (ln 𝑛) 𝑛 ) + 𝑛2−𝜃√𝜅 ) . (62) Taking 𝜅 large enough, the last term is bounded by 𝐶/𝑛4 . This completes the proof of(c). This completes the proof of Proposition 10. Proof of Theorem 3. Theorem 3 can be proved by combining arguments of ([36], Theorem 5.1) and ([51], Theorem 4.2). It is close to ([30], Proof of Theorem 2) by taking 𝜃 → ∞. The interested reader can find the details below. We consider the following wavelet decomposition for 𝑓: ∞ 𝑓 (𝑥) = ∑ 𝑐𝑗0 ,𝑘 𝜙𝑗0 ,𝑘 (𝑥) + ∑ ∑ 𝑑𝑗,𝑘 𝜓𝑗,𝑘 (𝑥) , 𝑗=𝑗0 𝑘∈Λ 𝑗 𝑘∈Λ 𝑗0 1 (63) 1 where 𝑐𝑗0 ,𝑘 = ∫0 𝑓(𝑥)𝜙𝑗0 ,𝑘 (𝑥)𝑑𝑥 and 𝑑𝑗,𝑘 = ∫0 𝑓(𝑥)𝜓𝑗,𝑘 (𝑥)𝑑𝑥. Using the orthonormality of the wavelet basis B, the MISE of 𝑓̂ can be expressed as 󵄩2 󵄩 E (󵄩󵄩󵄩󵄩𝑓̂ − 𝑓󵄩󵄩󵄩󵄩2 ) = 𝑃 + 𝑄 + 𝑅, (64) where 𝑙 V (∑𝑈𝑖 ) 𝑖=1 (60) 𝑙 = V (∑𝑞 (𝜓𝑗,𝑘 , 𝑉𝑖 )) ≤ 𝐶22𝜌𝑗 (𝑙 + 𝑙2 2−𝑗 ) ≤ 𝐶22𝜌𝑗 𝑙. 󵄨2 󵄨 𝑃 = ∑ E (󵄨󵄨󵄨󵄨𝑐̂𝑗0 ,𝑘 − 𝑐𝑗0 ,𝑘 󵄨󵄨󵄨󵄨 ) , 𝑘∈Λ 𝑗0 𝑗1 󵄨󵄨2 󵄨󵄨 𝑄 = ∑ ∑ E (󵄨󵄨󵄨𝑑̂𝑗,𝑘 1{\|𝑑̂𝑗,𝑘 \|≥𝜅𝜆 𝑗 } − 𝑑𝑗,𝑘 󵄨󵄨󵄨 ) , 󵄨 󵄨 𝑗=𝑗 𝑘∈Λ 𝑖=1 0 Therefore 𝐷𝑚 = 𝑙 ∞ 2𝜌𝑗 max V (∑𝑈𝑖 ) ≤ 𝐶2 𝑙∈{1,...,2𝑚} 𝑖=1 𝑗 𝑚. (61) 𝑅= ∑ 2 . ∑ 𝑑𝑗,𝑘 𝑗=𝑗1 +1 𝑘∈Λ 𝑗 (65) 8 Journal of Probability and Statistics Let us now investigate sharp upper bounds for 𝑃, 𝑅 and 𝑄 successively. Upper Bound for 𝑃. The point (a) of Proposition 10 and 2𝑠/(2𝑠 + 2𝜌 + 1) < 1 yield 𝑃≤𝐶 ln 𝑛 ln 𝑛 2𝑠/(2𝑠+2𝜌+1) 2𝑗0 . ≤𝐶 ≤ 𝐶( ) 𝑛 𝑛 𝑛 (66) Upper Bound for 𝑅. Upper Bound for 𝑄1 + 𝑄3 . Owing to the inequalities 1{\|𝑑̂𝑗,𝑘 \|<𝜅𝜆 𝑗 ,\|𝑑𝑗,𝑘 \|≥2𝜅𝜆 𝑗 } ≤ 1{\|𝐷̂𝑗,𝑘 \|>𝜅𝜆 𝑗 /2} , 1{\|𝑑̂𝑗,𝑘 \|≥𝜅𝜆 𝑗 ,\|𝑑𝑗,𝑘 \|<𝜅𝜆 𝑗 /2} ≤ 1{\|𝐷̂𝑗,𝑘 \|>𝜅𝜆 𝑗 /2} and 1{\|𝑑̂𝑗,𝑘 \|<𝜅𝜆 𝑗 ,\|𝑑𝑗,𝑘 \|≥2𝜅𝜆 𝑗 } ≤ 1{\|𝑑𝑗,𝑘 \|≤2\|𝐷̂𝑗,𝑘 \|} , the Cauchy-Schwarz inequality, and the points(b) and (c) of Proposition 10, we have 𝑗 1 󵄨 ̂ 󵄨󵄨2 𝑄1 + 𝑄3 ≤ 𝐶 ∑ ∑ E (󵄨󵄨󵄨󵄨𝐷 ̂𝑗,𝑘 \|>𝜅𝜆 𝑗 /2} ) 𝑗,𝑘 󵄨󵄨󵄨 1{\|𝐷 𝑗=𝑗0 𝑘∈Λ 𝑗 𝑠 𝑠 (i) For 𝑟 ≥ 1 and 𝑝 ≥ 2, we have 𝑓 ∈ 𝐵𝑝,𝑟 (𝑀) ⊆ 𝐵2,∞ (𝑀). Using 2𝑠/(2𝑠 + 2𝜌 + 1) < 2𝑠/(2𝜌 + 1), we obtain ∞ −2𝑗𝑠 𝑅≤𝐶 ∑ 2 𝑗=𝑗1 +1 ≤ 𝐶( (ln 𝑛)3 ≤ 𝐶( ) 𝑛 1/2 󵄨󵄨 ̂ 󵄨󵄨4 1/2 󵄨󵄨 ̂ 󵄨󵄨 𝜅𝜆 𝑗 󵄨 󵄨 󵄨 󵄨 ≤ 𝐶 ∑ ∑ (E (󵄨󵄨𝐷𝑗,𝑘 󵄨󵄨 )) (P (󵄨󵄨𝐷𝑗,𝑘 󵄨󵄨 > )) 2 𝑗=𝑗0 𝑘∈Λ 𝑗1 2𝑠/(2𝜌+1) 𝑗 𝑗 ≤𝐶 (67) ln 𝑛 2𝑠/(2𝑠+2𝜌+1) . ) 𝑛 1 1 𝑗(1+2𝜌) 1 ln 𝑛 2𝑠/(2𝑠+2𝜌+1) 2 ≤ 𝐶 . ≤ 𝐶( ) ∑ 𝑛2 𝑗=𝑗 𝑛 𝑛 0 (72) 𝑠 (𝑀) ⊆ (ii) For 𝑟 ≥ 1 and 𝑝 ∈ [1, 2), we have 𝑓 ∈ 𝐵𝑝,𝑟 𝑠+1/2−1/𝑝 𝐵2,∞ (𝑀). The condition 𝑠 > (2𝜌 + 1)/𝑝 implies that (𝑠 + 1/2 − 1/𝑝)/(2𝜌 + 1) > 𝑠/(2𝑠 + 2𝜌 + 1). Thus Upper Bound for 𝑄2 . It follows from the point(a) of Proposition 10 that ∞ 𝑅 ≤ 𝐶 ∑ 2−2𝑗(𝑠+1/2−1/𝑝) 𝑗 1 󵄨 ̂ 󵄨󵄨2 𝑄2 ≤ ∑ ∑ E (󵄨󵄨󵄨󵄨𝐷 𝑗,𝑘 󵄨󵄨󵄨 ) 1{\|𝑑𝑗,𝑘 \|≥𝜅𝜆 𝑗 /2} 𝑗=𝑗1 +1 𝑗=𝑗0 𝑘∈Λ 𝑗 2(𝑠+1/2−1/𝑝)/(2𝜌+1) (ln 𝑛)3 ≤ 𝐶( ) 𝑛 ≤ 𝐶( (68) ln 𝑛 2𝑠/(2𝑠+2𝜌+1) . ) 𝑛 Let us now introduce the integer 𝑗∗ defined by 2𝑗∗ = [( (69) ̂𝑗,𝑘 = 𝑑̂𝑗,𝑘 − 𝑑𝑗,𝑘 , Upper Bound for 𝑄. Adopting the notation 𝐷 𝑄 can be written as 𝑛 1/(2𝑠+2𝜌+1) ]. ) ln 𝑛 (74) Note that 𝑗∗ ∈ {𝑗0 , . . . , 𝑗1 } for 𝑛 large enough. Then 𝑄2 can be bounded as 𝑄2 ≤ 𝑄2,1 + 𝑄2,2 , 4 𝑄 = ∑𝑄𝑖 , 𝑗 0 Hence, for 𝑟 ≥ 1, {𝑝 ≥ 2 and 𝑠 > 0} or {𝑝 ∈ [1, 2), and 𝑠 > (2𝜌 + 1)/𝑝}, we have ln 𝑛 2𝑠/(2𝑠+2𝜌+1) . 𝑅 ≤ 𝐶( ) 𝑛 (73) 𝑗 1 1 ≤ 𝐶 ∑ 22𝜌𝑗 ∑ 1{\|𝑑𝑗,𝑘 \|>𝜅𝜆 𝑗 /2} . 𝑛 𝑗=j 𝑘∈Λ (75) (70) where 𝑖=1 where 𝑗 𝑗1 𝑄2,1 󵄨 ̂ 󵄨󵄨2 𝑄1 = ∑ ∑ E (󵄨󵄨󵄨󵄨𝐷 𝑗,𝑘 󵄨󵄨󵄨 1{\|𝑑̂𝑗,𝑘 \|≥𝜅𝜆 𝑗 ,\|𝑑𝑗,𝑘 \|<𝜅𝜆 𝑗 /2} ) , 0 𝑗=𝑗0 𝑘∈Λ 𝑗 𝑗1 𝑄2,2 𝑗=𝑗0 𝑘∈Λ 𝑗 (76) (71) On the one hand we have 2 E (𝑑𝑗,𝑘 1{\|𝑑̂𝑗,𝑘 \|<𝜅𝜆 𝑗 ,\|𝑑𝑗,𝑘 \|≥2𝜅𝜆 𝑗 } ) , 𝑗1 2 1{\|𝑑̂𝑗,𝑘 \|<𝜅𝜆 𝑗 ,\|𝑑𝑗,𝑘 \|<2𝜅𝜆 𝑗 } ) . 𝑄4 = ∑ ∑ E (𝑑𝑗,𝑘 𝑗=𝑗0 𝑘∈Λ 𝑗 1 1 = 𝐶 ∑ 22𝜌𝑗 ∑ 1{\|𝑑𝑗,𝑘 \|>𝜅𝜆 𝑗 /2} . 𝑛 𝑗=𝑗∗ +1 𝑘∈Λ 𝑗 𝑗=𝑗0 𝑘∈Λ 𝑗 𝑄3 = ∑ ∑ 𝑗 𝑗 󵄨 ̂ 󵄨󵄨2 𝑄2 = ∑ ∑ E (󵄨󵄨󵄨󵄨𝐷 𝑗,𝑘 󵄨󵄨󵄨 1{\|𝑑̂𝑗,𝑘 \|≥𝜅𝜆 𝑗 ,\|𝑑𝑗,𝑘 \|≥𝜅𝜆 𝑗 /2} ) , 𝑗1 1 ∗ = 𝐶 ∑ 22𝜌𝑗 ∑ 1{\|𝑑𝑗,𝑘 \|>𝜅𝜆 𝑗 /2} , 𝑛 𝑗=𝑗 𝑘∈Λ 𝑗 𝑄2,1 ≤ 𝐶 ln 𝑛 ∗ 𝑗(1+2𝜌) ln 𝑛 2𝑠/(2𝑠+2𝜌+1) ≤ 𝐶( . ) ∑2 𝑛 𝑗=𝑗0 𝑛 On the other hand, we have the following. (77) Journal of Probability and Statistics 9 (i) For 𝑟 ≥ 1 and 𝑝 ≥ 2, the Markov inequality and 𝑓 ∈ 𝑠 𝑠 (𝑀) ⊆ 𝐵2,∞ (𝑀) yield 𝐵𝑝,𝑟 ∞ 𝑗 𝑄2,2 ∞ ln 𝑛 1 2𝜌𝑗 1 2 2 ≤𝐶 𝑑 ≤ 𝐶 ∑ 2 ∑ ∑ ∑ 𝑑𝑗,𝑘 𝑗,𝑘 𝑛 𝑗=𝑗∗ +1 𝜆2𝑗 𝑘∈Λ 𝑗=𝑗∗ +1𝑘∈Λ 𝑗 𝑗 ln 𝑛 2𝑠/(2𝑠+2𝜌+1) ≤ 𝐶( . ) 𝑛 ∞ ≤ 𝐶 ∑ 2−2𝑗𝑠 𝑗=𝑗∗ +1 𝑗 𝑄2,2 ≤ 𝐶( ≤ 𝐶( ≤ 𝐶( 𝑗 𝑗=𝑗∗ +1 𝑘∈Λ 𝑗 (ii) For 𝑟 ≥ 1, 𝑝 ∈ [1, 2) and 𝑠 > (2𝜌 + 1)/𝑝, owing to the 𝑠 (𝑀) and (2𝑠 + 2𝜌 + 1)(2 − Markov inequality, 𝑓 ∈ 𝐵𝑝,𝑟 𝑝)/2 + (𝑠 + 1/2 − 1/𝑝 + 𝜌 − 2𝜌/𝑝)𝑝 = 2𝑠, we get 𝑗1 ln 𝑛 ) 𝑛 󵄨 󵄨𝑝 ∑ 󵄨󵄨󵄨󵄨𝑑𝑗,𝑘 󵄨󵄨󵄨󵄨 𝑘∈Λ 𝑗 𝑗 . (79) ≤ 𝐶( ln 𝑛 (2−𝑝)/2 ∞ 𝑗𝜌(2−𝑝) −𝑗(𝑠+1/2−1/𝑝)𝑝 2 ) ∑ 2 𝑛 𝑗=𝑗 +1 (2−𝑝)/2 ≤ 𝐶( ln 𝑛 ) 𝑛 ≤ 𝐶( ln 𝑛 2𝑠/(2𝑠+2𝜌+1) . ) 𝑛 So, for 𝑟 ≥ 1, {𝑝 ≥ 2 and 𝑠 > 0} or {𝑝 ∈ [1, 2), and 𝑠 > (2𝜌 + 1)/𝑝}, we have 𝑄4 ≤ 𝐶( (81) 𝑄 ≤ 𝐶( Let 𝑗∗ be the integer (74). Then 𝑄4 can be bound as (82) where 𝑗∗ 2 1{\|𝑑𝑗,𝑘 \|<2𝜅𝜆 𝑗 } , 𝑄4,1 = ∑ ∑ 𝑑𝑗,𝑘 𝑗=𝑗0 𝑘∈Λ 𝑗 𝑄4,2 = ∑ (83) ∑ 𝑗=𝑗∗ +1 𝑘∈Λ 𝑗 2 𝑑𝑗,𝑘 1{\|𝑑𝑗,𝑘 \|<2𝜅𝜆 𝑗 } . On the one hand, we have 𝑗∗ 𝑄4,1 ≤ 𝐶 ∑ 2𝑗 𝜆2𝑗 = 𝐶 𝑗=𝑗0 𝑗 ln 𝑛 ∗ 𝑗(1+2𝜌) ln 𝑛 2𝑠/(2𝑠+2𝜌+1) ≤ 𝐶( . ) ∑2 𝑛 𝑗=𝑗 𝑛 0 (84) On the other hand, we have the following. ln 𝑛 2𝑠/(2𝑠+2𝜌+1) . ) 𝑛 (87) Putting (70), (72), (80), and (87) together, for 𝑟 ≥ 1, {𝑝 ≥ 2 and 𝑠 > 0} or {𝑝 ∈ [1, 2) and 𝑠 > (2𝜌 + 1)/𝑝}, we obtain 𝑗=𝑗0 𝑘∈Λ 𝑗 𝑗1 (86) 2−𝑗∗ (𝑠+1/2−1/𝑝+𝜌−2𝜌/𝑝)𝑝 (80) 𝑗1 𝑄4 ≤ 𝑄4,1 + 𝑄4,2 , 𝑗 󵄨󵄨 󵄨󵄨𝑝 󵄨󵄨𝑑𝑗,𝑘 󵄨󵄨 󵄨 󵄨 ∗ Upper Bound for 𝑄4 . We have 2 1{\|𝑑𝑗,𝑘 \|<2𝜅𝜆 𝑗 } . 𝑄4 ≤ ∑ ∑ 𝑑𝑗,𝑘 ln 𝑛 (2−𝑝)/2 1 𝑗𝜌(2−𝑝) ) ∑ 2 ∑ 𝑛 𝑗=𝑗 +1 𝑘∈Λ ∗ . 2𝑠/(2𝑠+2𝜌+1) 2−𝑝 𝑄4,2 ≤ 𝐶 ∑ 𝜆 𝑗 = 𝐶( Therefore, for 𝑟 ≥ 1, {𝑝 ≥ 2 and 𝑠 > 0} or {𝑝 ∈ [1, 2), and 𝑠 > (2𝜌 + 1)/𝑝}, we have 𝑄2 ≤ C( (85) 𝑗=𝑗∗ +1 ln 𝑛 (2−𝑝)/2 −𝑗∗ (𝑠+1/2−1/𝑝+𝜌−2𝜌/𝑝)𝑝 2 ) 𝑛 ln 𝑛 ) 𝑛 𝑗=𝑗∗ +1 ln 𝑛 2𝑠/(2𝑠+2𝜌+1) . ) 𝑛 (78) 󵄨󵄨 󵄨󵄨𝑝 󵄨󵄨𝑑𝑗,𝑘 󵄨󵄨 󵄨 󵄨 ln 𝑛 (2−𝑝)/2 ∞ 𝑗𝜌(2−𝑝) −𝑗(𝑠+1/2−1/𝑝)𝑝 2 ) ∑ 2 𝑛 𝑗=𝑗∗ +1 2𝑠/(2𝑠+2𝜌+1) ∞ 2 ≤ 𝐶 ∑ 2−2𝑗𝑠 ≤ 𝐶( ∑ 𝑑𝑗,𝑘 𝑄4,2 ≤ ∑ (ii) For 𝑟 ≥ 1, 𝑝 ∈ [1, 2) and 𝑠 > (2𝜌 + 1)/𝑝, the Markov 𝑠 (𝑀), and (2𝑠 + 2𝜌 + 1)(2 − 𝑝)/2 + inequality, 𝑓 ∈ 𝐵𝑝,𝑟 (𝑠 + 1/2 − 1/𝑝 + 𝜌 − 2𝜌/𝑝)𝑝 = 2𝑠 imply that ln 𝑛 1 2𝜌𝑗 1 ≤𝐶 ∑ 2 𝑝 ∑ 𝑛 𝑗=𝑗∗ +1 𝜆 𝑗 𝑘∈Λ 𝑠 𝑠 (𝑀) ⊆ 𝐵2,∞ (𝑀), (i) For 𝑟 ≥ 1 and 𝑝 ≥ 2, since 𝑓 ∈ 𝐵𝑝,𝑟 we have ln 𝑛 2𝑠/(2𝑠+2𝜌+1) . ) 𝑛 (88) Combining (64), (66), (69), and (88), we complete the proof of Theorem 3. Proof of Theorem 4. The proof of Theorem 4 is a direct application of Theorem 3: under (G1)–(G5), the function 𝑞 defined by (24) satisfies (H1) see ([42], equation (2)) and (H2): (i) see ([42], Lemma 6), (ii) see, ([42], equation (11)) and (iii) see ([30], Proof of Proposition 6.1), with 𝜌 = 𝛿. Proof of Theorem 5. The proof of Theorem 5 is a consequence of Theorem 3: under (J1)–(J4), the function 𝑞 defined by (30) satisfies (H1) and (H2): (i)-(ii) see ([31], Proposition 1) and (iii) see ([52], equation (26)), with 𝜌 = 𝛿. Proof of Theorem 6. Set V(𝑥) = 𝑓(𝑥)𝑔(𝑥). Following the methodology of [49], we have 𝑓̂ (𝑥) − 𝑓 (𝑥) = 𝑆 (𝑥) − 𝑇 (𝑥) , (89) 10 Journal of Probability and Statistics (ii) using the boundedness of 𝑌1 (Ω), then (K4), we have where 2 󵄨 󵄨2 E (󵄨󵄨󵄨󵄨𝑞(𝛾𝑗,𝑘 , 𝑉1 )󵄨󵄨󵄨󵄨 ) = E (𝑌12 (𝛾𝑗,𝑘 (𝑋1 )) ) 𝑆 (𝑥) = 1 (̂V (𝑥) − V (𝑥) 𝑔̂ (𝑥) 2 ≤ 𝐶E ((𝛾𝑗,𝑘 (𝑋1 )) ) (90) +𝑓 (𝑥) (𝑔 (𝑥) − 𝑔̂ (𝑥))) 1{\|𝑔(𝑥)\|≥𝑐 ̂ ∗ /2} 1 1 0 (91) ̂ ̂ It follows from {\|𝑔(𝑥)\| < 𝑐∗ /2} ∩ {\|𝑔(𝑥)\| > 𝑐∗ } ⊆ {\|𝑔(𝑥) − 𝑔(𝑥)\| > 𝑐∗ /2}, (K3), (K4), and the Markov inequality that (iii) using the boundedness of 𝑌1 (Ω) and making the change of variables 𝑦 = 2𝑗 𝑥 − 𝑘, we obtain 󵄨󵄨 󵄨 󵄨󵄨𝑞 (𝛾𝑗,𝑘 , 𝑥)󵄨󵄨󵄨 𝑑𝑥 󵄨 󵄨 𝑉 (Ω) ∫ 1 (92) 1 󵄨 󵄨 \|𝑥\| 𝑑𝑥) (∫ 󵄨󵄨󵄨󵄨𝛾𝑗,𝑘 (𝑥∗ )󵄨󵄨󵄨󵄨 𝑑𝑥∗ ) −𝑀 0 = (∫ The triangular inequality yields 󵄨 󵄨󵄨 ̂ 󵄨󵄨𝑓 (𝑥) − 𝑓 (𝑥)󵄨󵄨󵄨 ≤ 𝐶 (\|̂V (𝑥) − V (𝑥)\| + 󵄨󵄨󵄨󵄨𝑔̂ (𝑥) − 𝑔 (𝑥)󵄨󵄨󵄨󵄨) . (93) 󵄨 󵄨 The elementary inequality (𝑎 + 𝑏)2 ≤ 2(𝑎2 + 𝑏2 ) implies that 󵄩2 󵄩 󵄩2 󵄩 E (󵄩󵄩󵄩󵄩𝑓̂ − 𝑓󵄩󵄩󵄩󵄩2 ) ≤ 𝐶 (E (‖̂V − V‖22 ) + E (󵄩󵄩󵄩𝑔̂ − 𝑔󵄩󵄩󵄩2 )) . 2 ≤ 𝐶 ∫ (𝛾𝑗,𝑘 (𝑥)) 𝑑𝑥 ≤ 𝐶 Using (K3) and the indicator function, we have 󵄨 󵄨 ≤ 𝐶 󵄨󵄨󵄨𝑔̂ (𝑥) − 𝑔 (𝑥)󵄨󵄨󵄨 . \|𝑇 (𝑥)\| ≤ 𝐶1{\|𝑔(𝑥)−𝑔(𝑥)\|>𝑐 ̂ ∗ /2} (97) 0 𝑇 (𝑥) = 𝑓 (𝑥) 1{\|𝑔(𝑥)\|<𝑐 . ̂ ∗ /2} 󵄨 󵄨 \|𝑆 (𝑥)\| ≤ 𝐶 (\|̂V (𝑥) − V (𝑥)\| + 󵄨󵄨󵄨𝑔̂ (𝑥) − 𝑔 (𝑥)󵄨󵄨󵄨) . 2 = 𝐶 ∫ (𝛾𝑗,𝑘 (𝑥)) 𝑔 (𝑥) 𝑑𝑥 (94) 𝑀 (98) 1 󵄨 󵄨 = 𝑀2 ∫ 󵄨󵄨󵄨󵄨𝛾𝑗,𝑘 (𝑥)󵄨󵄨󵄨󵄨 𝑑𝑥 ≤ 𝐶2−𝑗/2 . 0 We conclude by applying Lemma 2 with 𝜌 = 0; (K5) and (K6) imply (F1), and the previous inequality implies (F2). 𝑠 (𝑀) with 𝑟 ≥ 1, {𝑝 ≥ 2 Therefore, assuming that V ∈ 𝐵𝑝,𝑟 and 𝑠 ∈ (0, 𝑁)} or {𝑝 ∈ [1, 2) and 𝑠 ∈ (1/𝑝, 𝑁)}, Theorem 3 proves the existence of a constant 𝐶 > 0 satisfying We now bound this two MISEs via Theorem 3. Upper Bound for the MISE of ̂V. Under (K1)–(K6), the function 𝑞 defined by (38) satisfies the following. (H1) With V instead of 𝑓: since 𝜉1 and 𝑋1 are independent with E(𝜉1 ) = 0, E (𝑞 (𝛾𝑗,𝑘 , 𝑉1 )) = E (𝑌1 𝛾𝑗,𝑘 (𝑋1 )) = E (𝑓 (𝑋1 ) 𝛾𝑗,𝑘 (𝑋1 )) 1 1 0 0 (95) = ∫ 𝑓 (𝑥) 𝛾𝑗,𝑘 (𝑥) 𝑔 (𝑥) 𝑑𝑥 = ∫ V (𝑥) 𝛾𝑗,𝑘 (𝑥) 𝑑𝑥, 󵄨󵄨 󵄨 󵄨󵄨𝑞 (𝛾𝑗,𝑘 , (𝑥, 𝑥∗ ))󵄨󵄨󵄨 󵄨 󵄨 sup (𝑥,𝑥∗ )∈[−𝑀,𝑀]×[0,1] 󵄨󵄨 󵄨 󵄨󵄨𝑥𝛾𝑗,𝑘 (𝑥∗ )󵄨󵄨󵄨 󵄨 󵄨 󵄨 󵄨 ≤ 𝑀 sup 󵄨󵄨󵄨󵄨𝛾𝑗,𝑘 (𝑥∗ )󵄨󵄨󵄨󵄨 ≤ 𝐶2𝑗/2 𝑥 ∈[0,1] ∗ (99) ̂ Under (K1)–(K6), proceeding Upper Bound for the MISE of 𝑔. as the previous point, we show that the function 𝑞 defined by (39) satisfies (H1) with 𝑔 instead of 𝑓 and 𝑋𝑡 instead of 𝑉𝑡 , and(H2): (i)-(ii)-(iii) with 𝜌 = 0. 𝑠 (𝑀) with 𝑟 ≥ 1, {𝑝 ≥ 2 Therefore, assuming that 𝑔 ∈ 𝐵𝑝,𝑟 and 𝑠 ∈ (0, 𝑁)} or {𝑝 ∈ [1, 2), and 𝑠 ∈ (1/𝑝, 𝑁)}, Theorem 3 proves the existance of a constant 𝐶 > 0 satisfying 2𝑠/(2𝑠+1) (i) since 𝑌1 (Ω) is bounded thanks to (K2) and (K3), say \|𝑌1 \| ≤ 𝑀 with 𝑀 > 0, we have = ln 𝑛 2𝑠/(2𝑠+1) . ) 𝑛 ln 𝑛 󵄩 󵄩2 ) E (󵄩󵄩󵄩𝑔̂ − 𝑔󵄩󵄩󵄩2 ) ≤ 𝐶( 𝑛 (H2): (i)-(ii)-(iii) with 𝜌 = 0: sup (𝑥,𝑥∗ )∈𝑉1 (Ω) E (‖̂V − V‖22 ) ≤ 𝐶( . (100) Combining (94), (99), and (100), we end the proof of Theorem 6. Conflict of Interests The author declare that there is no conflict of interests regarding the publication of this paper. (96) Acknowledgment The author is thankful to the reviewers for their comments which have helped in improving the presentation. Journal of Probability and Statistics References [1] P. Doukhan, Mixing. Properties and Examples, vol. 85 of Lecture Notes in Statistics, Springer, New York, NY, USA, 1994. [2] M. Carrasco and X. Chen, “Mixing and moment properties of various GARCH and stochastic volatility models,” Econometric Theory, vol. 18, no. 1, pp. 17–39, 2002. [3] R. C. Bradley, Introduction to Strong Mixing Conditions, Volume 1, 2, 3, Kendrick Press, 2007. [4] P. M. Robinson, “Nonparametric estimators for time series,” Journal of Time Series Analysis, vol. 4, pp. 185–207, 1983. [5] G. G. Roussas, “Nonparametric estimation in mixing sequences of random variables,” Journal of Statistical Planning and Inference, vol. 18, no. 2, pp. 135–149, 1987. [6] G. G. Roussas, “Nonparametric regression estimation under mixing conditions,” Stochastic Processes and their Applications, vol. 36, no. 1, pp. 107–116, 1990. [7] Y. K. Truong and C. J. Stone, “Nonparametric function estimation involving time series,” The Annals of Statistics, vol. 20, pp. 77–97, 1992. [8] L. H. Tran, “Nonparametric function estimation for time series by local average estimators,” The Annals of Statistics, vol. 21, pp. 1040–1057, 1993. [9] E. Masry, “Multivariate local polynomial regression for time series: uniform strong consistency and rates,” Journal of Time Series Analysis, vol. 17, no. 6, pp. 571–599, 1996. [10] E. Masry, “Multivariate regression estimation local polynomial fitting for time series,” Stochastic Processes and their Applications, vol. 65, no. 1, pp. 81–101, 1996. [11] E. Masry and J. Fan, “Local polynomial estimation of regression functions for mixing processes,” Scandinavian Journal of Statistics, vol. 24, no. 2, pp. 165–179, 1997. [12] D. Bosq, Nonparametric Statistics for Stochastic Processes. Estimation and Prediction, vol. 110 of Lecture Notes in Statistics, Springer, New York, NY, USA, 1998. [13] E. Liebscher, “Estimation of the density and the regression function under mixing conditions,” Statistics and Decisions, vol. 19, no. 1, pp. 9–26, 2001. [14] D. L. Donoho and I. M. Johnstone, “Ideal spatial adaptation by wavelet shrinkage,” Biometrika, vol. 81, no. 3, pp. 425–455, 1994. [15] D. L. Donoho and I. M. Johnstone, “Adapting to unknown smoothness via wavelet shrinkage,” Journal of the American Statistical Association, vol. 90, pp. 1200–1224, 1995. [16] D. Donoho, I. Johnstone, G. Kerkyacharian, and D. Picard, “Wavelet shrinkage: asymptopia?” Journal of the Royal Statistical Society B, vol. 57, pp. 301–369, 1995. [17] D. L. Donoho, I. M. Johnstone, G. Kerkyacharian, and D. Picard, “Density estimation by wavelet thresholding,” Annals of Statistics, vol. 24, no. 2, pp. 508–539, 1996. [18] A. Antoniadis, “Wavelets in statistics: a review (with discussion),” Journal of the Italian Statistical Society, vol. 6, pp. 97–144, 1997. [19] W. Hardle, G. Kerkyacharian, D. Picard, and A. Tsybakov, Wavelet, Approximation and Statistical Applications, vol. 129 of Lectures Notes in Statistics, Springer, New York, NY, USA, 1998. [20] F. Leblanc, “Wavelet linear density estimator for a discrete-time stochastic process: 𝐿 𝑝 -losses,” Statistics and Probability Letters, vol. 27, no. 1, pp. 71–84, 1996. [21] K. Tribouley and G. Viennet, “𝐿 𝑝 adaptive density estimation in a alpha-mixing framework,” Annales de l’institut Henri Poincare B, vol. 34, no. 2, pp. 179–208, 1998. 11 [22] E. Masry, “Wavelet-based estimation of multivariate regression functions in besov spaces,” Journal of Nonparametric Statistics, vol. 12, no. 2, pp. 283–308, 2000. [23] P. N. Patil and Y. K. Truong, “Asymptotics for wavelet based estimates of piecewise smooth regression for stationary time series,” Annals of the Institute of Statistical Mathematics, vol. 53, no. 1, pp. 159–178, 2001. [24] H. Doosti, M. Afshari, and H. A. Niroumand, “Wavelets for nonparametric stochastic regression with mixing stochastic process,” Communications in Statistics, vol. 37, no. 3, pp. 373– 385, 2008. [25] H. Doosti and H. A. Niroumand, “Multivariate stochastic regression estimation by wavelet methods for stationary time series,” Pakistan Journal of Statistics, vol. 27, no. 1, pp. 37–46, 2009. [26] H. Doosti, M. S. Islam, Y. P. Chaubey, and P. Góra, “Twodimensional wavelets for nonlinear autoregressive models with an application in dynamical system,” Italian Journal of Pure and Applied Mathematics, no. 27, pp. 39–62, 2010. [27] J. Cai and H. Liang, “Nonlinear wavelet density estimation for truncated and dependent observations,” International Journal of Wavelets, Multiresolution and Information Processing, vol. 9, no. 4, pp. 587–609, 2011. [28] S. Niu and H. Liang, “Nonlinear wavelet estimation of conditional density under left-truncated and 𝛼-mixing assumptions,” International Journal of Wavelets, Multiresolution and Information Processing, vol. 9, no. 6, pp. 989–1023, 2011. [29] F. Benatia and D. Yahia, “Nonlinear wavelet regression function estimator for censored dependent data,” Journal Afrika Statistika, vol. 7, no. 1, pp. 391–411, 2012. [30] C. Chesneau, “On the adaptive wavelet deconvolution of a density for strong mixing sequences,” Journal of the Korean Statistical Society, vol. 41, no. 4, pp. 423–436, 2012. [31] C. Chesneau, “Wavelet estimation of a density in a GARCHtype model,” Communications in Statistics, vol. 42, no. 1, pp. 98– 117, 2013. [32] C. Chesneau, “On the adaptive wavelet estimation of a multidimensional regression function under alpha-mixing dependence: beyond the standard assumptions on the noise,” Commentationes Mathematicae Universitatis Carolinae, vol. 4, pp. 527–556, 2013. [33] Y. P. Chaubey and E. Shirazi, “On MISE of a nonlinear wavelet estimator of the regression function based on biased data under strong mixing ,” Communications in Statistics. In press. [34] M. Abbaszadeh and M. Emadi, “Wavelet density estimation and statistical evidences role for a garch model in the weighted distribution,” Applied Mathematics, vol. 4, no. 2, pp. 10–416, 2013. [35] B. Delyon and A. Juditsky, “On minimax wavelet estimators,” Applied and Computational Harmonic Analysis, vol. 3, no. 3, pp. 215–228, 1996. [36] G. Kerkyacharian and D. Picard, “Thresholding algorithms, maxisets and well-concentrated bases,” Test, vol. 9, no. 2, pp. 283–344, 2000. [37] I. Daubechies, Ten Lectures on Wavelets, SIAM, 1992. [38] A. Cohen, I. Daubechies, and P. Vial, “Wavelets on the interval and fast wavelet transforms,” Applied and Computational Harmonic Analysis, vol. 1, no. 1, pp. 54–81, 1993. [39] S. Mallat, A Wavelet Tour of Signal Processing, The sparse way, with contributions from Gabriel Peyralphae, Elsevier/Academic Press, Amsterdam, The Netherlands, 3rd edition, 2009. 12 [40] Y. Meyer, Ondelettes et Opalphaerateurs, Hermann, Paris, France, 1990. [41] V. Genon-Catalot, T. Jeantheau, and C. Larédo, “Stochastic volatility models as hidden Markov models and statistical applications,” Bernoulli, vol. 6, no. 6, pp. 1051–1079, 2000. [42] J. Fan and J. Koo, “Wavelet deconvolution,” IEEE Transactions on Information Theory, vol. 48, no. 3, pp. 734–747, 2002. [43] A. B. Tsybakov, Introduction alphaa l’estimation nonparamalphaetrique, Springer, New York, NY, USA, 2004. [44] E. Masry, “Strong consistency and rates for deconvolution of multivariate densities of stationary processes,” Stochastic Processes and their Applications, vol. 47, no. 1, pp. 53–74, 1993. [45] R. Kulik, “Nonparametric deconvolution problem for dependent sequences,” Electronic Journal of Statistics, vol. 2, pp. 722– 740, 2008. [46] F. Comte, J. Dedecker, and M. L. Taupin, “Adaptive density deconvolution with dependent inputs,” Mathematical Methods of Statistics, vol. 17, no. 2, pp. 87–112, 2008. [47] H. van Zanten and P. Zareba, “A note on wavelet density deconvolution for weakly dependent data,” Statistical Inference for Stochastic Processes, vol. 11, no. 2, pp. 207–219, 2008. [48] W. Hardle, Applied Nonparametric Regression, Cambridge University Press, Cambridge, UK, 1990. [49] V. A. Vasiliev, “One investigation method of a ratios type estimators,” in Proceedings of the 16th IFAC Symposium on System Identialphacation, pp. 1–6, Brussels, Belgium, July 2012, (in progress). [50] Y. Davydov, “The invariance principle for stationary processes,” Theory of Probability and Its Applications, vol. 15, no. 3, pp. 498– 509, 1970. [51] C. Chesneau, “Wavelet estimation via block thresholding: a minimax study under 𝐿𝑝 -risk,” Statistica Sinica, vol. 18, no. 3, pp. 1007–1024, 2008. [52] C. Chesneau and H. Doosti, “Wavelet linear density estimation for a GARCH model under various dependence structures,” Journal of the Iranian Statistical Society, vol. 11, no. 1, pp. 1–21, 2012. Journal of Probability and Statistics Advances in Operations Research Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Advances in Decision Sciences Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Journal of Applied Mathematics Algebra Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Journal of Probability and Statistics Volume 2014 The Scientific World Journal Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 International Journal of Differential Equations Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014 Submit your manuscripts at http://www.hindawi.com International Journal of Advances in Combinatorics Hindawi Publishing Corporation http://www.hindawi.com Mathematical Physics Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Journal of Complex Analysis Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 International Journal of Mathematics and Mathematical Sciences Mathematical Problems in Engineering Journal of Mathematics Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Discrete Mathematics Journal of Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Discrete Dynamics in Nature and Society Journal of Function Spaces Hindawi Publishing Corporation http://www.hindawi.com Abstract and Applied Analysis Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 International Journal of Journal of Stochastic Analysis Optimization Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014
https://openalex.org/W3041460430	https://europepmc.org/articles/pmc7353807?pdf=render	English	null	High multiple mutations of Plasmodium falciparum-resistant genotypes to sulphadoxine-pyrimethamine in Lagos, Nigeria	Infectious diseases of poverty	2,020	cc-by	8,903	(2020) 9:91 (2020) 9:91 Quan et al. Infectious Diseases of Poverty (2020) 9:91 https://doi.org/10.1186/s40249-020-00712-4 Open Access High multiple mutations of Plasmodium falciparum-resistant genotypes to sulphadoxine-pyrimethamine in Lagos, Nigeria Hong Quan1,2, Uche Igbasi3, Wellington Oyibo4, Sunday Omilabu5, Shen-Bo Chen1,2, Hai-Mo Shen1,2, Chukwuma Okolie6, Jun-Hu Chen1,2* and Xiao-Nong Zhou1,2 * Correspondence: chenjh@nipd.chinacdc.cn * Correspondence: chenjh@nipd.chinacdc.cn 1National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention, Chinese Center for Tropical Diseases Research, WHO Collaborating Center for Tropical Diseases, National Centre for International Research on Tropical Diseases, Ministry of Science and Technology, Key Laboratory of Parasite and Vector Biology, Ministry of Health, Shanghai 200025, People’s Republic of China a d e e t o , C ese Ce te o op ca seases esea c , O Collaborating Center for Tropical Diseases, National Centre for International Research on Tropical Diseases, Ministry of Science and Technology, Key Laboratory of Parasite and Vector Biology, Ministry of Health, Shanghai 200025, People’s Republic of China 2National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention⁃Shenzhen Center for Disease Control and Prevention Joint Laboratory for Imported Tropical Disease Control, Shanghai 200025, People’s Republic of China Full list of author information is available at the end of the article Abstract Background: Plasmodium falciparum-resistance to sulphadoxine-pyrimethamine (SP) has been largely reported among pregnant women. However, the profile of resistance markers to SP dihydrofolate reductase (dhfr) and dihydropteroate synthase (dhps) in the general population are varied and not frequently monitored. Currently, SP is used as partner drug for artemisinin combination therapy (SP-artesunate) in some sub-Saharan African countries or as a prophylactic drug in intermittent preventive treatment of malaria during pregnancy and infants and in seasonal malaria chemoprevention (SMC). Profiling of P. falciparum-resistant genotypes to SP is dynamic and critical in providing data that would be useful for malaria control programmes. This study assessed the profile of dhfr and dhps genes genotypes among individuals with malaria in Lagos, Nigeria. Methods: Molecular markers of SP resistance were identified by nested PCR and sequenced among malaria positive dried blood spots (DBS) that were collected from individuals attending health facilities from January 2013 to February 2014 and during community surveys from October 2010 to September 2011 across different Local Government Areas of Lagos State, Nigeria. (Continued on next page) Background 96.9% prevalence of SP-resistant mutations in the last decade [8–10]. The efficacy of antimalarial medicines is critical to the implementation of effective malaria case management where patients confirmed to have malaria parasites are treated promptly. Consequently, failing antimalaria med- icines due to parasite resistance will greatly affect the at- tainment of the case management goal. Resistance to antimalarial drugs has been described for Plasmodium falciparum, the predominant Plasmodium species in Africa [1]. Over a decade, sulphadoxine-pyrimethamine (SP) was the second-line treatment medicine while chloroquine (CQ) served as the first-line antimalarial medicine for the treatment of uncomplicated P. falcip- arum malaria [2]. P. falciparum, unfortunately developed resistance to both widely used medicines and are not currently recommended for the treatment of malaria as monotherapies in the general population. The malaria parasite’s resistance to SP is due to point mutations in target enzymes, dihydrofolate reductase (dhfr) and dihy- dropteroate synthase (dhps) [3]. Resistance to SP and CQ were reported at different times in the history of anti-malarial medicine resistance [4, 5]. SP acts primarily on the schizonts during the hepatic and erythrocytic phases of the plasmodia life cycle [11], by inhibiting enzymes necessary for parasite folate biosyn- thesis. Pyrimethamine acts by inhibiting dhfr in the para- site [12], thus preventing the biosynthesis of purines and pyrimidines, while sulphadoxine binds the enzyme dhps [13], inhibiting the use of para-aminobenzoic acid during the synthesis of dihydropteroic acid. When combined the two key stages in DNA synthesis in the plasmodia are pre- vented consequently, cell division and reproduction are halted. As these two drugs target the same pathway and act synergistically, they are usually given in combination as SP but referred to as monotherapy [14]. Mutations in the dhfr and dhps genes of P. falciparum parasites have been associated with decreased parasite sensitivity to the anti-folate drugs. In both genes, each successive mutation has been shown to incrementally in- crease the parasite’s tolerance to the drug in vitro [14]. A change from wild type serine 108 to asparagine108 (S108N) in dhfr is sufficient to cause low level pyrimeth- amine resistance both in vitro and in vivo [15]; this repre- sents the initial and critical mutation for pyrimethamine. Additional mutation(s) at positions 50: C50R, 51: N51I, 59: C59R and I164L synergistically increase the levels of resistance [16, 17]. © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Page 2 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 (2020) 9:91 (Continued from previous page) p p g Results: A total of 242 and 167 samples were sequenced for dhfr and dhps, respectively. Sequence analysis of dhfr showed that 95.5% (231/242), 96.3% (233/242) and 96.7% (234/242) of the samples had N51I, C59R and S108N mutant alleles, respectively. The prevalence of dhps mutation at codons A437G, A613S, S436A, A581G, I431V and K540E were 95.8% (160/167), 41.9% (70/167), 41.3% (69/167), 31.1% (52/167), 25.1% (42/167), and 1.2% (2/167) respectively. The prevalence of triple mutations (CIRNI) in dhfr was 93.8% and 44.3% for the single dhps haplotype mutation (SGKAA). Partial SP-resistance due to quadruple dhfr-dhps haplotype mutations (CIRNI-SGKAA) and octuple haplotype mutations (CIRNI-VAGKGS) with rate of 42.6% and 22.0%, respectively has been reported. Conclusions: There was increased prevalence in dhfr triple haplotype mutations when compared with previous reports in the same environment but aligned with high prevalence in other locations in Nigeria and other countries in Africa. Also, high prevalence of dhfr and dhps mutant alleles occurred in the study areas in Lagos, Nigeria five to eight years after the introduction of artemisinin combination therapy underscores the need for continuous monitoring. Keywords: Plasmodium falciparum, Antimalarial drug resistance, Sulphadoxine–pyrimethamine, Dihydrofolate reductase, Dihydropteroate synthase Background Currently, SP is used as a partner drug for antimalarial drug resistance (ACT, such as SP-artesunate), seasonal malaria chemoprevention (SMC) in areas where it is rec- ommended, intermittent preventive treatment of malaria in infants and children (IPTi & c) in some sub-Saharan African countries, and as intermittent preventive treat- ment of malaria in pregnancy (IPTp) [6]. The spread of SP resistance may compromise the effectiveness of inter- mittent preventive treatment of malaria in pregnancy (MiP) with SP (IPTp-SP) and other interventions includ- ing SMC across Africa. In West Africa, SP resistant genes of P. falciparum have been reported, and IPTp-SP remains the interventional strategy for the prevention of malaria in pregnancy [7]. Studies in Nigeria have re- ported varying mutant combinations with over 50.0– Furthermore, mutation from alanine to glycine at codon 437 (A437G) for dhps represents the critical mu- tation for sulphadoxine resistance and additional muta- tion(s) at positions 436 (S436A/F), 540 (K540E), 581 (A581G), and 613 (A613S/T) have been associated with decreased parasite sensitivity to the sulpha drugs includ- ing sulphadoxine and dapsone [13, 18, 19]. Mutations at codons 437 and 540 of dhps play the most significant role in sulphadoxine resistance among African parasites. In East and South Africa, mutations at the 437 and 540 Page 3 of 11 Page 3 of 11 Page 3 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Quan et al. Infectious Diseases of Poverty codons are found together while in West and Central Africa the 437 is found on its own [20]. However, muta- tion at codon 431 (I431V) has been scarcely reported. It was first reported among imported malaria infections that originated from Nigeria in 2009 [21] and pregnant women from Cameroon in 2015 [22, 23], though its ef- fect on parasite susceptibility to SP is yet to be fully de- scribed [10]. (Fig. 1) and has an estimated population of over 10 mil- lion inhabitants, which is more than 10% of the total population of Nigeria. The state is a low-lying coastal State and Nigeria’s centre of commerce, accounting for more than 70% of the nation’s industrial and commercial establishments. Lagos is a centre of commerce with very diverse and fast-growing population, with high migration to its cities from all parts of Nigeria as well as neigh- bouring and foreign countries. Study population and sites The samples used for this study were obtained from chil- dren and adult patients that presented with fever/symp- toms of malaria in the last 48 h in a cross-sectional case management study in health facilities (January 2013 to February 2014) and from asymptomatic persons in com- munity surveys (October 2010 to September 2011). The health facilities were: Ijede General Hospital (H1), Imota Primary Health Centre (H2), Bayioku Primary Health Centre (H3), Agura Primary Health Centre (H4). These four health centres are situated in Ikorodu LGA. St. Kizito Primary Health Centre, Lekki in Ibeju Lekki LGA (H5), St Mathew Primary Health Centre (Catholic), Ajegunle, Amukoko in Ajeromi Ifelodu LGA (H6), and in Shomolu General Hospital, Shomolu LGA (H7) (Fig. 1). The asymp- tomatic study was conducted in persons aged 2 months and above in communities that were randomly and pur- posively selected based on malaria endemicity data, and enrolment was done using multi-stage and stratified sam- pling in households. The communities were: 1) Ipakodo in Ikorodu LGA, 2) Agbowa, 3) Ajagunmagbe, 4) Noforija in Epe LGA, 5) Lekki in Ibeju Lekki LGA, 6) Maidian Com- munity, 7) Owode/Ajegunle, 8) Owode/Elede in Kosofe LGA, 9) Badagry town, 10) Apa in Badagry LGA, 11) Sura in Lagos Island LGA, 12) Ijora, 13) Abete, 14) Akere, 15) Alayabiagba in Apapa LGA, 16) Otto in Lagos Mainland LGA, 17) Monkey Village, 18) Ijegun in Amuwo Odofin LGA, 19) Aburo in Alimosho LGA, 20) Tedi-Muwo in Ojo LGA (Fig. 1). SP-resistant parasites could be classified as “partially resistant”, “fully resistant” and “super resistant” [28]. The parasites are classified based on the combination of mu- tations they carry in the two genes (dhfr and dhps). The quadruple combination of triple mutation, dhfr N51I, C59R, S108N and dhps A437G, confers partial resist- ance; the quintuple combination of triple mutations, dhfr N51I, C59R, S108N and double mutation, dhps A437G, K540E, confers full resistance; and the sextuple combin- ation of triple mutation, dhfr N51I, C59R, S108N and triple mutation, dhps A437G, K540E, A581G, confers super resistance [28]. These haplotype mutations affect the outcome of IPTp and IPTi [28] .. Molecular genotyping and characterization of single nucleotide polymorphisms (SNPs) used in drug resist- ance monitoring could provide red flags of threats to continued use of SP in strategies planned by countries. Methods y The study was conducted in Lagos State, Nigeria. Lagos State is an African megacity located in south-western Nigeria on the west coast of Africa, within latitudes 6023′N and 6041′N and longitudes 2042′E and 3042′E Background It has been demonstrated that the accumulation of sin- gle nucleotide polymorphisms (SNPs) in dhfr and dhps genes increases the levels of SP resistance in vivo [23]. In West and Central Africa, a triple mutant genotype of dhfr (N51I, C59R and S108N) combined with the A437G mutation in the dhps gene has been related to SP treatment failure [24]. Another significant predictor of SP treatment failure is the quintuple mutant geno- type, which includes the dhfr triple mutations (N51I, C59R and S108N) combined with the dhps double muta- tions (A437G + K540E) [25–27]. There are 20 Local Government Areas (LGAs) from where 50 Local Council Development Areas (LCDAs) were carved for ease of administration. The land surface generally slopes gently downwards from north to south and is naturally made up of depositional landforms which include: wetlands, barrier islands, beaches, low- lying tidal flats and estuaries [29]. Furthermore, Lagos is hypo-endemic for malaria during the dry season with moderate and stable transmission but peaks during the wet season due to increase in the population of mosqui- toes [30]. In Nigeria, high prevalence of triple mutant genotype of dhfr (N51I, C59R and S108N) combined with A437G mutation in the dhps gene have been reported [10], but reports of quintuple dhfr/dhps mutation comprising of (N51I, C59R and S108N) plus (A437G + K540E) is scarce [10]. Study population and sites This study provides data for trend profiling of molecular markers of resistance to antifolate drugs from isolates of Plasmodium falciparum from stored patients’ DBS ob- tained between 2010 and 2014 in Lagos, South-West, Nigeria. Preparation of dry blood spots Dry blood spots were prepared from approximately 3–4 drops of blood collected from the study participants Quan et al. Infectious Diseases of Poverty (2020) 9:91 Page 4 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Fig. 1 Location of health facilities and communities that were studied in Lagos State, Nigeria spot was placed into a 1.5 ml micro centrifuge tube. The genomic DNA from the DBSs was extracted using the QIAamp® DNA Mini kit (Qiagen, Germany) according to the manufacturer’s instructions. The DNA extracted was finally eluted using 50 μl elution buffer and kept at −20 °C. (both in the health facility and community) on filter paper (Whatman® filter paper #3, Whatman Inter- national Ltd., Maidstone, England). Thick and thin mal- aria blood films (MBFs) for malaria microscopy were also prepared for each person. The prepared blood spots were air-dried and kept in a zip-lock bag with desiccant and stored at 2–8 °C for molecular analysis. Nested PCR was used to amplify the extracted DNA. The primary and secondary amplification were done in a 25 μl reaction mixture that comprised of 2 μl of template genomic DNA, 1 μl of primer F, 1 μl of primer R, 12.5 μl of Taq 2x DNA master mix (Sangon Bio Inc., Shanghai, China) and 8.5 μl of double distilled water (ddH2O) using specific primers and cycling condition for the dif- ferent genes. Malaria microscopy Malaria microscopy on the prepared MBFs was done using standard protocol and read by independent micro- scopist to identify malaria positive smears [31]. The re- sults of the malaria microscopy were used in the selection of DBS for the molecular profiling of P. falcip- arum-resistant genes and haplotypes to SP. The dhfr genes were amplified using the primer as previously described [32]; Forward: 5′-TCCTTTTTAT GATGGAACAAG-3′, Reverse: 5′-AGTATATACA TCGCTAACAGA-3′, and cycling conditions; initial de- naturation for 5 min at 94 °C; 35 cycles at 95 °C for 30 s, 50 °C for 30 s and 68 °C for 1 min, finial extension at 68 °C for 5 min. 2 μl of the PCR product was used in the second round amplification with specified primers; For- ward: 5′-TTTATGATGGAACAAGTCTGC-3′, Reverse: DNA extraction and PCR amplification A subset of 404 dB from the samples positive for P. fal- ciparum mono infection was used for this study. Three milliliter diameter punches were made from the DBS with single-hole paper puncher. Sterilization of the puncher was done after every punch of each patient’s DBS with 70% soaked alcohol swab. The punched blood Page 5 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Quan et al. Infectious Diseases of Poverty 5′-ACTCATTTTCATTTATTTCTGG-3′, and cycling conditions; 5 min at 94 °C, 30 cycles at 94 °C for 30 s, 52 °C for 30 s and 68 °C for 1 min, finial extension at 68 °C for 5 min. obtained from symptomatic and asymptomatic individ- uals respectively from the study areas in Lagos, South – West, Nigeria between 2010 and 2014. The age of the participants were two months and above (range: 2 months–65 years; mean ± SD, 16.8 ± 14.1 years) (Table 1). The asexual parasitaemia of the study individuals ranged from 63 to 202 010 parasites/μl of blood (geometric mean parasite density of 7615 parasites/μl of blood). The dhps genes were amplified using the primer as previously described [33]; Forward: 5′-AACCTAAACG TGCTGCTGTTCAA-3′, Reverse: 5′-AATTGTGTGA TTTGTCCACAA-3′, and cycling conditions; initial de- naturation for 5 min at 95 °C, 35 cycles at 95 °C for 30 s, 50 °C for 30 s and 68 °C for 1 min, finial extension at 68 °C for 5 min. 2 μl of the PCR product was used in the second round amplification with specified primers; For- ward: 5′-ATGATAAATGAAGGTGCTAG-3′, Reverse: 5′-TCATTTTGTTGTTCATCATGT-3′, and cycling conditions; 5 min at 95 °C, 30 cycles at 95 °C for 30 s, 52 °C for 30 s and 68 °C for 1 min, finial extension at 68 °C for 5 min. The expected PCR product is 647 bp. Prevalence of individual point mutations in dhfr and dhps The sequence analyses of dhfr at each codon showed that the 51I mutations appeared in 95.5% of the P. fal- ciparum isolates. The prevalence of 59R and 108 N mu- tations was 96.3% and 96.7% respectively. However, the prevalence of P. falciparum isolates with wild dhfr was low and ranged from 3.3% (108 N) to 4.5% (51I) (Table 2). Positive controls were obtained from BEI Resources, USA (parasite Genomic DNA IPC 3663/3D7 strains and Dd2_R539T/Dd2 strains) and were used as references for susceptible and resistant genotypes, respectively. DNA extraction and PCR amplification Nuclease-free water was used as a negative control. Furthermore, the sequence analyses of dhps showed that the most prevalent mutation of the cohort of indi- viduals examined was 437G (95.8%). Other mutations were: A613S (41.9%), S436A (41.3%), A581G (34.1%), I431V (25.1%) and K540E (1.2%) (Table 2). The genetic data from this study is deposited at the National Center for Biotechnology Information (NCBI). Accession num- bers- BankIt2291211: MN985140–MN985306 (167 se- quences) and BankIt2298803: MT140637–MT140882 (241 Sequences). g The nested PCR products for dhfr and dhps were loaded on a 1% agarose gel containing 0.5 μg/ml ethidium brom- ide. Amplified bands of DNA were visualized under ultra- violet illumination and positive samples were selected for sequencing. The amplicon for the different genes were se- quenced on an Applied Biosystem 3130 x l Genetic analyzer (Applied Biosystems, Foster City, CA, USA). Se- quence alignment was done with DNASTAR 7.1 software and analyzed using the reference: Plasmodium falciparum 3D7 dhps sensitive strain (NCBI reference Sequence: XM_ 001349382.1) and dhfr sensitive strain (NCBI Reference Sequence: XM_001351443.1), respectively. The frequency of dhfr and dhps mutations in the health facilities and communities occurred in varied pro- portions (Tables 3, 4, 5, 6). The health facility situated in Ijede, Ikorodu LGA had the highest number of samples and as well as the highest frequency of dhfr gene muta- tion at codons 51I, 59R and 108 N (Table 3) and dhps gene mutation at codons 431 V, 436A, 437G, 581G and 613S (Table 4). Analysis of dhfr gene mutation among P falciparum isolates from asymptomatic individuals from different communities showed that Lekki located in Ethical considerations The participants (in the health facility and community) gave written consent to participate and for their blood samples to be used for further malaria testing. Partici- pants who presented at the screening for malaria that did not agree to participate were also attended to and standard care provided appropriately. All samples had only study identification numbers that could not be linked with personal details of the participants. The study protocol was approved by the Research Grants and Experimentation Ethics Committee, College of Medicine, University of Lagos, Nigeria and the Ethics Committee of the National Institute of Parasitic Diseases (NIPD), China. Table 1 Demographic characteristics of the cohort of individuals with positive dry blood spots that were used in the study in Lagos, Nigeria Table 1 Demographic characteristics of the cohort of individuals with positive dry blood spots that were used in the study in Lagos, Nigeria Description n = 338 Sex, n (%) Male 152 (45.0) Female 186 (55.0) Age category, n (%) ≤5 years 68 (20.1) 6–10 years 116 (34.3) > 10 years 154 (45.6) Parasite density (parasites/μl of blood) Range 63–202 010 Geometric mean 7615 Prevalence of dhfr and dhps haplotypes Seven and seventeen distinct haplotypes occurred in dhfr and dhps, respectively in the cohorts of individuals in the study areas in Lagos (Table 7). Dhfr and dhps allele combinations A total of 19 haplotypes were seen in the dhfr and dhps combinations (Table 8). Quadruple mutations occurred in the combined dhfr (triple mutations) + dhps (single muta- tion) haplotype (CIRNI + SGKAA) and was the most prevalent (42.6%, 60/141). This was followed by octuple mutations in CIRNI + VAGKGS (22.0%, 31/141). Sextuple mutations (CIRNI + AGKAS) and quintuple mutations (CIRNI + AGKAA) among the isolates were 9.9% and 5.0% respectively. Mutations in the haplotypes of the other combined genotypes were generally low (Table 8). SNPs Single nucleotide polymorphisms interval. Ibeju-Lekki LGA had the highest frequency of mutations at codons 51I, 59R and 108 N (Table 5) and dhps muta- tions at codons 431, 436, 437, 540, 581 and 613 (Table 6). Generally, there was no considerable difference in the occurrence of dhfr and dhps mutation across the sites in Lagos State. Demographic characteristics Point mutations in dhps and dhfr were evaluated in 242 and 167 out of a cohort of 338 malaria-positive DBS Quan et al. Infectious Diseases of Poverty (2020) 9:91 Page 6 of 11 Table 2 Prevalence of dhfr and dhps SNPs among Plasmodium falciparum isolates from Lagos, Nigeria Gene SNPs Wild type n (%) Mutation n (%) dhfr (n = 242) 51 11 (4.5) 231 (95.5) 59 9 (3.7) 233 (96.3) 108 8 (3.3) 234 (96.7) 83 241 (99.6) 1 (0.4) 122 241 (99.6) 1 (0.4) 160 241 (99.6) 1 (0.4) dhps (n = 167) 431 125 (74.9) 42 (25.1) 436 95 (56.9) 69 (41.3) 437 7 (4.2) 160 (95.8) 540 165 (98.8) 2 (1.2) 581 115 (68.9) 52 (31.1) 613 97 (58.1) 70 (41.9) SNPs Single nucleotide polymorphisms interval. Table 2 Prevalence of dhfr and dhps SNPs among Plasmodium falciparum isolates from Lagos, Nigeria positions: 83R, 122 K and 160E together with the triple CIRNI mutations (Table 7). Discussion P. falciparum drug resistance remains a challenge to ef- fective malaria case management and prevention. This has made continuous monitoring of molecular markers of antimalarial drug resistance imperative in malaria- endemic countries to track trends and distribution of relevant resistant genes and haplotypes to ensure that threats to existing artemisinin combination therapies and drug-dependent interventions are identified and ad- dressed promptly. Information on these threats will also further guide National Malaria Control Programmes to adopt the most suitable interventions using the appro- priate drug combinations. Our study showed high fre- quencies of P. falciparum isolates with mutant dhfr and dhps in circulation in Lagos, Nigeria. LGA Local Government Area. Dhps haplotypes Mutations in the dhps haplotype occurred in different proportions: single mutation in SGKAA was 44.3% (74/ 167). Quintuple mutations in VAGKGS (19.8%, 33/167) was the most prevalent of the multiple mutations com- pared to the triple, double and quadruple mutations that were recorded (Table 7). Dhfr haplotypes Mutations in dhps haplotypes at 437 occurred in 95.8%, 31.1% and 1.2% at codons 581 and 540 respect- ively in the samples analyzed. Amino acid changes at position 437 (A437G) represented the critical mutation for sulphadoxine resistance. Additional mutation(s) at positions 436 (S436A/F), 540 (K540E), 581 (A581G), and 613 (A613S/T) are associated with decreased parasite sensitivity to the sulpha drugs including sulphadoxine and dapsone [13, 18, 19]. Two (1.2%) dhps double haplo- type mutation consisting of A437G and K540E were seen and have been consistently associated with in vivo clinical failure independently [26, 38]. Similarly, 581G dhps haplotype mutation has also been shown to be as- sociated with important modulatory role in resistance [39] .The World Health Organization (WHO) recom- mends that when the frequency of this mutation is above 10.0%, IPTp with SP may not be able to protect pregnant women from delivering infants with low birth weight [40]. The 540E and 581G haplotype mutations have also comparable with some other reported studies among pregnant women in Nigeria and in Sub Saharan Af- rica [9, 10, 34]. These reported studies indicating the authors, the mutation prevalance and the time of sample collection in Nigeria included: Agomo et al., 66.7% in Lagos in 2008/2009 [8], Iwalokun et al., 50.0% (Lagos) (2011) [9], and Oguike et al., 100.0% (Ibadan) (2003), 81.3% (Maiduguri) (2010), 90.2% (Enugu) (2010) and 98.7% in Benin city (2014/2015) [10]; and in Guinea, Jiang et al., 86.8% (Bioko Island) (2013/2014) [34]. There was an increase in dhfr triple haplotype mutations in Lagos within about 2 years. This mutation is associated with high-level resistance to pyrimethamine [35–37] and increased risk of SP resistance if it occurreds concurrently with dhps mu- tations [26, 27, 36, 37]. Mutations such as 16 V+ 164 L in dhfr that are associated with high resistance to cycloguanil, the active form of proguanil [37, 38] were not observed in our study in Lagos, Nigeria. Table 5 Distribution of dhfr mutations at the community locations in Lagos, Nigeria LGA/ Community Name of community No. Dhfr haplotypes Triple mutations in the dhfr haplotype (CIRNI) was the most prevalent (93.8%, 227/242) while 3.3% (8/242) of the isolates were wild type haplotype (CNCSI). The prevalence of double mutations in CICNI and CNRNI was 0.4% (1/242) and 1.2% (3/242) respectively while 1.2% of the samples had quadruple mutations at The dhfr triple haplotype mutation (CIRNI) was highly prevalent at all study sites in our study, which was Table 3 Distribution of dhfr mutations at the health facilities in Lagos, Nigeria LGA/ Health facility Location of the health facilities No. of samples sequenced for dhfr n (%) dhfr mutation n (%) Wild N51I C59R S108N Ikorudu Ijede [H1] 72 (34.0) 3 (1.4) 69 (32.5) 69 (32.5) 69 (32.5) Imota [H2] 27 (12.7) 1 (0.5) 26 (12.3) 26 (12.3) 26 (12.3) Bayioku [H3] 7 (3.3) 0 7 (3.3) 7 (3.3) 7 (3.3) Agura [H4] 52 (24.5) 3 (1.4) 49 (23.1) 48 (22.6) 49 (23.1) Ajeromi -Ifelodun Amukoko [H6] 35 (16.5) 0 35 (16.5) 33(15.6) 35 (16.5) Shomolu Shomolu [H7] 19 (9.0) 0 19 (9.0) 19 (9.0) 19 (9.0) Total 212 (100) 7 (3.3) 205(96.7) 202(95.3) 205(96.7) LGA Local Government Area. Table 3 Distribution of dhfr mutations at the health facilities in Lagos, Nigeria Page 7 of 11 Page 7 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Table 4 Distribution of dhps mutations at the health facilitie in Lagos, Nigeria LGA/ Health facility Location of health facilities No. of samples sequenced for dhps n (%) dhps mutation n (%) I431V A436S A437G K540E A581G A613S Ikorudu Ijede [H1] 50 (34.0) 12 (8.2) 28 (19.0) 45 (30.6) 0 14 (9.5) 20 (13.6) Imota [H2] 17 (11.6) 5 (3.4) 9 (6.1) 16 (10.9) 0 5 (3.4) 8 (5.4) Bayioku [H3] 5 (3.4) 0 4 (2.7) 5 (3.4) 0 1 (0.7) 1 (0.7) Agura [H4] 26 (17.7) 5 (3.4) 16 (10.9) 24 (16.3) 0 7 (4.8) 13 (8.8) Ajeromi- Ifelodun Amukoko [H6] 31 (21.1) 9 (6.1) 15 (10.2) 28 (19.0) 0 11 (7.5) 15 (10.2) Shomolu Shomolu [H7] 18 (12.2) 2 (1.4) 14 (9.5) 18 (12.2) 2 (1.4) 4 (2.7) 5 (3.4) Total 147 (100) 33 (22.4) 83 (56.5) 136 (92.5) 2 (1.4) 42 (28.6) 62 (42.2) LGA Local Government Area. Table 4 Distribution of dhps mutations at the health facilitie in Lagos, Nigeria LGA Local Government Area. LGA Local Government Area. Dhfr haplotypes of samples sequenced for dhfr n (%) dhfr mutation n (%) Wild N51I C59R S108N Ikorodu Ipakodo [1] 2 (6.7) 0 2 (6.7) 2 (6.7) 2 (6.7) Epe Agbowa [2] 5 (16.7) 0 5 (16.7) 5 (16.7) 5 (16.7) Noforija [4] 1 (3.3) 0 1 (3.3) 1 (3.3) 1 (3.3) Ibeju Lekki Lekki [5] 10 (33.3) 1 (3.3) 9 (30.0) 9 (30.0) 9 (30.0) Kosofe Madian Community [6] 1 (3.3) 0 1 (3.3) 1 (3.3) 1 (3.3) Owode/Ajegunle [7] 3 (10.0) 0 3 (10.0) 3 (10.0) 3 (10.0) Owode/Elede [8] 1 (3.3) 0 1 (3.3) 1 (3.3) 1 (3.3) Lagos Island Sura [11] 1 (3.3) 0 1 (3.3) 1 (3.3) 1 (3.3) Apapa Ijora [12] 1 (3.3) 0 1 (3.3) 1 (3.3) 1 (3.3) Akere [14] 1(3.3) 0 1 (3.3) 1 (3.3) 1 (3.3) Amowo Odofin Monkey Village [17] 0 1 (3.3) 1 (3.3) 1 (3.3) Ijegun [18] 0 3 (10.0) 3 (10.0) 3 (10.0) Total 30 (100) 1 (3.3) 29 (96.7) 29 (96.7) 29 (96.7) LGA Local Government Area. Table 5 Distribution of dhfr mutations at the community locations in Lagos, Nigeria Page 8 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Table 6 Distribution of dhps mutations at the community locations in Lagos, Nigeria LGA/ Community Name of community No. of samples sequenced for dhps n (%) dhps mutation n (%) I431V A436S A437G K540E A581G A613S Ikorudu Ipakodo [1] 1 (5.0) 0 0 1 (5.0) 0 0 1 (5.0) Epe Agbowa [2] 3 (15.0) 2 (10.0) 1 (5.0) 3 (15.0) 0 2 (10.0) 2 (10.0) Ibeju- Lekki Lekki [5] 9 (45.0) 2 (10.0) 4 (20.0) 9 (45.0) 0 2 (10.0) 5 (25.0) Kosofe Owode/Ajegunle [7] 1 (5.0) 0 1 (5.0) 1 (5.0) 0 0 0 Owode /Elede [8] 1 (5.0) 0 1 (5.0) 1 (5.0) 0 0 0 Apapa Akere [14] 1 (5.0) 0 1 (5.0) 1 (5.0) 0 0 1 (5.0) Alayabiagba [15] 1 (5.0) 0 1 (5.0) 1 (5.0) 0 0 0 Amuwo- Odofin Monkey village [17] 2 (10.0) 0 2 (10.0) 2 (10.0) 0 0 0 Ijegun [18] 1 (5.0) 0 1 (5.0) 1 (5.0) 0 0 0 Total 20 (100) 4 (20.0) 11(55.0) 20 (100.0) 0 4 (20.0) 9 (45.0) LGA Local Government Area. Mutated alleles are underlined. Dhfr haplotypes Table 6 Distribution of dhps mutations at the community locations in Lagos, Nigeria been shown to have important implications for the ef- fectiveness of SP in children less than 5 years of age and in pregnant women [41]. Reports from previous studies within Nigeria and in Africa were: 37.5% and 22.5% of A437G and K540E haplotype mutations respectively in Lagos (2011) [9] and 96.4% of 437G haplotype and no mutation at K540 codon in Calabar (2013/2014) [42], in pregnant women [41]. Reports from previous studies within Nigeria and in Africa were: 37.5% and 22.5% of A437G and K540E haplotype mutations respectively in Lagos (2011) [9] and 96.4% of 437G haplotype and no mutation at K540 codon in Calabar (2013/2014) [42], Table 7 Prevalence of dhfr and dhps haplotypes in Plasmodium falciparum isolates from Lagos, Nigeria Gene Category Haplotype n (%) dhfr Wild type CNCSI 8 (3.3) Double mutation CICNI 1 (0.4) CNRNI 3 (1.2) Triple mutation CIRNI 227 (93.8) Quadruple mutation CIRNIR 1 (0.4) CIRNIK 1 (0.4) CIRNIE 1 (0.4) dhps Single mutation SGKAA 74 (44.3) AAKAA 4 (2.4) Double mutations AGKAA 7 (4.2) FAKAS 3 (1.8) SGKGA 2 (1.2) AAKGA 1 (0.6) SGKAS 6 (3.6) SGEAA 2 (1.2) VSGKAA 1 (0.6) Triple mutations SGKGS 9 (5.4) VAGKAA 1 (0.6) AGKGA 1 (0.6) AGKAS 14 (8.4) Quadruple AGKGS 2 (1.2) VAGKAS 3 (1.8) VAGKGA 4 (2.4) Quintuple VAGKGS 33 (19.8) Mutations in the haplotype are underlined. Table 8 Prevalence of combined dhfr and dhps haplotypes combinations in Plasmodium falciparum isolates from Lagos, Nigeria Gene Category Haplotype n (%) dhfr/dhps (n = 141) Triple mutant CNRNI + SGKAA 2 (1.4) CNRNI + AAKAA 1 (0.7) Quadruple Mutant CIRNI + SGKAA 60 (42.6) CIRNI + AAKAA 2 (1.4) Quintuple mutant CIRNI + AGKAA 7 (5.0) CIRNI + SGKAS 5 (3.5) CIRNI + SGKGA 2 (1.4) CIRNI + FAKAS 2 (1.4) CIRNI + VSGKAA 1 (0.7) Sextuple mutant CIRNI + AGKAS 14 (9.9) CIRNI + SGKGS 4 (2.8) CIRNI + VAGKAA 1 (0.7) CIRNI + AGKGA 1 (0.7) CIRNI + AGKGA 1 (0.7) CIRNI + AGKGA 2 (1.4) Septuple mutant CIRNI + VAGKGA 4 (2.8) CIRNI + VAGKAS 2 (1.4) CIRNI + AGKGS 2 (1.4) Octuple mutant CIRNI + VAGKGS 31 (22.0) Mutated alleles are underlined. LGA Local Government Area. Mutations in the haplotype are underlined. Dhfr haplotypes Table 7 Prevalence of dhfr and dhps haplotypes in Plasmodium falciparum isolates from Lagos, Nigeria Table 7 Prevalence of dhfr and dhps haplotypes in Plasmodium falciparum isolates from Lagos, Nigeria Gene Category Haplotype n (%) dhfr Wild type CNCSI 8 (3.3) Double mutation CICNI 1 (0.4) CNRNI 3 (1.2) Triple mutation CIRNI 227 (93.8) Quadruple mutation CIRNIR 1 (0.4) CIRNIK 1 (0.4) CIRNIE 1 (0.4) dhps Single mutation SGKAA 74 (44.3) AAKAA 4 (2.4) Double mutations AGKAA 7 (4.2) FAKAS 3 (1.8) SGKGA 2 (1.2) AAKGA 1 (0.6) SGKAS 6 (3.6) SGEAA 2 (1.2) VSGKAA 1 (0.6) Triple mutations SGKGS 9 (5.4) VAGKAA 1 (0.6) AGKGA 1 (0.6) AGKAS 14 (8.4) Quadruple AGKGS 2 (1.2) VAGKAS 3 (1.8) VAGKGA 4 (2.4) Quintuple VAGKGS 33 (19.8) Mutations in the haplotype are underlined. Table 8 Prevalence of combined dhfr and dhps haplotypes combinations in Plasmodium falciparum isolates from Lagos, Ni i Mutated alleles are underlined. Page 9 of 11 Page 9 of 11 Page 9 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Nigeria. In Mukono District (Uganda), high frequency of mutation in dhps codon 437G (99.1%) and 540E (98.2%) (2010–2012) [43] was reported within the same period in which the samples in our study were collected. residue 164 is almost always found, but is rarely seen sub-Saharan Africa despite extensive use of the drug [48]. Moderate level of resistance conferred by dhfr and dhps polymorphisms is typically found in West Africa with the absence of I164L polymorphism that is associ- ated with very high-level SP resistance (up to 20 000-fold decrease in susceptibility in comparison with the wild type) [50]. I164L polymorphism have been variously re- ported in parts of East Africa [51], some parts of South Africa [52] and Asia [53]. There is dearth of information on why I164L mutation does not occur in Africa despite extensive drug pressure. It was suggested that this amino acid change carries a high fitness cost to the parasite, such that it is unable to survive the immune response of “malaria-experienced” hosts in West Africa [50]. Never- theless, though SP is ineffective in treating symptomatic disease in malaria-naïve children in many parts of Africa, it has retained some efficacy in preventing malaria in pregnant women [49]. Dhfr haplotypes The prevalence of 581G dhps haplotype mutation was (31.4%) and is associated with important modulatory role in resistance [39]. The World Health Organization (WHO) recommends that when the frequency of the 540 dhps haplotype mutation exceeds 95%, IPTp should not be implemented, because it could fail [40]. WHO also recommends that when the frequency of dhps Ala581Gly haplotype is above 10.0%, IPTp with SP may not be able to protect pregnant women from delivering infants with low birth weight [40]. Our study was in the general population and retrospectively, there were no existing data on birth outcomes to correlate this finding in the study areas where the samples were collected be- tween 2010 and 2014. Nevertheless, it underscored the need for regular molecular marker studies in areas where SP is used for malaria interventions. In addition, It is worth noting that the occurrence of the dhps 540E and Ala581Gly haplotype mutations are rare in West Af- rica, but common in east and southern Africa [39, 44]. For consideration that the samples we used in our study were collected nearly 10 years ago, and SP has been restricted to be used in malaria control among gen- eral population for several years in Nigeria due to severe drug resistance developed in P. falciparum, recovery of wild type of SP sensitive parasite could probably be ex- pected. Similar situation has ever occurred that chloro- quine sensitivity of P. falciparum reappeared after long time stopping of drug use in malaria control [54]. Recent investigation of SP sensitivity of malaria parasite in local area in Nigeria should be proceeded. The I431V occurred in combination with other dhps haplotype mutations, and the most frequent was VAGKGS haplotype. There were similar reports on this emerging mu- tation on dhps in Nigeria and Cameroun [10, 21, 22]. The occurrence of I431V mutation in dhps over the years may suggest conferment of selective advantage in the presence of SP drug pressure and displacement of the more sensitive haplotypes. Since SP is used as IPTp, and also readily avail- able in the Nigerian market for treatment of malaria [45], an indication that ongoing SP drug pressure is strong. An- other plausible explanation is that dhps haplotype mutation (431 V) has arisen by chance and provided an improvement in the fitness of parasites carrying the 437, 581 and/or 613 mutations, but does not change susceptibility to sulphadox- ine [10]. Conclusion This study showed a high prevalence of dhfr and dhps mutant alleles in Plasmodium falciparum isolates in Lagos, Nigeria, indicating that SP resistant parasites were in circulation five to 8 years after the introduction of ACT regimen. There was increased prevalence in dhfr triple haplotype mutations when compared with previ- ous reports in the same environment but aligned with high prevalence in other locations in Nigeria and other countries in Africa. Mutation in dhps, particularly 540E that is scarcely reported was low in this study. Partiallyt dhfr-dhps haplotype mutations were reported while I164L mutation that is consistently associated with SP resistance was not seen. This study has added to the rep- ertoire of SP haplotype research for analyses of trends and monitoring of threats to continued use of SP. Mo- lecular marker studies on resistant genotypes and haplo- types of SP remains invaluable where the medicine is used in various interventions by national malaria programmes. Partial resistance to SP [28] in dhfr-dhps haplotypes combinations were described in our study. The preva- lence was 42.6% and 22.0% for quadruple haplotype mu- tations, CIRNI-SGKAA and octuple haplotype (CIRNI + VAGKGS) respectively. Specifically, CIRNI-SGKAA was highly associated with sub-optimal IPTp-SP effectiveness in previous studies [46]. However, K540E haplotype mu- tation was not found in any of the dhfr-dhps combina- tions in our study. In West Africa wild dhps K540 commonly occur with triple dhfr mutations and single 437G dhps mutation [47, 48]. The highest levels and spread of antifolate re- sistance are found in Southeast Asia and South America [48–50]. In these two regions, a polymorphism at dhfr Dhfr haplotypes Further studies are needed to assess the effect of this mutation on the phenotype of parasites carrying this haplotype. SP: Sulphadoxine–pyrimethamine; ACT: Artemisinin-based combination therapy; Pfdhfr: Plasmodium falciparum dihydrofolate reductase; Abbreviations SP: Sulphadoxine–pyrimethamine; ACT: Artemisinin-based combination therapy; Pfdhfr: Plasmodium falciparum dihydrofolate reductase; Page 10 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 Quan et al. Infectious Diseases of Poverty (2020) 9:91 (2020) 9:91 dhps: dihydropteroate synthase; IPT: Intermittent preventive treatment; MiP: Malaria in pregnancy; SNPs: Single nucleotide polymorphism dhps: dihydropteroate synthase; IPT: Intermittent preventive treatment; MiP: Malaria in pregnancy; SNPs: Single nucleotide polymorphism University of Lagos, Lagos, Nigeria. 6Department of Surveying and Geoinformatics, Faculty of Engineering, University of Lagos, Lagos, Nigeria. Funding h k This work was supported by the National Research and Development Plan of China (Grant No. 2018YFE0121600), the Project of Shanghai Science and Technology Commission (Grant No. 18490741100), the National Sharing Service Platform for Parasite Resources (Grant No. TDRC-2019-194-30), the Foundation of National Science and Technology Major Program (Grant no. 2018ZX10734–404 and 2016ZX10004222–004), and the National Natural Science Foundation of China (Grant No. 81101266) and the Open Project of Key Laboratory of Parasite and Vector Biology, Ministry of Health (Grant No. WSBKTKT201405). 8. Agomo CO, Oyibo WA, Sutherland C, Hallet R, Oguike M. Assessment of markers of antimalarial drug resistance in Plasmodium falciparum isolates from pregnant women in Lagos, Nigeria. PLoS One. 2016;11(1):e0146908. 9. Iwalokun BA, Iwalokun SO, Adebodun V, Balogun M. Carriage of mutant dihydrofolate reductase and dihydropteroate synthase genes among Plasmodium falciparum isolates recovered from pregnant women with asymptomatic infection in Lagos, Nigeria. Med Princ Pract. 2015;24(5):436–43. 10. Oguike MC, Falade CO, Shu E, Enato IG, Watila I, Baba ES, et al. Molecular determinants of sulfadoxine-pyrimethamine resistance in Plasmodium falciparum in Nigeria and the regional emergence of dhps 431V. Int J Parasitol Drugs Drug Resist. 2016;6(3):220–9. Received: 13 December 2019 Accepted: 1 July 2020 Received: 13 December 2019 Accepted: 1 July 2020 Ethics approval and consent to participate 12. Dieckmann A, Jung A. Mechanisms of sulfadoxine resistance in Plasmodium falciparum. Mol Biochem Parasitol. 1986;19(2):143–7. Research Grants and Experimentation Ethics Committee, College of Medicine, University of Lagos, Nigeria and the Ethics Committee of the National Institute of Parasitic Diseases (NIPD), China CDC approved this study protocol. All the research participants / caregivers (in the case of children) gave informed consent/ assent. Patients that declined to participate in the study were not denied access to the available routine care in the health facility. Generally, the study was conducted in line with the principles of Good Clinical Laboratory Practice and ethical considerations were duly followed. 13. Triglia T, Menting JG, Wilson C, Cowman AF. Mutations in dihydropteroate synthase are responsible for sulfone and sulfonamide resistance in Plasmodium falciparum. Proc Natl Acad Sci U S A. 1997;94(25):13944–9. 14. Chulay JD, Watkins WM, Sixsmith DG. Synergistic antimalarial activity of pyrimethamine and sulfadoxine against Plasmodium falciparum in vitro. Am J Trop Med Hyg. 1984;33(3):325–30. 15. Cowman AF, Morry MJ, Biggs BA, Cross GA, Foote SJ. Amino acid changes linked to pyrimethamine resistance in the dihydrofolate reductase- thymidylate synthase gene of Plasmodium falciparum. Proc Natl Acad Sci U S A. 1988;85(23):9109–13. Acknowledgements We acknowledge the support received from the Lagos State Ministry of Health that gave approval to conduct this study at the facilities and communities from where samples were collected. Sincere thanks the patients that participated in this study and the management team of the various health facilities namely Ijede General Hospital, Imota Primary Health Centre, Bayeku Primary Health Centre, Agura Primary Health Centre, Ikorodu Lagos; Shomolu General Hospital, Shomolu, and St. Matthew’s Primary Health Centre, Amukoko, Lagos. We thank the Bio Venture for Global Health (BVGH) that fostered the collaboration through the WIPO Re: Search programme. References 1. WHO. Global Report on Antimalarial Drug Efficacy and Drug Resistance: 2000–2010. 2010. 2. FMOH: Federal ministry of health N, PP. le27. National Malaria Control Programme in Nigeria. Annual Report 2005. 3. Uhlemann AC, Krishna S. Antimalarial multi-drug resistance in Asia: mechanisms and assessment. Curr Top Microbiol Immunol. 2005;295:39–53 3. Uhlemann AC, Krishna S. Antimalarial multi drug resistance in Asia: mechanisms and assessment. Curr Top Microbiol Immunol. 2005;295:39–53. 4. Wellems TE, Plowe CV. Chloroquine-resistant malaria. J Infect Dis. 2001;184: 770–6. g p g The molecular characterization study was sponsored by the Chinese Ministry of Health, China. 4. Wellems TE, Plowe CV. Chloroquine-resistant malaria. J Infect Dis. 2001;184: 770–6. 4. Wellems TE, Plowe CV. Chloroquine-resistant malaria. J Infect Dis. 2001;184: 770–6. Availability of data and materials 11. Wang P, Lee CS, Bayoumi R, Djimde A, Doumbo O, Swedberg G, et al. Resistance to antifolates in Plasmodium falciparum monitored by sequence analysis of dihydropteroate synthetase and dihydrofolate reductase alleles in a large number of field samples of diverse origins. Mol Biochem Parasitol. 1997;89(2):161–77. Genetic data from this study deposited at the National Center for Biotechnology Information (NCBI). Accession numbers- BankIt2291211: MN985140 - MN985306 (167 sequences) and BankIt2298803: MT140637 - MT140882 (241 Sequences). Authors’ contributions 5. Wongsrichanalai C, Pickard AL, Wernsdorfer WH, Meshnick SR. Epidemiology of drug-resistant malaria. Lancet Infect Dis. 2002;2:209–18. 6. WHO. World Health Malaria Report. World Health Organization. 2014. HQ, UI and HMS collected the data and wrote the first draft; SO, SBC and CO collected the samples and analyzed the data; and WO, JHC and XNZ HQ, UI and HMS collected the data and wrote the first draft; SO, SBC and CO collected the samples and analyzed the data; and WO, JHC and XNZ designed the experiments, guided the English writing and revised the first draft. All authors read and approved the final manuscript. 7. Desai M, Gutman J, Taylor SM, Wiegand RE, Khairallah C, Kayentao K, et al. Impact of sulfadoxine-pyrimethamine resistance on effectiveness of intermittent preventive therapy for malaria in pregnancy at clearing infections and preventing low birth weight. Clin Infect Dis. 2016;62(3):323–33. Consent for publication Not applicable. 16. Contreras CE, Cortese JF, Caraballo A, Plowe CV. Genetics of drug-resistant Plasmodium falciparum malaria in the Venezuelan state of bolivar. Am J Trop Med Hyg. 2002;67(4):400–5. Competing interests 17. Khalil I, Rønn AM, Alifrangis M, Gabar HA, Satti GM, Bygbjerg IC. Dihydrofolate reductase and dihydropteroate synthase genotypes associated with in vitro resistance of Plasmodium falciparum to pyrimethamine, trimethoprim, sulfadoxine, and sulfamethoxazole. Am J Trop Med Hyg. 2003;68(5):586–9. The authors declare that they have no competing interests. Author details 1 1National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention, Chinese Center for Tropical Diseases Research, WHO Collaborating Center for Tropical Diseases, National Centre for International Research on Tropical Diseases, Ministry of Science and Technology, Key Laboratory of Parasite and Vector Biology, Ministry of Health, Shanghai 200025, People’s Republic of China. 2National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention⁃Shenzhen Center for Disease Control and Prevention Joint Laboratory for Imported Tropical Disease Control, Shanghai 200025, People’s Republic of China. 3Center for Infectious Diseases Research, Microbiology Department, Nigerian Institute of Medical Research, 6 Edmund Crescent, Yaba, Lagos, Nigeria. 4ANDI Center of Excellence for Malaria Diagnosis, Department of Medical Microbiology and Parasitology, College of Medicine, University of Lagos, Lagos, Nigeria. 5Department of Medical Microbiology and Parasitology, College of Medicine, 18. Triglia T, Wang P, Sims PF, Hyde JE, Cowman AF. Allelic exchange at the endogenous genomic locus in Plasmodium falciparum proves the role of dihydropteroate synthase in sulfadoxine-resistant malaria. EMBO J. 1998; 17(14):3807–15. 19. Berglez J, Iliades P, Sirawaraporn W, Coloe P, Macreadie I. Analysis in Escherichia coli of Plasmodium falciparum dihydropteroate synthase (DHPS) alleles implicated in resistance to sulfadoxine. Int J Parasitol. 2004;34(1):95–100. 20. Pearce RJ, Pota H, Evehe MS, Bâ EH, Mombo-Ngoma G, Malisa AL, Ord R, et al. Multiple origins and regional dispersal of resistant dhps in African Plasmodium falciparum malaria. PLoS Med. 2009;6(4):e1000055. 21. Sutherland CJ, Fifer H, Pearce RJ, Reza F, Nicholas M, Haustein T, Njimgye- Tekumafor NE, et al. Novel pfdhps haplotypes among imported cases of Page 11 of 11 Quan et al. Infectious Diseases of Poverty (2020) 9:91 (2020) 9:91 overty (2020) 9:91 efficacy of intermittent preventive treatment in pregnancy: systematic review and meta-analysis. Tropical Med Int Health. 2015;20(12):1621–33. efficacy of intermittent preventive treatment in pregnancy: systematic review and meta-analysis. Tropical Med Int Health. 2015;20(12):1621–33. Plasmodium falciparum malaria in the United Kingdom. Antimicrob Agents Chemother. 2009;53(8):3405–10. 42. Esu E, Tacoli C, Gai P, Berens-Riha N, Pritsch M, Loescher T, et al. Prevalence of the Pfdhfr and Pfdhps mutations among asymptomatic pregnant women in Southeast Nigeria. Parasitol Res. 2018;117(3):801–7. 22. Chauvin P, Menard S, Iriart X, Nsango SE, Tchioffo MT, Abate L, et al. Prevalence of Plasmodium falciparum parasites resistant to sulfadoxine/ pyrimethamine in pregnant women in Yaoundé, Cameroon: emergence of highly resistant pfdhfr/pfdhps alleles. J Antimicrob Chemother. 2015;70(9): 2566–71. 43. Author details 1 2001;17(12):582–8. 29. Odusanya OO, Akinyinka MR, Oluwole EO, Odugbemi BA, Bakare OQ, Adeniran A. How does the public perceive healthcare workers in Lagos? A comparison of health workers in public and private health facilities. Niger Postgrad Med J. 2018;25(3):177–85. 49. ter Kuile FO, van Eijk AM, Filler SJ. Effect of sulfadoxine-pyrimethamine resistance on the efficacy of intermittent preventive therapy for malaria control during pregnancy: a systematic review. JAMA. 2007;297(23):2603–16. 30. National Population Commision (NPC) (Nigeria) NMCPNNaII. Nigeria Malaria Indicator Survey 2010. Abuja, Nigeria and ICF International. 50. Nzila AM, Mberu EK, Nduati E, Ross A, Watkins WM, Sibley CH. Genetic diversity of Plasmodium falciparum parasites from Kenya is not affected by antifolate drug selection. Int J Parasitol. 2002;32(12):1469–76. 31. WHO. Basic Laboratory methods in Medical Parasitology. Geneva: World Health Organization; 1991. 32. Alam MT, de Souza DK, Vinayak S, Griffing SM, Poe AC, Duah NO, et al. Selective sweeps and genetic lineages of Plasmodium falciparum drug -resistant alleles in Ghana. J Infect Diss. 2011;203:220–7. 51. Lynch CA, Pearce R, Pota H, Egwang C, Egwang T, Bhasin A, et al. Travel and the emergence of high-level drug resistance in Plasmodium falciparum in Southwest Uganda: results from a population-based study. Malar J. 2017; 16(1):150. 33. Vinayak S, Alam MT, Mixson-Hayden T, McCollum AM, Sem R, Shah NK, et al. Origin and evolution of sulfadoxine resistant Plasmodium falciparum. PLoS Pathog. 2010;6:e1000830. 52. Kaingona-Daniel EP, Gomes LR, Gama BE, Almeida-de-Oliveira NK, Fortes F, Ménard D, et al. Low-grade sulfadoxine-pyrimethamine resistance in Plasmodium falciparum parasites from Lubango, Angolia. Malar J. 2016;15:309. 34. Jiang T, Chen J, Fu H, Wu K, Yao Y, Eyi JUM, et al. High prevalence of Pfdhfr- Pfdhps quadruple mutations associated with sulfadoxine-pyrimethamine resistance in Plasmodium falciparum isolates from Bioko Island, Equatorial Guinea. Malar J. 2019;18(1):101. 53. Basuki S, Fitriah RPM, Kasmijati AP, Riyanto S, et al. Origins and spread of novel genetic variants of sulfadoxine-pyrimethamine resistance in Plasmodium falciparum isolates in Indonesia. Malar J. 2018;17(1):475. 35. Basco LK, Eldin de Pécoulas P, Wilson CM, Le Bras J, Mazabraud A. Point mutations in the dihydrofolate reductase-thymidylate synthase gene and pyrimethamine and cycloguanil resistance in Plasmodium falciparum. Mol Biochem Parasitol. 1995;69(1):135–8. 54. Egan TJ, Kaschula CH. Strategies to reverse drug resistance in malaria. Curr Opin Infect Dis. 2007;20(6):598–604. 36. Nzila-Mounda A, Mberu EK, Sibley CH, Plowe CV, Winstanley PA, Watkins WM. Author details 1 Mbonye AK, Birungi J, Yanow SK, Shokoples S, Malamba S, Alifrangis M, et al. Prevalence of Plasmodium falciparum resistance markers to sulfadoxine-pyrimethamine among pregnant women receiving intermittent preventive treatment for malaria in Uganda. Antimicrob Agents Chemother. 2015;59(9):5475–82. 23. Plowe CV. The evolution of drug-resistant malaria. Trans R Soc Trop Med Hyg. 2009;103(Suppl 1):S11–4. 24. Kun JF, Lehman LG, Lell B, Schmidt-Ott R, Kremsner PG. Low-dose treatment with sulfadoxine-pyrimethamine combinations selects for drug-resistant Plasmodium falciparum strains. Antimicrob Agents Chemother. 1999;43(9): 2205–8. 44. van Eijk AM, Larsen DA, Kayentao K, Koshy G, Slaughter DEC, Roper C, et al. Effect of Plasmodium falciparum sulfadoxine-pyrimethamine resistance on the effectiveness of intermittent preventive therapy for malaria in pregnancy in Africa: a systematic review and meta-analysis. Lancet Infect Dis. 2019;19(5):546–56. 25. Kublin JG, Dzinjalamala FK, Kamwendo DD, Malkin EM, Cortese JF, Martino LM, Mukadam RAG, et al. Molecular markers for failure of sulfadoxine- pyrimethamine and chlorproguanil-dapsone treatment of Plasmodium falciparum malaria. J Infect Dis. 2002;185(3):380–8. 45. Ugwu EO, Iferikigwe ES, Obi SN, Ugwu AO, Agu PU, Okezie OA. Anti-malaria prescription in pregnancy among general practitioners in Enugu state, south East Nigeria. Niger Med J. 2013;54(2):96–9. 26. Happi CT, Gbotosho GO, Folarin OA, Akinboye DO, Yusuf BO, Ebong OO, et al. Polymorphisms in Plasmodium falciparum dhfr and dhps genes and age related in vivo sulfadoxine-pyrimethamine resistance in malaria-infected patients from Nigeria. Acta Trop. 2005;95(3):183–93. 46. Minja DT, Schmiegelow C, Mmbando B, Boström S, Oesterholt M, Magistrado P, et al. Plasmodium falciparum mutant haplotype infection during pregnancy associated with reduced birthweight, Tanzania. Emerg Infect Dis. 2013;19(9):1446–54. 27. Afonso A, Hunt P, Cheesman S, Alves AC, Cunha CV. Do Rosário V, Cravo P, et al. Malaria parasites can develop stable resistance to artemisinin but lack mutations in candidate genes atp6 (encoding the sarcoplasmic and endoplasmic reticulum Ca2+ ATPase), tctp, mdr1, and cg10. Antimicrob Agents Chemother. 2006;50(2):480–9. 47. Mugittu K, Ndejembi M, Malisa A, Lemnge M, Premji Z, Mwita A, et al. Therapeutic efficacy of sulfadoxine-pyrimethamine and prevalence of resistance markers in Tanzania prior to revision of malaria treatment policy: Plasmodium falciparum dihydrofolate reductase and dihydropteroate synthase mutations in monitoring in vivo resistance. Am J Trop Med Hyg. 2004;71(6):696–702. 28. Naidoo I, Roper C. Mapping 'partially resistant', 'fully resistant', and 'super resistant' malaria. Trends Parasitol. 2013;29(10):505–15. 48. Sibley CH, Hyde JE, Sims PF, Plowe CV, Kublin JG, Mberu EK, et al. Pyrimethamine-sulfadoxine resistance in Plasmodium falciparum: what next? Trends Parasitol. 54. Egan TJ, Kaschula CH. Strategies to reverse drug resistance in malaria. Curr Opin Infect Dis. 2007;20(6):598–604. Author details 1 Kenyan Plasmodium falciparum field isolates: correlation between pyrimethamine and chlorcycloguanil activity in vitro and point mutations in the dihydrofolate reductase domain. Antimicrob Agents Chemother. 1998; 42(1):164–9. 37. Foote SJ, Kyle DE, Martin RK, Oduola AM, Forsyth K, Kemp DJ, Cowman AF, et al. Several alleles of the multidrug-resistance gene are closely linked to chloroquine resistance in Plasmodium falciparum. Nature. 1990;345(6272): 255–8. 38. Peterson DS, Milhous WK, Wellems TE. Molecular basis of differential resistance to cycloguanil and pyrimethamine in Plasmodium falciparum malaria. Proc Natl Acad Sci U S A. 1990;87(8):3018–22. 39. Gutman J, Kalilani L, Taylor S, Zhou Z, Wiegand RE, Thwai KL, et al. The A581G mutation in the gene encoding Plasmodium falciparum dihydropteroate synthetase reduces the effectiveness of sulfadoxine- pyrimethamine preventive therapy in Malawian pregnant women. J Infect Dis. 2015;211(12):1997–2005. 40. WHO. WHO Evidence Review Group on Intermittent Preventive Treatment (IPT) of malaria in pregnancy. Geneva: World Health Organization; 2013. 41. Chico RM, Cano J, Ariti C, Collier TJ, Chandramohan D, Roper C, et al. Influence of malaria transmission intensity and the 581G mutation on the
https://openalex.org/W3139789650	https://wiyatamandala.e-journal.id/JBM/article/download/154/111	Swahili	null	PENGARUH SOCIAL IDENTITY, EWOM, PERCEIVED RISK, DAN TRUST TERHADAP PURCHASE INTENTION DAN DAMPAKNYA TERHADAP PURCHASE DECISION PADA E-COMMERCE	Jurnal Bina Manajemen	2,021	cc-by	6,404	Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce ABSTRACT ABSTRACT This zstudy zwas zdesigned zto ztest zempirical ztruths zbetween ztwo zor zmore zresearch zvariables zthat zhave zbeen zformulated zin zthe zhypothesis. zThe zresearch zvariables zused zwere zsocial zidentity, zEWOM, zperceived zrisk, ztrust, zpurchase zintention, zand zpurchase zdecision. zExogenous zvariables zin zthis zstudy zare zsocial zidentity, zEWOM, zand zperceived zrisk. zEndogenous zvariables zin zthis zstudy zare ztrust, zpurchase zintention, zand zpurchase zdecision. zMethods zof zcollecting zdata zusing zquestionnaires zand zdata zanalysis zmethods zusing zpartial zleast zsquare. zThe zresults zof zthe zstudy zconcluded zthat z zhypothesis z4 zand z5 zwas znot zsignificant. zThere zwas zno zeffect zof zperceived zrisk zon ztrust zand zthere zwas zno zeffect zof zperceived zrisk zon z zpurchase zintention. Keywords: zSocial zIdentity, zEWOM, zPerceived zRisk, zTrust, zPurchase zIntention, zPurchase zDecision. Keywords: zSocial zIdentity, zEWOM, zPerceived zRisk, zTrust, zPurchase zIntention, zPurchase zDecision. Kata kunci: Identitas Sosial, EWOM, Persepsi Resiko, Kepercayaan, Niat Membeli, Keputusan Pembelian. PENGARUH zSOCIAL zIDENTITY, zEWOM, zPERCEIVED zRISK, zDAN zTRUST zTERHADAP zPURCHASE zINTENTION zDAN zDAMPAKNYA zTERHADAP zPURCHASE zDECISION zPADA zE-COMMERCE Tessa zHandra z Universitas zMultimedia zNusantara z tessa.handra@lecturer.umn.ac.id z Felix zSutisna z Universitas zMultimedia zNusantara z felix.sutisna@lecturer.umn.ac.id ABSTRAK Penelitian ini dirancang untuk menguji kebenaran empiris antara dua atau lebih variabel penelitian yang telah dirumuskan dalam hipotesis. Variabel zpenelitian yang digunakan adalah identitas sosial, EWOM, persepsi risiko, kepercayaan, niat membeli, dan keputusan pembelian. Variabel eksogen dalam penelitian ini adalah identitas sosial, EWOM, dan persepsi risiko. Variabel endogen dalam penelitian ini adalah kepercayaan, niat membeli, dan keputusan pembelian. Metode pengumpulan data menggunakan kuesioner dan metode analisis data menggunakan partial least square. Hasil penelitian menyimpulkan bahwa hipotesis 4 dan 5 tidak signifikan. Tidak ada pengaruh persepsi risiko terhadap kepercayaan dan tidak zada pengaruh persepsi risiko terhadap niat membeli. Kata kunci: Identitas Sosial, EWOM, Persepsi Resiko, Kepercayaan, Niat Membeli, Keputusan Pembelian. 126 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 PENDAHULUAN z tersebut zmaka zakan zbanyak ziklan zyang zmenawarkan produk-produk zdari zpenjual. zSelain zitu zkeuntungan zbagi zpembeli zdapat zmelakukan zperbandingan zharga zdan zragam zjenis zbarang zserta ztidak zperlu zlagi zmengeluarkan zwaktu,zbiaya zdan ztenaga zuntuk zmenemukan zbarang zkebutuhan zsemua zsudah zdigantikan zdengan ze-commerce. zPerkembangan ze-commerce zdi zIndonesia zbegitu zcepat zdan zberkembang zsangat zluas, zangka zpenjualan ze-commerce ztersebut zjuga zmelihatkan ztanda- tanda zyang zke zarah zpositif. zHal zini zsemakin zdidukung zdengan zmunculnya zberbagai ze-commerce zseperti zTokopedia, zShopee, zBlibli, zBukalapak, zJD.id zdan ze-commerce zlainnya. zSelain zitu zperkembangan ze-commerce zdi zIndonesia zmasih zmenjadi zbisnis zyang zsangat zmenguntungkan. zSemakin zbanyaknya ze-commerce zyang zada zdi zIndonesia zakan zsemakin zmemudahkan zdan zmemberikan zbanyak zpilihan zbagi zmasyakarat zIndonesia zdalam zmembeli zbarang. zSelan zitu zkonsumen zjuga zsemakin zterbantu zdalam zmendapatkan Zaman zsekarang zini ztransaksi zantar zpedagang zdan zpembeli ztidak zlagi zharus zbertatap zmuka, zsemua zbisa zdilakukan zdi zmana zsaja zdan zkapan zsaja. zKemajuan zdan zkemutakhiran zteknologi zsekarang zini zmempengaruhi zgaya zhidup zkonsumen zluas. zDimulai zdari zbidang zkomunikasi zyaitu zmunculnya zberbagai zhandphone zpintar, zlalu zbermunculan zmedia zsosial zdan zyoutube zdan zyang zsekarang zbanyak zdilirik zpara zpebisnis zdi zIndonesia zyaitu ze- commerce. zE-commerce zadalah zsuatu zproses zberbisnis zdengan zmenggunakan zteknologi zelektronik zyang zmenghubungkan zantara zperusahaan, zkonsumen zdan zkonsumen zdalam zbentuk ztransaksi zelektronik zdan zpertukaran/penjualan zbarang, zservis, zdan zinformasi zsecara zelektronik. zE-commerce zakan zsangat zmembantu zsekali zdalam zmenghemat zwaktu, zbiaya zdan ztenaga zbagi zpenjual zdan zdari zsisi zpembeli, zpembeli zakan zdimudahkan zuntuk zmencari zsegala zkebutuhan zyang zdiperlukan zdengan zmasuk zke zweb ze-commerce z Zaman zsekarang zini ztransaksi zantar zpedagang zdan zpembeli ztidak zlagi zharus zbertatap zmuka, zsemua zbisa zdilakukan zdi zmana zsaja zdan zkapan zsaja. zKemajuan zdan zkemutakhiran zteknologi zsekarang zini zmempengaruhi zgaya zhidup zkonsumen zluas. zDimulai zdari zbidang zkomunikasi zyaitu zmunculnya zberbagai zhandphone zpintar, zlalu zbermunculan zmedia zsosial zdan zyoutube zdan zyang zsekarang zbanyak zdilirik zpara zpebisnis zdi zIndonesia zyaitu ze- commerce. zE-commerce zadalah zsuatu zproses zberbisnis zdengan zmenggunakan zteknologi zelektronik zyang zmenghubungkan zantara zperusahaan, zkonsumen zdan zkonsumen zdalam zbentuk ztransaksi zelektronik zdan zpertukaran/penjualan zbarang, zservis, zdan zinformasi zsecara zelektronik. PENDAHULUAN z zKonsumen zsadar zakan zkebutuhan zsuatu zproduk zyang zmenjadi zjawaban zatas zmasalah zyang zdihadapi zkonsumen, ztinggal zproduk zmana zyang zharus zdibeli. zPemilihan zproduk zberdasarkan zinformasi-informasi zyang zdigali zoleh zkonsumen zmelalui zberbagai zmedia. zSetelah zitu zkonsumen zmengevaluasi zdan zmempertimbangkan zproduk zyang zakan zdibeli znanti zberdasarkan zinformasi ztersebut zsehingga zakhirnya zmemutuskan zsuatu zproduk zyang ztepat zsesuai zkebutuhan zdan zmanfaat zproduk ztersebut. zMenurut zSchiffman zdan zKanuk z(2004) zkeputusan zpembelian zsebagai zpemilihan zdari zdua zatau zlebih zalternatif zpilihan zkeputusan zpembelian. zKonsumen zharus zmenentukan zpilihan zdari zberbagaizalternatif zyang zada. z konsumen zuntuk zmelakukan zpembelian zsuatu zproduk zdiawali zoleh zadanya zkesadaran zatas zpemenuhan zkebutuhan zdan zkeiinginan. zKonsumen zsadar zakan zkebutuhan zsuatu zproduk zyang zmenjadi zjawaban zatas zmasalah zyang zdihadapi zkonsumen, ztinggal zproduk zmana zyang zharus zdibeli. zPemilihan zproduk zberdasarkan zinformasi-informasi zyang zdigali zoleh zkonsumen zmelalui zberbagai zmedia. zSetelah zitu zkonsumen zmengevaluasi zdan zmempertimbangkan zproduk zyang zakan zdibeli znanti zberdasarkan zinformasi ztersebut zsehingga zakhirnya zmemutuskan zsuatu zproduk zyang ztepat zsesuai zkebutuhan zdan zmanfaat zproduk ztersebut. zMenurut zSchiffman zdan zKanuk z(2004) zkeputusan zpembelian zsebagai zpemilihan zdari zdua zatau zlebih zalternatif zpilihan zkeputusan zpembelian. zKonsumen zharus zmenentukan zpilihan zdari zberbagaizalternatif zyang zada. z PENDAHULUAN z zE-commerce zakan zsangat zmembantu zsekali zdalam zmenghemat zwaktu, zbiaya zdan ztenaga zbagi zpenjual zdan zdari zsisi zpembeli, zpembeli zakan zdimudahkan zuntuk zmencari zsegala zkebutuhan zyang zdiperlukan zdengan zmasuk zke zweb ze-commerce z 127 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce informasi zbaik zharga, zkualitas, zdan promo. zNamun zuntuk zmenjadi ze- commerce zpilihan zbanyak zorang perlu zusaha zdan zperforma zyang baik dari ze-commerce zdalam zmelayani konsumen zIndonesia. zDalam menentukan zpemilihan zhingga pembeli zmemutuskan zpembelian zdi e-commerce zyang zakan zdigunakan, konsumen zakan zmengumpulkan banyak zinformasi zsebagai zbahan pertimbangan zuntukzmembuat keputusan zpembelian. zOleh zkarena itu zanalisis zmengenai zkeputusan pembelian zsangat zdibutuhkan zuntuk menjawab zmasalah zfaktor zapa zsaja yang zmempengaruhi zkonsumen dalam zmemutuskan zmembeli zpada salah zsatu ze-commerce. Untuk menganalisis zpermasalahan keputusan zpembelian zmaka pendekatan zyang zdapat zdigunakan adalah zpendekatan ztheory zofzplanned behavior z(TPB) zyang zmempunyai asumsi zbahwa zsuatu zperilaku manusia zdidahului zoleh zniat zatau kepentingan zyaitu zperlu zdilakukan atau ztidak zdilakukan zperilaku. Dengan zdemikian, zniat zbeli zdan keputusan zaktual zmembeli zmemiliki hubungan zyang zproporsional. Beberapa zpenelitian zmendukung zpenelitian zini zyaitu zpenelitian zEndhar zet zal. z(2016), zSony z(2016), zSanda z(2017), zHana, zet zal. z(2016), zDwi zdan zMahendra z(2013). zBerdasarkan zbeberapa zpenelitian zterdahulu zmaka zpenelitian zini zmengambil zvariabel zsocial zidentity, zEWOM, zperceived zrisk, ztrust, zdanzpurchase zintention zsebagai zvariabel zyang zmempengaruhi zpurchase zdecision. zOleh zkarena zitu ztujuan zpenelitian zini zadalah zmenganalisis zpengaruh zsocial zidentity, zEWOM, zperceived zrisk, ztrust, zdan zpurchase zintention zterhadap zpurchase zdecision. Gunakan ztabel zdan zgambar zdengan zpenyesuaian zterhadap zpanjang zlebar zkolom. zPada zhalaman zterakhir zmakalah zAnda, zatur zpanjang zlebarnya zagar zsama zbesar. zGunakan ztanda zhubung zotomatis zdan zpemeriksa zejaan z(bila ztersedia). zTiap zkolom zdiatur zrata zkiri-kanan. zGunakan ztabel zdan zgambar zdengan zpenyesuaian zterhadap zpanjang zlebar zkolom. zPada zhalaman zterakhir zmakalah zAnda, zatur zpanjang zlebarnya zagar zsama zbesar. zGunakan ztanda zhubung zotomatis zdan zpemeriksa zejaan z(bila ztersedia). 128 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 Tiap zkolom zdiatur zrata zkiri-kanan. Gunakan ztabel zdan zgambar zdengan penyesuaian zterhadap zpanjang zlebar kolom. zPada zhalaman zterakhir zmakalah zAnda, zatur zpanjang zlebarnya zagar zsama zbesar. zGunakan ztanda zhubung zotomatis zdan zpemeriksa zejaan z(bila ztersedia). konsumen zuntuk zmelakukan zpembelian zsuatu zproduk zdiawali zoleh zadanya zkesadaran zatas zpemenuhan zkebutuhan zdan zkeiinginan. TELAAH ZLITERATUR Z Keputusan zPembelian z(Purchase zdecision), zmerupakan zproses zdimana konsumen zmembuat zkeputusan zuntukzmembeli zberbagai zproduk zdan zmerek yang zdimulai zdengan zpengenalan zkebutuhan, zpencarian zinformasi,evaluasi zinformasi, zmembuat zpembelian zdan zkemudian zmengevaluasi zkeputusan zsetelah zmembeli. zBeberapa zfaktor zyangzberasal zdari zlingkungan zseperti zbudaya, zkelas zsosial, zpengaruh zkelompok zdan zkeluarga zdapat zmempengaruhi zproses zpengambilan zkeputusan zseseorang. zAdapun zbeberapa zhal zyang zdapat zmempengaruhi zkeputusan zpembelian zseperti zsumber zdaya zkonsumen, zmotivasi, zdan zketerlibatan, zpengetahuan, zsikap zdan zkepribadian zserta zgaya zhidup zdan zdemografi z(Engel zet zal,1995). zKeputusan zpembelian zadalah zpengambilan zkeputusan zoleh z 129 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Electronic zWord zof zMouth Electronic zword zof zMouth zmerupakan zpernyataan zpositif zatau znegatif zyang dibuat zoleh zpelanggan zpotensial,zpelanggan zaktual zdan zmantan zpelanggan tentang zproduk zatau zperusahaan z zmelalui zinternet z(Malik zet zal., z2013). z Menurut zHennig-Thurau z(2004), zelectronic zword zof zmouth z(EWOM) zmerujuk pada zpernyataan zpositif zatau znegatif zdari zpotential, zaktual zatau zkonsumen pendahulu zmengenai zsuatu zproduk zatau zperusahaan zvia zinternet. zWeb ztelah menciptakan zkesempatan zkepada zEWOM zberkomunikasi zmelalui zberbagai macam zmedia zseperti zforum zdiskusi, zelectronic zbulletin zboard, znewsgroup, blog, zdan zsocial znetworking z(Goldsmith, z2006). zMenurut zHennig-Thurau zet al., z(2004), zEWOM zmerupakan zbentuk zkomunikasi zpemasaran zyang zberisi tentang zpernyataan zpositif zatau znegatif zyang zdilakukan zpelanggan z zpotensial, pelanggan zmaupun zmantan zpelanggan ztentang zsuatu zproduk zatau zperusahaan, yang ztersedia zbagi zbanyak zorang zatau zlembaga zmelalui zmedia zinternet. Electronic zWord zof zMouth Electronic zword zof zMouth zmerupakan zpernyataan zpositif zatau znegatif zyang dibuat zoleh zpelanggan zpotensial,zpelanggan zaktual zdan zmantan zpelanggan tentang zproduk zatau zperusahaan z zmelalui zinternet z(Malik zet zal., z2013). z Menurut zHennig-Thurau z(2004), zelectronic zword zof zmouth z(EWOM) zmerujuk pada zpernyataan zpositif zatau znegatif zdari zpotential, zaktual zatau zkonsumen pendahulu zmengenai zsuatu zproduk zatau zperusahaan zvia zinternet. zWeb ztelah menciptakan zkesempatan zkepada zEWOM zberkomunikasi zmelalui zberbagai macam zmedia zseperti zforum zdiskusi, zelectronic zbulletin zboard, znewsgroup, blog, zdan zsocial znetworking z(Goldsmith, z2006). zMenurut zHennig-Thurau zet al., z(2004), zEWOM zmerupakan zbentuk zkomunikasi zpemasaran zyang zberisi tentang zpernyataan zpositif zatau znegatif zyang zdilakukan zpelanggan z zpotensial, pelanggan zmaupun zmantan zpelanggan ztentang zsuatu zproduk zatau zperusahaan, yang ztersedia zbagi zbanyak zorang zatau zlembaga zmelalui zmedia zinternet. TELAAH ZLITERATUR Z Intensi zPembelian Purchase zintention zatau zminat zbeli zmenurut zAssael z(1998) zmerupakan zkecenderungan zkonsumen zuntuk zmembeli zsuatu zmerek zatau zmengambil ztindakan zyang zberhubungan zdengan zpembelian zyang zdiukur zdengan ztingkat zkemungkinan zkonsumen zmelakukan zpembelian. zMenurut zKotler z& zKeller z(2006), zpurchase zintention zadalah zseberapa zbesar zkemungkinan zkonsumen zmembeli zsuatu zmerek zatau zseberapa zbesar zkemungkinan zkonsumen zuntuk zberpindah zdari zsatu zmerek zke zmerek zlainnya. zSchiffman zdan zkanuk z(2004), zmengemukakan zbahwa zminat zmembeli zmerupakan zpikiran zyang ztimbul zkarena zadanya zperasaan ztertarik zdan zingin zmemiliki zterhadap zsuatu zbarang zatau zjasa zyang zdiharapkan. Penulis zmemiliki zdefinisi zsendiri tentang zpurchase zintention z zyaitu suatu keadaan zyang zdimana seseorang zsadar zuntuk zbertahan hidup zharus zmemenuhi segala kebutuhan zpokoknya zdengan melakukan zpembelian zuntuk melanjutkan hidupnya Intensi zPembelian Purchase zintention zatau zminat zbeli zmenurut zAssael z(1998) zmerupakan zkecenderungan zkonsumen zuntuk zmembeli zsuatu zmerek zatau zmengambil ztindakan zyang zberhubungan zdengan zpembelian zyang zdiukur zdengan ztingkat zkemungkinan zkonsumen zmelakukan zpembelian. zMenurut zKotler z& zKeller z(2006), zpurchase zintention zadalah zseberapa zbesar zkemungkinan zkonsumen zmembeli zsuatu zmerek zatau zseberapa zbesar zkemungkinan zkonsumen zuntuk zberpindah zdari zsatu zmerek zke zmerek zlainnya. zSchiffman zdan zkanuk z(2004), zmengemukakan zbahwa zminat zmembeli zmerupakan zpikiran zyang ztimbul zkarena zadanya zperasaan ztertarik zdan zingin zmemiliki zterhadap zsuatu zbarang zatau zjasa zyang zdiharapkan. Penulis zmemiliki zdefinisi zsendiri tentang zpurchase zintention z zyaitu suatu keadaan zyang zdimana seseorang zsadar zuntuk zbertahan hidup zharus zmemenuhi segala kebutuhan zpokoknya zdengan melakukan zpembelian zuntuk l j k hid Intensi zPembelian 130 Jurnal Bina Manajemen, Maret 2021, Vol.8 No.2 Hal 126 - 145 zjumlah zyang zbesar zpesan yang zada zdi zEWOM zsecara zonline zdan zmenganalisis zkarakteristik zmereka seperti zjumlah zkata zsentimental zyang zdigunakan zposisi zpesan, zgaya zpesan, dan sejenisnya. zNamun zada zkekurangan zdari zEWOM zdimana ztidak zdapat melakukan zpenilaian zkreadibilitas zkomunikator zmelalui zsistem zreputasi zonline. Menurut zGoldsmith zand zHorowitz z(2006), zberpendapat zbahwa zdi zdunia online, ada zbermacam-macam zcara zdimana zkonsumen zdapat zbertukar zinformasi. Pengguna zinternet zdapatzmelakukan zElectronic zword zof zmouth zmelalui berbagai zsaluran zonline, ztermasuk zblog, zmikroblog, zemail, zsitus zulasan (review) zkonsumen, zforum, zkomunitas zkonsumen zvirtual, zdan zsitus zjejaring sosial. zPenulis zberpendapat zbahwa zEWOM zadalah zsuatu zpengalaman zyang dirasakan zoleh zpengguna zbaik zitu zproduk zmaupun zjasa zbaik zpengalaman zyang mereka zdapat zadalah zpengalaman zbaik zatau zburuk zdari zproduk zatau zjasa tersebut zyang zdisebarluaskan zoleh zpengguna zmenggunakan zmedia zinternet yang zdapat zdiakses zoleh z EWOM zmemiliki zjangkauan zyang sangat zluas zdan zkecepatan zdalam penyebaran zinformasinya zsangat cepat zdan zbahkan ztak zterkendali karena menggunakan zberbagai macam zmedia zdiinternet zseperti web, zblog, zmedia sosial, zyoutube, dan ztwitter. zberbeda zdengan zword of zmouth zyang ztraditional yang hanya zmengandalkan zrekomendasi dari zorang-orang zyang zkita zkenal saja atau zdari zkomunitas zyang zkita miliki zdan zdisampaikan zsecara langsung melalui ztatap zmuka zatau berkomunikasi zsecara zlangsung zdan tidak ztersebar zdi media zinternet. Informasi zdalam zbentuk z zEWOM tidak zperlu zada zpertukaran informasi zantar zsesama zpembaca atau zkomunikator zhadir. Contohnya, pengguna zforum zdapat membaca zdan menulis zkomentar orang zlain zsetelah topik zpembicaraan dibuat. zEWOM lebih ztetap zdan mudah zdiakses kapan zsaja dan dimana zsaja. z Hampir zsebagian zbesar zEWOM ditulis zdalam zbentuk zteks zyang ditampilkan di internet zdiarsipkan sehingga zakan ztersedia zdalam zwaktu yang ztidak zterbatas. Peneliti zdapat dengan zmudah zmengambil zdalam 131 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce zdirepresentasikan zsebagai zkognitif zyang ztergantung zpada konteks zprototipe zyang zmenangkap zsifat zkhas zkelompok.Menurut zPadilla z& Perez z(2003), zteori zidentitas zsosial zmenyatakan zbahwa zindividu zberpikir, merasa, zdan zbertindak zsebagaimana zyang zdilakukan zoleh zanggota zkelompok yang zdiikuti. zPendapat zlain zmenyatakan zbahwa zindividu zmemutuskan bergabung zdengan zkelompok zsosial zyang zsesuai zdengan zdirinya zdan zdapat meningkatkan zstatus zsosialnya. zBernad zet zal z(2016). Intensi zPembelian z zPenulis zmemiliki pengertian zlain ztentang zidentitas zsosial zyaitu zSebuah zkomunitas zyang memiliki kesamaan zbaik zhobby, zpekerjaan, zgaya zhidup ataupunzkesamaan-kesamaan tertentu zyang zbergabung zmenjadi zsatu zdan zmembuat zsuatu zidentitas zsendiri baik zdiakui zatau ztidak zdiakui zoleh zmasyarakat zluas. semua zorang zyang zmempunyai zakses zinternet. Identitas zsosial Identifikasi zdengan zkelompok zsosial zadalah zkeadaan zpsikologis zyang sangatzberbeda zdari zkategori zsosial ztertentu zdan zmemiliki zhasil zevaluasi zdiri yang zpenting. zsecara zkeseluruhan, zidentitas zsosial zdapat zmendukung zketika dalam zkeadaan ztertekan. zTeori zini zdikemukakan zoleh zDholakia zet zal z(2004). Menurut zTajfel z(1982), zidentitas zsosial zadalah zbagian zdari zkonsep zdiri seseorang zyang zberasal zdari zpengetahuan zmereka ztentang zkeanggotaan zdalam suatu zkelompom zsosial zbersamaan zdengan zsignifikansi znilai zdan zemosional dari zkeanggotaan ztersebut. zIdentitas zmerupakan zpersepsi zatas zanggota kelompok zterhadap zkelompoknya zyangzdianggap zmemiliki zkesamaan zdan membentuk zsuatu zikatan zemosional zantar zanggota zkelompok zdan zkelompok itu sendiri.Hogg z& zReid z(2006), zmengartikan zbahwa zidentitas zsosial zdasar zuntuk sejumlah zfenomena zkomuni-katif zyang znyata, zmenjelaskan bagaimana norma kelompok yang Identitas zsosial Identifikasi zdengan zkelompok zsosial zadalah zkeadaan zpsikologis zyang sangatzberbeda zdari zkategori zsosial ztertentu zdan zmemiliki zhasil zevaluasi zdiri yang zpenting. zsecara zkeseluruhan, zidentitas zsosial zdapat zmendukung zketika dalam zkeadaan ztertekan. zTeori zini zdikemukakan zoleh zDholakia zet zal z(2004). Menurut zTajfel z(1982), zidentitas zsosial zadalah zbagian zdari zkonsep zdiri seseorang zyang zberasal zdari zpengetahuan zmereka ztentang zkeanggotaan zdalam suatu zkelompom zsosial zbersamaan zdengan zsignifikansi znilai zdan zemosional dari zkeanggotaan ztersebut. zIdentitas zmerupakan zpersepsi zatas zanggota kelompok zterhadap zkelompoknya zyangzdianggap zmemiliki zkesamaan zdan membentuk zsuatu zikatan zemosional zantar zanggota zkelompok zdan zkelompok itu sendiri.Hogg z& zReid z(2006), zmengartikan zbahwa zidentitas zsosial zdasar zuntuk sejumlah zfenomena zkomuni-katif zyang znyata, zmenjelaskan zbagaimana znorma kelompok zyang Peneriman zRisiko Oglethorpe zdan zMonroe z(1994) zmengekumakan zbahwa zkonsumen zmenghadapi ketidakpastian zdanzkonsekuensi-konsekuensi znegatif zyang zmungkin zditerima atas zpembelian zsuatu zproduk zatau zjasa. Peneriman zRisiko Oglethorpe zdan zMonroe z(1994) zmengekumakan zbahwa zkonsumen zmenghadapi ketidakpastian znegatif zyang zmungkin zditerima atas zpembelian zsuatu zproduk zatau zjasa. 131 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 Konsumen zgiat zmencari zinformasi tambahan zketika zdihadapkan zpada pembelian zyang zberisiko ztinggi. zLui dan Jamieson z(2003) zmenyatakan tingkat zrisiko zdalam zberbelanja secara zonline tergantung zpada persepsi zkonsumen zdalam memperkirakan ztinggi zrendahnya risiko zyang zakan zdialami zketika menggunakan zinternet zuntuk berbelanja. Konsumen zmemiliki persepsi zresiko zyang zlebih ztinggi ketika zmelakukan pembelian zsecara online zjika zdibandingkan zketika mereka zmelakukan pembelian melalui toko. z(Suresh zdan zShashikala, 2011). Intensi zPembelian Ada zpendapat zlain yang zmengatakan bahwa zpersepsi zakan zresiko zdinilai sebagai ztingkat anggapan zpelanggan akan zhasil znegatif zyang zmungkin terjadi zketika melakukan ztransaksi secara zonline z(Featherman zand Pavlou, z2002). zDalam melakukan pembelian zsecara zonlineada ztiga macam zresiko zyang zakan dihadapi oleh zkonsumen zyaitu zrisiko zproduk, risiko ztransaksi, zdan zrisiko psikologis. Risiko zproduk zmengacu pada zketidakpastian zbahwa zproduk yang zdibeli zsecara online tidak sesuai dengan zharapan zkarena ztidak zpernah melihat zlangsung secara zfisik zatau zmemegangnya zsecara zlangsung. zRisikoztransaksi zadalah resiko ketidakpastian zdalam zpembayaran zapakah zuang zyang zdikirim zsampai zpada orang zyang ztepat zatau ztidak. zRisiko zpsikologis zadalah zrasa zketakutan zdan khawatir zyang ztimbul zselama zproses zpembelian zsampai zproses zpengiriman barang ztersebut zsampai zditangan zkonsumen.zKondisi zini zmenghasilkan pengambilan zkeputusan zyang zkompleks. zKonsumen zsecara ztidak zlangsung akan melakukan zevaluasi zsecara zdetail zterhadap zproduk zdan zmerek zyang zmenjadi pesaing zdari zproduk ztersebut. zPenulis zmengartikan z zPerceived zof zrisk zadalah persepsi zatau zpemikiran ztentang zrisiko zyang zakan zdialami zoleh zkonsumen, suatu zketidakpastian zdan zkonsekuensi- konsekuensi zyang zbisa zterjadi zyang akan diterima zoleh zkonsumen. Kepercayaan Kepercayaan Trust zdiartikan zsebagai zkesediaan zmengandalkan zkemampuan, zintergritas zdan motivasi z zpihak zlain zuntuk zbertindak zdalam zrangka zmemuaskan zkebutuhan dan zkepentingan zseseorang zsebagaimana zdisepakati zbersama zsecara zimplesit Trust zdiartikan zsebagai zkesediaan zmengandalkan zkemampuan, zintergritas zdan motivasi z zpihak zlain zuntuk zbertindak zdalam zrangka zmemuaskan zkebutuhan dan zkepentingan zseseorang zsebagaimana zdisepakati zbersama zsecara zimplesit 133 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce maupun zeksplesit z(Sheth zdan zMittal, z2004). zTrust zatau zkepercayaan dipandang sebagai zpersepsi zakan zkehandalan zdari zsudut zpandangzkonsumen zdidasarkan pada zpengalaman, zatau zlebih zpada zurutan-urutan ztransaksi zatau zinteraksi yangzdicirikan zoleh zterpenuhinya zharapan zakan zkinerja zproduk zdan kepuasaan.zGuillen zet zal. z(2003), zmendefinisikan ztrust z(kepercayaan zmerek) sebagai zsuatu zperasaan zaman zyang zdimiliki zkonsumen zakibat zdari zinteraksinya dengan zsebuah zmerek, zyang zberdasarkan zpersepsi zbahwa zmerek tersebut dapat zdiandalkan zdan zbertanggung zjawab zatas zkepentingan zdan zkeselamatan dari zkonsumen. Pengalaman zmenjadikan zsumber zbagi zkonsumen zbagi zterciptanya suatuzkepercayaan zpada zsuatu zproduk zatau zmerek zyang ztimbul zsecara zalami yang zbisa zdisebabkan zkarena zkeinginan zyang zdimiliki zkonsumen zhanya zbisa dijawab zoleh zproduk zatau zmerek ztersebut. zIni zyang zmembuat zkepercayaan zitu tumbuh zsehingga zkonsumen ztidak zakan zberalih zlagi zkeproduk zlain zkarena kepercayaan zakan zmerek zdan zproduk ztersebut zmemiliki zkemampuan zuntuk menjawab zkebutuhan zkonsumen zbahkan zmemberikan zlebih zdari zkonsumen butuhkan. zKepercayaan zadalah zrasa zyang ztimbul zsecara zalami zberdasarkan pengalaman zseseorang zyang zdimana zproduk zatau zjasa zyang zdipilih zoleh konsumen zdapat zdiandalkan zdan zdapat zdipertanggungjawabkan. Pengaruh zpurchase zintention terhadap zpurchase zdecision Minat zbeli zmampu zmenghasilkan zkeputusan zpembelian zdari zminat zyang dialami zoleh zkonsumen. zPada ztahapan ztimbulnya zminat, zkonsumen zmenyadari bahwa zmereka zmenyukai zproduk ztertentu zdan zingin zmemiliki zproduk ztersebut sehingga zapabila zkeyakinan zterhadap zproduk zpositif zmaka zakan zmenimbulkan keputusan zuntuk zmelakukan zpembelian z(Endhar zet zal., z2016). zAri zet zal (2013) menyimpulkan zbahwa zminat zbeli zmerupakan zvariabel zyang zmempengaruhi keputusan zpembelian. zHasil zyang zserupa zjuga zdisimpulkan zoleh zpenelitian zTri et zal z(2013). zBerdasarkan zuraian ztersebut zmaka zhipotesis zpenelitian: 134 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 zanggota zkomunitas zmereka zlebih zbisa zdiandalkan dan dapat zmemperlihatkan zevaluasi zyang zpositif. zOleh zkarena zitu zanggota zyang memiliki zrasa zmemiliki zkomunitas zdan zmemahami zdirinya zadalah zbagian zdari komunitas zakan zmemiliki zrasa zkepercayaan zyang zlebih ztinggi zkepada komunitas zdan zanggota zdi zdalamnya. Kepercayaan zDengan zdemikian ztingkat zidentitas sosial diyakini zmemiliki zpengaruh zterhadap ztingkat zkepercayaan zyang zdimiliki anggota zkomunitas z(Sony, z2016). zPenelitian zRahmania z(2015) zmenyimpulkan bahwa zterdapat zpengaruh zpositif zidentitas zsosial zterhadap zkepercayaan. zCindy (2014)zmenyimpulkan zbahwa ztidak zterdapat zpengaruh zidentitas zsosial terhadapkepercayaan. zBerdasarkan zuraian ztersebut zmaka zhipotesis zpenelitian: Hipotesis z3: zTerdapat zpengaruh zsocial zidentity zterhadap ztrust Pengaruh zEWOM zterhadap ztrust EWOM mempengaruhi z zkepercayaan konsumen zdalam zmembuat keputusan pembelian. Z Semakin positif informasi zyang disebarluaskan zdari konsumen kepada konsumen zlainnya maka zakan menimbulkan ztingkat kepercayaan yang baik. zKonsumen cenderung mempercayai zopini zdari konsumen lain zyang telahzmengkonsumsi produk. zOpini yang zpositif cenderung meningkatkan kepercayaan konsumen terhadap zkinerja zproduk (Syafaruddin zet zal., z2016). Penelitian Annisa zdan zTony z(2016) menemukan zbahwa zEWOM berpengaruh terhadap ztrust. zHasil zini sesuai zdengan zpenelitian zHatane dan Adi z(2014). Berdasarkan zuraian tersebut zmaka zhipotesis zpenelitian: Hipotesis z2: zTerdapat zpengaruh EWOM zterhadap ztrust Pengaruh zsocial zidentity zterhadap trust Identitas z zsosial zpada zkomunitas Pengaruh zEWOM zterhadap ztrust EWOM mempengaruhi z zkepercayaan konsumen zdalam zmembuat keputusan pembelian. Z Semakin positif informasi zyang disebarluaskan zdari konsumen kepada konsumen zlainnya maka zakan menimbulkan ztingkat kepercayaan yang baik. zKonsumen cenderung mempercayai zopini zdari konsumen lain zyang telahzmengkonsumsi produk. zOpini yang zpositif cenderung meningkatkan kepercayaan konsumen terhadap zkinerja zproduk (Syafaruddin zet zal., z2016). Penelitian Annisa zdan zTony z(2016) menemukan zbahwa zEWOM berpengaruh terhadap ztrust. zHasil zini sesuai zdengan zpenelitian zHatane dan Adi z(2014). Berdasarkan zuraian tersebut zmaka zhipotesis zpenelitian: Hipotesis z2: zTerdapat zpengaruh EWOM zterhadap ztrust Pengaruh zsocial zidentity zterhadap trust Identitas sosial pada komunitas Pengaruh zperceived zrisk zterhadap ztrust Pengaruh zperceived zrisk zterhadap ztrust Persepsi zrisiko zmempengaruhi zkepercayaan zkonsumen. zsemakin ztinggi zrisiko dalam zmelakukan zpembelian zonline zmaka zkonsumen z Persepsi zrisiko zmempengaruhi zkepercayaan zkonsumen. zsemakin ztinggi zrisiko dalam zmelakukan zpembelian zonline zmaka zkonsumen z Identitas z zsosial zpada zkomunitas yang zkuat zdapat zmengarahkan individu kepada zanggapan zbahwa 135 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce zpenting zdalam zpemrosesan zinformasi zyang zdilakukan zoleh konsumen. zKonsumen zterdorong zuntuk zmencari ztambahan zinformasi zketika dihadapkan zpada zpembelian zproduk zdengan zrisiko ztinggi. zPersepsi zrisiko zyang rendah zakan zmenyebabkan ztingginya zminat zuntuk zbertransaksi zsecara zonline, dan zjuga zsebaliknya zpersepsi zrisiko zyang ztinggi zakan zmenyebabkan zrendahnya minat zuntuk zbertransaksi zsecara zonline z(Ratna zet zal, z2016). zPenelitian zHana, et zal. Kepercayaan z(2016) zmenyimpulkan zbahwa zterdapat zpengaruh znegatif zperceived zrisk terhadap zpurchase zintention zsesuai zdengan zpenelitian zArif zdan zNoor z(2018). Berdasarkan zuraian ztersebut zmaka zhipotesis zpenelitian: akan zsemakin ztidak percaya zterhadap zsitus zjual zbeli zonline ztersebut. zBeberapa zrisiko zyang zada dalam zmelakukan zpembelian zonline zseperti zresiko zterhadap zmetode pembayaran, zresiko zterhadap zpenyalahgunaan zinformasi zdata zpribadi, zdan zjuga risiko zterhadap zketidaksesuaian zbarang zyang zditerima zdengan zbarang zyang dipasang zdi zsitus zjual zbeli z(Sanda, z2017). zPenelitian zAwliya zet zal z(2014) menemukan zbahwa zterdapat zpengaruh znegatif zperceived zrisk zterhadap ztrust sejalan zdengan zhasil zpenelitian zHana, zet zal. z(2016). zBerdasarkan zuraian tersebut zmaka zhipotesis zpenelitian: Hipotesis z4: zTerdapat zpengaruh zperceived zrisk zterhadap ztrust Hipotesis z5: zTerdapat zpengaruh zperceived zrisk zterhadap zpurchase zintention. Hipotesis z5: zTerdapat zpengaruh zperceived zrisk zterhadap zpurchase zintention. Pengaruh zperceived zrisk zterhadap zpurchase zintention Semakin ztinggi zpersepsi zrisiko zmakazakan zmenyebabkan zseseorang mempunyai ketakutan zlebih ztinggi zsaat zbertransaksi zonline, zbegitu zjuga zsebaliknya. Persepsi zterhadap zrisiko zdiprediksi zakan zberpengaruh znegatif zterhadap zniat individu zuntuk zmenggunakan zonline zshop. zPersepsi zrisiko zmenjadi zsalah zsatu komponen Pengaruh ztrust zterhadap zpurchase zintention Trust zyang zsemakin ztinggi ztentu zakan zdapat zdijadikan zukuran untukzmenumbuhkan zpurchase zintention zsecara zonline. zSemakin ztinggi trustzmaka zsemakin ztinggi zpurchase zintention. zTrust zsecara z Pengaruh ztrust zterhadap zpurchase zintention Pengaruh ztrust zterhadap zpurchase zintention Trust zyang zsemakin ztinggi ztentu zakan zdapat zdijadikan zukuran untukzmenumbuhkan zpurchase zintention zsecara zonline. zSemakin ztinggi trustzmaka zsemakin ztinggi zpurchase zintention. zTrust zsecara z 136 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 positif mempengaruhi zniat zuntuk zberbelanja zsecara zonline zkarena zkonsumen zyakin bahwa zperusahaan zmampu zmenjalankan zkegiatan zonline-nya z(karena kompetensi) zdan zdapat zmengirimkan zproduk-produk zyang zdibeli zkepada konsumen. zJika zkonsumen zmempercayai zonline zstore zyang zdisediakan zoleh perusahaan, zmaka zhal ztersebut zmemungkinkan zmereka zmeningkatkan zniatnya untuk zmelakukan zpembelian zsecara zonline. zPemahaman zini zsecara zumum mengontrol ztransaksi zonline zyang zberpengaruh zpositif zterhadap zniat zkonsumen untuk zmelakukan zpembelian z(Dwi zdan zMahendra, z2013). zPenelitian zArif zdan Noor z(2018) zmenyimpulkan zbahwa zterdapat zpengaruh zpositif zdan zsignifikan trust zterhadap zpurchase zintention. zHasil zini zsesuai zdengan zpenelitian zAntoni dan zMeyzi z(2017). zBerdasarkan zuraian ztersebut zmaka zhipotesis zpenelitian: Hipotesis z6: zTerdapat zpengaruh ztrust zterhadap zpurchase zintention Hipotesis z6: zTerdapat zpengaruh ztrust zterhadap zpurchase zintention zdecision. zMetode pengumpulan zdata zmenggunakan zkuesioner zdan zmetodezanalisis zdata menggunakan zpartial zleast zsquare. zSampel zpenelitian zsebanyak z100 responden yang zmerupakan zkonsumen ze- commerce zX zselama zperiode zbulan zDesember 2020. Kepercayaan zPengambilan z100 zresponden zini zdidasarkan zoleh zpendapat zAgus z(2015) yang zmenyatakan zbahwa zdalam zpenggunaan zpartial zleast zsquare zsebaiknya sampel zyang zdigunakan zminimal zsebanyak z100 zobservasi METODOLOGI zPENELITIAN z Penelitian zini zdidesain zuntuk zmenguji zkebenaran zempiris zantara zdua zatau lebih variabel zpenelitian zyang ztelah zdirumuskan zdalam zhipotesis. zVariabelzpenelitian yang zdigunakan zadalah zsocial zidentity, zEWOM, zperceived zrisk, ztrust,zpurchase intention, zdan zpurchase zdecision. zVariabel zeksogen zdalam zpenelitian zini adalah social zidentity, zEWOM, zdan zperceived zrisk. z zVariabel zendogen zdalam penelitian zini zadalah zpurchase zintention zdan zpurchase 137 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengujian zValiditas dengan 0,3. zSedangkanzdiscriminant zvalidity zdilihat zdari znilai zAVE zdan zdinyatakan valid zapabila zlebih zbesar zatau zsama zdengan z0,5 z(Agus, z2015). zHasil pengujian zvaliditas zkonvergen zdapat zdilihat zpada ztabel zberikut Pengujian zValiditas Menurut zAgus z(2015) zpengujian zvaliditas zkonvergen zyang zdikoreksi dinyatakan zvalid zapabila znilai zloading zfactor zlebih zbesar zatau zsama z Menurut zAgus z(2015) zpengujian zvaliditas zkonvergen zyang zdikoreksi dinyatakan zvalid zapabila znilai zloading zfactor zlebih zbesar zatau zsama z Tabel z1 Output zSmartpls zuntuk zUji zValiditas Tabel z1 Output zSmartpls zuntuk zUji zValiditas Butir pernyataan EWOM memiliki nilai loading factor tertinggi 0,922 dan terendah 0,768. Butir pernyataan perceived risk memiliki nilai loading Butir pernyataan EWOM memiliki nilai loading factor tertinggi 0,922 dan terendah 0,768. Butir pernyataan perceived risk memiliki nilai loading dan terendah 0,768. Butir pernyataan perceived risk memiliki nilai loading Butir pernyataan EWOM memiliki nilai loading factor tertinggi 0,922 130 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce factor tertinggi 0,836 dan terendah 0,736. Butir pernyataan purchase decision memiliki nilai loading factor tertinggi 0,966 dan terendah 0,827. Butir pernyataan purchase intention memiliki nilai loading factor tertinggi 0,952 dan terendah 0,726. Butir pernyataan social identity memiliki nilai loading factor tertinggi 0,857 dan terendah 0,788. Butir pernyataan trust memiliki nilai loading factor tertinggi 0,958 dan terendah 0,746. Setelah diuji convergent validity selanjutnya diuji discriminant validity dengan melihat nilai AVE (Agus, 2015). Tabel 2 Average Variance Extracted (AVE) dan Cronbach Alpha Variabel AVE Cronbach's zAlpha EWOM 0,728 0,873 Perceived zRisk 0,637 0,717 Purchase zDecision 0,848 0,908 Purchase zIntention 0,77 0,844 Social zIdentity 0,684 0,771 Trust 0,785 0,858 Gambar 1 Output SmartPLS Tabel 2 Average Variance Extracted (AVE) dan Cronbach Alpha Tabel 2 Average Variance Extracted (AVE) dan Cronbach Alpha Variabel AVE Cronbach's zAlpha EWOM 0,728 0,873 Perceived zRisk 0,637 0,717 Purchase zDecision 0,848 0,908 Purchase zIntention 0,77 0,844 Social zIdentity 0,684 0,771 Trust 0,785 0,858 Gambar 1 Output SmartPLS Gambar 1 Output SmartPLS Gambar 1 Output SmartPLS Gambar 1 Output SmartPLS 130 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Tabel 3 Rangkuman Pengujian Hipotesis Tabel 3 Rangkuman Pengujian Hipotesis Berdasarkan tabel 3 nilai p-value untuk hipotesis 1 sebesar sebesar 0,000 < 0,05 artinya terdapat pengaruh purchase intention terhadap purchase decision. Nilai p-value untuk hipotesis 2 zsebesar 0,000 < 0,05 artinya terdapat pengaruh EWOM terhadap trust. Nilai p-value untuk hipotesis 3 sebesar sebesar 0,003 < 0,05 artinya terdapat pengaruh social identity terhadap Berdasarkan tabel 3 nilai p-value untuk hipotesis 1 sebesar sebesar 0,000 < 0,05 artinya terdapat pengaruh purchase intention terhadap purchase decision. Nilai p-value untuk hipotesis 2 zsebesar 0,000 < 0,05 artinya terdapat pengaruh EWOM terhadap trust. Nilai p-value untuk hipotesis 3 sebesar sebesar 0,003 < 0,05 artinya terdapat pengaruh social identity terhadap trust. Nilai p-value untuk hipotesis 4 zsebesar 0,108 > 0,05 artinya tidak terdapat pengaruh perceived risk terhadap trust Nilai p-value untuk hipotesis 5 sebesar sebesar 0,112 > 0,05 artinya tidak terdapat pengaruh perceived risk terhadap purchase intention. Nilai p-value untuk hipotesis 6 sebesar sebesar 0,112 < 0,000 artinya terdapat pengaruh trust terhadap purchase intention trust. Nilai p-value untuk hipotesis 4 zsebesar 0,108 > 0,05 artinya tidak terdapat pengaruh perceived risk terhadap trust Nilai p-value untuk hipotesis 5 sebesar sebesar 0,112 > 0,05 artinya tidak terdapat pengaruh perceived risk terhadap purchase intention. Nilai p-value untuk hipotesis 6 sebesar sebesar 0,112 < 0,000 artinya terdapat pengaruh trust terhadap purchase intention mereka zmenyukai zproduk ztertentu zdan zingin zmemiliki zproduk ztersebut sehingga zapabila zkeyakinan zterhadap zproduk zpositif zmaka zakan zmenimbulkan keputusan zuntuk zmelakukan zpembelian. zMeyediakan zproduk zyang zbanyak diminati zdan zdibutuhkan zkonsumen zdapat zmeningkatkan zminat zmembeli sehingga zdapat zmeningkatkan zpeluang zkeputusan zmembeli. Terdapat zpengaruh EWOM zterhadap ztrust. zHasil zini zsesuai zdengan SIMPULAN Terdapat zpengaruh zpurchase zintention zterhadap zpurchase zdecision. zHasil zini sesuai zdengan zpenelitian zEndhar zet zal z(2016), zAri zet zal z(2013), zdan zTri zet zal (2013).zMinat zbeli zmampu zmenghasilkan zkeputusan zpembelian zdari zminat yang dialami zoleh zkonsumen.zPada ztahapan ztimbulnya zminat, zkonsumen zmenyadari bahwa z Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 positif.zOleh zkarena zitu zanggota zyang zmemiliki zrasa zmemiliki zkomunitas zdan memahami zdirinya zadalah zbagian zdari zkomunitas zakan zmemiliki zrasa kepercayaan zyang zlebih ztinggi zkepada zkomunitas zdan zanggota zdi dalamnya.zDengan zdemikian ztingkat zidentitas zsosial zdiyakini zmemiliki pengaruh zterhadap ztingkat zkepercayaan zyang zdimiliki zanggota zkomunitas. Menyediakan zproduk zyang zdibutuhkan zkomunitas ztertentu zmerupakan zsuatu cara zmeningkatkan zidentitas zsosial zterhadap zkepercayaan. zE-commerce zyang menyediakan zhanya zproduk zbranded zdapat zmeyakinkan zkonsumen zbahwa produk zyang zdijualnya zmerupakan zproduk zpremium. Tidak zterdapat zpengaruh perceived zrisk zterhadap ztrust. zHasil zini zsesuai zdengan zpenelitian zSanda (2017), Awliya zet zal z(2014), zdan zHana, zet zal.z(2016). zHasil zini ztidak signifikan disebabkan zoleh zkonsumen zyang zsudah zberpengalaman zdalam melakukan transaksi zpada ze- commerce zsehingga zsudah zterbentuk zkepercayaan terhadap ze-commerce. zUmumnya zmemperkecil zresiko zpenelitian zTony z(2016), zdan Hatane zdan zAdi z(2014).zEWOM zmempengaruhi zkepercayaan zkonsumen zdalam membuat zkeputusan zpembelian.zSemakin zpositif zinformasi zyang zdisebarluaskan dari zkonsumen zkepada zkonsumen zlainnyazmaka zakan zmenimbulkan ztingkat kepercayaan zyang zbaik.zKonsumen zcenderung zmempercayai zopini zdari konsumen zlain zyang ztelah zmengkonsumsi zproduk.zOpini zyang zpositif cenderung zmeningkatkan zkepercayaan zkonsumen zterhadap zkinerja zproduk. EWOM zmewakili zopini zkonsumen. zOleh zkarena zitu zmeningkatkan zcitra positif produk zdapat zmeningkatkan zreview zdan ztestimony zpositif zyang zdapat meyakinkan zkonsumen zterhadap zsuatu zproduk. Terdapat zpengaruh zsocial identity zterhadap ztrust. zHasil zini ztidak zsesuai zdengan zpenelitian zSony z(2016), Rahmania z(2015), zdan zCindy z(2014).zIdentitas zsosial zpada zkomunitas zyang kuat zdapat zmengarahkan zindividu zkepada zanggapan zbahwa zanggota zkomunitas mereka zlebih zbisa zdiandalkan zdan zdapat 140 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce zAntoni zdan zMeyzi z(2017). zTrust zyang zsemakin ztinggi tentu akan zdapat zdijadikan zukuran zuntuk zmenumbuhkan zpurchase zintention zsecara online.zSemakin ztinggi ztrust zmaka zsemakin ztinggi zpurchase zintention.zTrust secara zpositif zmempengaruhi zniat zuntuk zberbelanja zsecara zonline zkarena konsumen zyakin zbahwa zperusahaan zmampu zmenjalankan zkegiatan zonline-nya (karena zkompetensi) zdan zdapat zmengirimkan zproduk-produk zyang zdibeli kepada zkonsumen.zJika zkonsumen zmempercayai zonline zstore zyang zdisediakan oleh zperusahaan, zmaka zhal ztersebut zmemungkinkan pembelian zdapat meningkatkan kepercayaan zkonsumen zterhadap suatu ze-commerce. zPenyediaan layanan pengembalian zuang zapabila produk zyang zditerima ztidak zsesuai merupakan salah satu zcara zmemperkecil zresiko zyang zberdampak zkepada kepercayaan konsumen. Tidak zterdapat zpengaruh zperceived zrisk zterhadap zpurchase zintention. zHasil zini sesuai zdengan zpenelitian zRatna zet zalz(2016), zHana, zet zal.z(2016), zdan zArif dan Noor z(2018). Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 zHasil zini ztidak zsignifikan zkarena zpromosi zharga zyang ditawarkan serta zproduk zyang zsangat zmenarik zsehingga zkonsumen ztidak zterlalu memperhatikan zresiko zyang zada. zKonsumen zpada zdasarnya zselalu zmemikirkan resiko zdalam zmembeli zsuatu zproduk.Apabila ze-commerce zdapat zmemperkecil resiko zini zmaka zminat zbeli zpun zakan zmeningkat. zKeamanan zdalam bertransaksi zonline zmerupakan zsalah zsatu zcara zmemperkecil zresiko zdalam berbelanja zdi ze-commerce. Terdapat zpengaruh ztrust zterhadap zpurchase intention. zHasil zini zsesuai zdengan zpenelitian zDwi zdan zMahendra z(2013), zArif dan zNoor z(2018), zmerekazmeningkatkan niatnya zuntuk zmelakukan zpembelian zsecara zonline.zPemahaman zini zsecara umum zmengontrol ztransaksi zonline zyang zberpengaruh zpositif zterhadap zniat konsumen zuntuk zmelakukan zpembelian. zKepercayaan zkonsumen zmerupakan faktor zutama zdalam zpembelian zproduk. zMembuat zkonsumen zpercaya zterhadap suatu zproduk, zmerek, zkinerja zproduk zharus zdilakukan zoleh zsales. zWalaupun e-commerce zmenjual zproduk zsecara zonlineznamun 141 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 – 145 zproduk zdari zpihak zlain. zOleh zkarena zitu zpenggunaan ztenaga zpenjual untuk zmemasarkan zproduk zdapat zmeningkatkan zkepercayaan zkonsumen. pembuktian zataupun testimony tenaga penjualan zdapat zmeningkatkan kepercayaan zkonsumen. zKonsumen akan percaya zterhadap zsuatu zproduk apabila zkonsumen zsudah memperoleh informasi mengenai Awliya zA. zet zal. z(2014). zAnalisis zpengaruh zpersepsi zteknologi, DAFTAR zPUSTAKA Ari, zM. zA. zet zal. z(2013). zPeran zMinat zBeli zKonsumen zSebagai zMediasi zKualitas zProduk zterhadap zKeputusan zPembelian zpada zPerusahaan zKue. zBinus zBusiness zReview, zVol. z4, zNo. z1. zdoi: z zhttps://doi.org/10.21512/bbr.v 4i1.1125. Agus zWidarjono. z(2015). zAnalisis zstatistika zmultivariat zterapan. zYogyakarta: zUPP zSTIM zYKPN. Annisa zM. zA. zDan zTony zW. z(2016). zPengaruh zEWOM zterhadap zbrand zimage zdan zbrand ztrust, zserta zdampaknya zpada zminat zbeli zproduk zsmartphone ziphone z(studi zpada zmasyarakat zdi zYogyakarta. zJurnal zManajemen zBisnis zIndonesia, zVol. z5, zNo. z3. z zRetrieved zfrom zhttp://journal.student.uny.ac.i d/ojs/index.php/jmbi/article/vi ew/5030/4696. Arif zR. zM. zdan zNoor zK. z(2018). zPengaruh zkepercayaan, zkemudahan, zdan zpersepsi zrisiko zterhadap zkeputusan zpembelian zjasa zGojek zdi zkota zSemarang zyang zdimediasi zminat zbeli zsebagai zvariabel zintervening. zEkobis, zVol. z19, zNo. z2. zdoi:zhttp://dx.doi.org/10.3065 9/ekobis.19.2.203-213. Antoni zdan zMeyzi zH. z(2017). zPengaruh zbrand zimage zdan zkepercayaan zmerek zterhadap zminat zbeli zkonsumen zpada zproduk zmobilzToyota zEtios zValco. zJOM zFISIP, zVol. z4, zNo. z1. zRetrieved zfrom zhttps://jom.unri.ac.id/index.p hp/JOMFSIP/article/view/123 00/11945. Assael, zH. z(1998). zConsumer zBehavior zand zMarketing zAction z6th zEdition. zNew zYork: zSouth zWestern zCollege zPublishing. Assael, 142 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce zserta zdampaknya zpada zkeputusan zpembelian. zJurnal zAdministrasi zBisnis, zVol. z38, zNo. z2. zRetrieved zfrom zhttp://administrasibisnis.stude ntjournal.ub.ac.id/index.php/j ab/article/view/1525/1909. zserta zdampaknya zpada zkeputusan zpembelian. zJurnal zAdministrasi zBisnis, zVol. z38, zNo. z2. zRetrieved zfrom zhttp://administrasibisnis.stude ntjournal.ub.ac.id/index.php/j ab/article/view/1525/1909. zpersepsi zresiko zterhadap zkepercayaan zdan zdampaknya zterhadap zkepuasan zbelanja zonline zmahasiswa zdi zPekanbaru. zJurnal zEkonomi, zVol. z22, zNo. z3. zdoi: zhttp://dx.doi.org/10.31258/je. 22.3.p.1-16. Engel, zJ. zF. zEt zal. z(1995). zPerilaku zKonsumen. zJakarta: zBina zRupa zAksara. Bernad, zM. zet zal z(2016). Social zIdentity zand zSocial zFree- Riding. zEuropean zEconomic zReview. zdoi: zhttp://dx.doi.org/10.1016/j.eu roecorev.2016.01.001. Featherman, zM. zS. zand zPavlou, z zP. zA. z(2002). zPredicting ze- service zAdoption: zA zPerceived zRisk zFacet zPerspective. zInternational zJournal zof zHuman-Computer zStudies, zVol. z59. zdoi: zhttps://doi.org/10.1016/S1071 -5819(03)00111-3. Cindy zF. zT. z(2014). zPengaruh zidentitas zmerek zterhadap zloyalitas zmerek zmelalui zcitra znerek zdan zkepercayaan zmerek zToyota. zJurnal zManajemen zPemasaran, zVol. z8, zNo. z2. zdoi: z10.9744/pemasaran.8.2.62- 69. Guillen, zM. zJ. zet zal. z(2003). z zDevelopment zand zValidation zofzBrand zTrust zScale. zInternational zJournal zof zMarket zResearch., zVol. z45, zNo. z1. zdoi: z zhttps://doi.org/10.1177%2F1 47078530304500103. Dholakia, zU. zM. zet zal. z(2004). zA zSocial zInfluence zModel zof zConsumer zParticipation zin zNetwork- zand zSmall-Group- Based zVirtual zCommunities. zInternational zJournal zof zResearch zinzMarketing, zVol. z21. zdoi: zhttps://doi.org/10.1016/j.ijres mar.2003.12.004. Goldsmith, zR. zE. zand zHorowitz, zD. z(2006). zMeasuring zMotivations zfor zOnline zOpinion zSeeking, zJournal zof zInteractive zAdvertising, zVol. z6, zNo. z2. zdoi: zhttps://doi.org/10.1080/15252 019.2006.10722114. Dwi zP. zJ. zA. zdan zMahendra zA. zN. zPengaruh ztrust zdan zperceived zof zrisk zterhadap zniat zuntuk zbertransaksi zmenggunakan ze-commerce. Jurnal zNominal, zVol. zII, zNo. z1. zdoi: zhttps://doi.org/10.21831/nomi nal.v2i1.1646. Hana zA. zet zal. z(2016). zPengaruh zpast zexperience zdan zperceived zrisk zterhadap zpurchase zintention zpada zsitus zRakuten.co.id zdi zJakarta zdengan zcustomer ztrust zsebagai zvariabel zmediator. zConference zon zManagement zand zBehavioral Endhar zW. zP. z(2016). zPengaruh zdiskon zterhadap zminat zbeli 143 Jurnal Bina Manajemen, Maret 2021, Vol.9 No.2 Hal 126 - 145 Studies. zRetrieved zfrom zhttp://repository.untar.ac.id/1 1931/. ntegrating_trust_and_risk_per ceptions_in_business-to consumer_electronic_commer ce_with_the_technology_acce ptance_model. Hatane zS. zdan zAdi zS. zL. z(2014). zAnlaisis zEWOM, zbrand zimage, zbrand ztrust zdan zminat zbeli zproduk zsmartphone zdi zSurabaya. zJurnal zManajemen zPemasaran, zVol. z8, zNo. z2. zdoi: z10.9744/pemasaran.8.2.47- 54. Malik, zM. zE. zet zal., z(2013). zImpact zof zBrand zImage zand zAdvertisement zon zConsumer zBuying zBehavior. zWorld zApplied zSciences zJournal, zVol. z23, zNo. z1. zdoi: z10.5829/idosi.wasj.2013.23.0 1.824. Oglethorpe z, zJ. zE. z zand zMonroe z, zK. zB. z(1994). Assael, zDeterminants zof zPerceived zHealth zand zSafety zRisks zof zSelected zHazardous zProducts zand zActivities. zThe zJournal zof zConsumer zAffairs, zVol. z28, zNo. z2. z zdoi: zhttps://doi.org/10.1111/j.1745 -6606.1994.tb00855.x. Hennig-Thurau, zT. zet zal. z(2004). zElectronic zword zof zmouth zvia zconsumer zopinion zplatforms: zWhat zmotivates zconsumers zto zarticulate zthemselves zon zthezinternet. zJournal zof zInteractive zMarketing, zVol. z18, zNo. z1. zdoi: zhttps://doi.org/10.1002/dir.10 073. Padilla, zA. zM. z zand zPerez z, zW. z(2003). zAcculturation, zSocial zIdentity, zand zSocial zCognition: zA zNew zPerspective. zHispanic zJournal zof zBehavioral zSciences., zVol. z25, zNo, z1. zdoi: z10.1177/0739986303251694. Hogg, zM. zA. zand zReid, zS. zA. z(2006). zSocial zIdentity, zSelf- Categorization, zand zthe zCommunication zGroup zNorms. zCommunication zTheory, zVol. z16. zdoi:10.1111/j.1468- 2885.2006.00003.x. Kotler, zP. zand zKeller, zK. zL. z(2006). zMarketing zManagement. z zNew zJersey: zPearson. Rahmania zA. z(2015). zPengaruh zdari zidentity zterhadap zbrand zvalue, zsatisfaction, ztrust zand zbrand zloyalty. zJurnal zManajemen zdan zPemasaran zJasa, zVol. z8, zno. z2. zRetrieved zfrom zhttps://trijurnal.lemlit.trisakti. ac.id/jasa/article/view/1593/1 383. Rahmania zA. z(2015). zPengaruh zdari zidentity zterhadap zbrand zvalue, zsatisfaction, ztrust zand zbrand zloyalty. zJurnal zManajemen zdan zPemasaran zJasa, zVol. z8, zno. z2. zRetrieved zfrom zhttps://trijurnal.lemlit.trisakti. ac.id/jasa/article/view/1593/1 383. Lui,zH.zK. and JamiesonzR.z(2003). zIntegrating zTrust zand zRisk zPerception zin zBusiness zto zConsumer zElectronic zCommerce zwith zTechnology zAcceptance zModel. zECIS z2003 zProceedings. zRetrieved zfromzhttps://www.researchgat e.net/publication/221408744_I 144 Pengaruh Social Identity, EWOM, Perceived Risk, Dan Trust Terhadap Purchase Intention Dan Dampaknya Terhadap Purchase Decision Pada E-Commerce t_and_Online_Helping_Behav ior_on_Facebook- based_Community. Ratna zM. zR. zet zal. z(2016). zPengaruh zkepercayaan zterhadap zminat zbeli zmelalui zpersepsi zrisiko zpada ztransaksi zjual zbeli zonline zmelalui zmedia zsosial. zJurnal zEkonomi zBisnis, zVol. z21, zNo. z2. zRetrieved zfrom zhttp://journal.um.ac.id/index. php/ekobis/article/view/7433/3 384. Suresh, zA. zM. zand zShashikala, zR. z(2011). zIdentifying zFactors zof zConsumer zPerceived zRisk ztoward zOne-commerce zin zIndia. z z3rd zInternational zConference zon zInformation zand zFinancialzEngineering zIPEDR, zVol.12. z zRetrieved zfrom http://www.ipedr.com/vol12/6 1-C137.pdf. Sanda zK. zS. z(2017). zPengaruh ze- commerce zknowledge, zrisiko, zdan zteknologi zterhadap zkepercayaan zdan zniat zbeli zonline. zJurnal zEkonomi zManajemen zSumber zDaya, zVol. z19, zNo. z2. zDOI: zhttps://doi.org/10.23917/daya saing.v19i2.5382. Syafaruddin zZ. z(2016). zPengaruh zkomunikasi zelectronic zword zof zmouth zterhadap zkepercayaan z(trust) zdan zniat zbeli z(purchase zintention) zserta zdampaknya zpada zkeputusan zpembelian. zJurnal zBisnis zdan zManajemen, zVol. z3, zNo. z1. zRetrieved zfrom zhttps://media.neliti.com/medi a/publications/75271-ID- none.pdf. Schiffman z& zKanuk. z(2004). zPerilaku zKonsumen z(edisi z7). zJakarta: zPrentice zHall. Sheth, zJ.N. zand zMittal, zB. z(2004) zCustomer zBehavior: zA zManagerial zPerspective. z2nd zEdition. zNew zYork: zSouth- Western. Tajfel, zH. z(1982). zSocial zPsychology zof zIntergroup zRelations. z zAnnual zReview zPsychology, zVol. z33, zNo. z1. Assael, zdoi: zhttps://doi.org/10.1146/annur ev.ps.33.020182.000245. Tajfel, Sony zK. z(2016). zSocial zidentity, zgroup znorm, ztrust, zand zonline zhelping zbehavior zon zFacebook-based zcommunity. zJurnal zIlmiah zManajemen, zVol. zVI, zNo. z2. zRetrieved zfrom https://www.researchgate.net/ publication/309662556_Social _Identity_Group_Norms_Trus 145
https://openalex.org/W2892252239	http://old.scielo.br/pdf/cadsc/v26n3/1414-462X-cadsc-1414-462X201800030075.pdf	Portuguese	null	Qualidade das notificações de violências interpessoal e autoprovocada no Estado do Rio de Janeiro, Brasil, 2009-2016	Cadernos Saúde Coletiva	2,018	cc-by	6,077	DOI: 10.1590/1414-462X201800030075 DOI: 10.1590/1414-462X201800030075 Artigo Original Resumo Introdução: Os dados dos Sistemas de Informação em Saúde são muito utilizados para a produção do conhecimento. Para a sua efetiva utilização, no entanto, é essencial assegurar que sejam válidos e confiáveis. Objetivo: Avaliar a qualidade das notificações de violências no estado do Rio de Janeiro, visando contribuir para o aprimoramento e melhoria da informação. Método: Estudo ecológico sobre a consistência das notificações de violência registradas no Sistema de Informação de Agravos de Notificação do Estado do Rio de Janeiro, entre 2009 a 2016. A tendência temporal das proporções de inconsistências, para cada grupo de campos relacionados, foi avaliada por regressão linear simples e respectiva significância estatística (p≤0,05). Resultados: No período foram notificados 103.841 casos de violência, mas 31,7% não foram identificados se interpessoal ou autoprovocada. As principais inconsistências foram: incongruência entre violência psicológica e meio da agressão, sexo e identidade de gênero, tipo de violência e idade da vítima, violência autoprovocada e características do agressor e tipo de violência; a maioria com estabilidade no período (p>0,05). Conclusão: Os resultados evidenciam inconsistências expressivas nas notificações de violência que comprometem a utilização da informação. Palavras-chave: sistema de informação em saúde; doenças e agravos de notificação compulsória; violência doméstica; tentativa de suicídio; direitos humanos. Quality of the notifications of interpersonal and self-inflicted violence in the state of Rio de Janeiro, Brazil, 2009-2016 Quality of the notifications of interpersonal and self-inflicted violence in the state of Rio de Janeiro, Brazil, 2009-2016 Vania Reis Girianelli1, Aldo Pacheco Ferreira1, Marcos Besserman Vianna1, Nair Teles1, Regina Maria de Carvalho Erthal1, Maria Helena Barros de Oliveira1 ▄ ▄MÉTODO Trata-se de estudo ecológico sobre a qualidade das notificações de violências no estado do Rio de Janeiro no período de 2009 a 2016. Os dados foram obtidos do Sinan, disponibilizado com acesso livre no site da Secretaria de Estado de Saúde do Rio de Janeiro. Foram elegíveis para o estudo apenas as violências notificadas nos municípios do estado. As violências autoprovocadas e interpessoais só passaram a integrar a LNNC em 20114. A notificação de violência interpessoal abrange doméstica/intrafamiliar; sexual; tráfico de pessoas; trabalho escravo; trabalho infantil; tortura; intervenção legal; violências homofóbicas; e também violência extrafamiliar/comunitária, mas restrita a populações vulneráveis (crianças, adolescentes, mulheres, idosos, deficientes, indígenas, lésbicas, gays, bissexuais, travestis e transexuais)5. Posteriormente, as notificações de violência sexual e de tentativa de suicídio passaram a ser de notificação imediata no nível municipal6, visando garantir o acesso às medidas de prevenção e assistência em tempo oportuno. Na penúltima versão da ficha de notificação7 foram incluídos os campos “orientação sexual”, “identidade de gênero” e “motivo da violência”; mas estes dados, por enquanto, só estão disponíveis no site do estado do Rio de Janeiro. A dimensão da qualidade analisada foi apenas a consistência interna12,13, que aprecia a existência de coerência entre os campos. Esta avaliação baseou-se na recomendação do instrutivo de preenchimento de notificação de violência preconizado pelo Ministério da Saúde14. Os campos avaliados foram: idade, sexo, gestação, escolaridade, situação conjugal/estado civil, orientação sexual, identidade de gênero, lesão autoprovocada, motivação da violência, tipo de violência, meio de agressão, tipo de violência sexual, procedimento realizado, sexo do provável agressor, número de agressores envolvidos e vínculo com a pessoa atendida. A investigação consistiu em identificar a existência de inconsistência entre os campos relacionados, sendo classificada em: A notificação de violências, no entanto, já havia sido implantada anteriormente, como previsto nas legislações (Estatuto da Criança e do Adolescente - ECA8, Estatuto do Idoso9, violência contra à mulher10 e prevenção de suicídio11), mas era restrito a serviços de referência (Doenças Sexualmente Transmissíveis/Síndrome da Imunodeficiência Adquirida – DST/Aids, ambulatórios especializados, maternidades, entre outros) definidos pelas secretarias municipais e estaduais de saúde. Cabe ressaltar que em 2009 a notificação de violências foi inserida no Sinan, mas de forma gradual conforme a adesão dos estados e municípios. a) Não conformidade - preenchimento inadequado, mas que não compromete totalmente a compreensão ou análise dos dados. Abstract Background: Data from Health Information Systems are widely used for the production of knowledge. For their effective use, however, it is essential to ensure that they are valid and reliable. Objective: Evaluate the quality of reports of violence in the state of Rio de Janeiro aiming to contribute to the improvement of information. Method: Ecological study on the consistency of notifications of violence recorded to the Information System on Diseases of Compulsory Declaration in the state of Rio de Janeiro from 2009 to 2016. The temporal trend of the proportions of inconsistencies, for each group of related fields, was evaluated by simple linear regression and its statistical significance (p≤0.05). Results: During the study period, 103,841 cases of violence were reported, but in 31.7% of them it was not possible to identify whether they were interpersonal or self-inflicted. The main inconsistencies were incongruence between psychological violence and aggression, gender identity and sex, type of violence and age of the victim, self-inflicted violence and characteristics of the aggressor and type of violence; most of them presented stability in the period (p>0.05). Conclusion: The results show significant inconsistencies in the reports of violence that compromise the use of this information. Keywords: health information systems; disease notification; domestic violence; attempted suicide; human rights. 1 Departamento de Direitos Humanos, Saúde e Diversidade Cultural (DIHS), Escola Nacional de Saúde Pública Sérgio Arouca (ENSP), Fundação Oswaldo Cruz (FIOCRUZ) - Rio de Janeiro (RJ), Brasil. Trabalho realizado na Escola Nacional de Saúde Pública Sérgio Arouca (ENSP) – Rio de Janeiro (RJ), Brasil. Endereço para correspondência: Vania Reis Girianelli – Departamento de Direitos Humanos, Saúde e Diversidade Cultural (DIHS), Escola Nacional de Saúde Pública Sérgio Arouca (ENSP), Fundação Oswaldo Cruz (FIOCRUZ), Avenida Brasil, 4036, Sala 905 – Manguinhos – CEP: 521040-361 – Rio de Janeiro (RJ), Brasil – Email: vaniagirianelli@yahoo.com.br Fonte de financiamento: nenhuma. Conflito de interesses: nada a declarar. Este é um artigo publicado em acesso aberto (Open Access) sob a licença Creative Commons Attribution, que permite uso, distribuição e reprodução em qualquer meio, sem restrições desde que o trabalho original seja corretamente citado. 318 Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 Qualidade das notificações de violência no Rio de Janeiro ▄ ▄INTRODUÇÃO Os dados dos SIS são muito utilizados para a produção do conhecimento. Para a sua efetiva utilização, no entanto, é essencial assegurar que sejam válidos e confiáveis. A ficha de notificação de violências, além de ser a mais recente a ser implantada, é a mais complexa do sistema. Nesta ficha foram unificadas diversas fichas de notificação existentes anteriormente (tentativa de suicídio, maus tratos da criança e adolescente, maus tratos do idoso e violência contra mulher), e também contém campos que envolvem conceitos atualmente discutidos na sociedade como identidade de gênero, comportamento sexual e sexismo; cujos significados podem ainda não ser de domínio dos profissionais que atendem à demanda, e consequentemente levar a erros de classificação. O Sistema de Informação em Saúde (SIS) do Brasil é constituído por vários subsistemas que contemplam dados referentes a mortalidade, nascimento, morbidade e assistência à saúde, que são fundamentais para formulação e avaliação das políticas, planos e programas de saúde, subsidiando o processo de tomada de decisões. Os dados dos SIS são disponibilizados no site do Departamento de Informática do Sistema Único de Saúde (DataSUS) do Ministério da Saúde (MS), e também no site institucional de alguns estados e municípios1.i O Sistema de Informação de Agravos de Notificação (Sinan) é um dos subsistemas do SIS que foi implantado na década de 1990. Este sistema é alimentado, principalmente, com dados que constam da ficha de notificação individual de doenças e agravos, incluídos na lista nacional de notificação compulsória (LNNC)2. Esta lista é composta de doenças e agravos relevantes tendo em vista a magnitude, potencial de disseminação, transcendência, vulnerabilidade, compromissos internacionais e regulamento sanitário internacional. Os estados e municípios também podem acrescentar, na sua região, outras questões de saúde pública relacionadas ao contexto local3. O presente estudo tem como objetivo avaliar a qualidade das notificações de violência no estado do Rio de Janeiro, visando contribuir para o aprimoramento e melhoria da informação. ▄ ▄MÉTODO Situação em que a classificação correta de um campo seria “Não se Aplica”, mas é classificado como “ignorado”, “não”, ou o campo não é preenchido. b) Incongruência: classificação de dois campos distintos de forma que não sejam simultaneamente verdadeiros, comprometendo a compreensão e consequentemente a análise de dados. Situação, por exemplo, em que uma vítima adulta tem a violência sofrida classificada como pornografia infantil, ou seja, um dos campos foi preenchido incorretamente. Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 319 eis Girianelli, Aldo Pacheco Ferreira, Marcos Besserman Vianna, Nair Teles, Regina Maria de Carvalho Erthal, Maria Helena Barros de Oliveira O percentual de inconsistência foi calculado considerando no numerador a quantidade fichas com inconsistência para determinadas categorias dos campos relacionados, conforme exemplificado anteriormente, e no denominador apenas o quantitativo de fichas de notificação que continham as categorias dos campos sob análise. As inconsistências relativas a não conformidade foram classificadas como baixa quando menor de 10%, moderada entre 10% a 30% e alta quando maior que 30%. Já as incongruências foram classificadas como baixa quando menor que 0,5%, moderada entre 0,5% a 10% e alta quando maior que 10%. no período para os campos sexo do agressor (β = 0,34; p = 0,004) e número de agressores envolvidos (β = 0,26; p = 0,020). A mediana das não conformidades foi moderada, variando de 11,2% para estupro a 14,3% para exploração sexual, com redução estatisticamente significativa no período para a maioria dos campos avaliados (p < 0,05). O campo atentado violento ao pudor não consta na nova versão da ficha de notificação, tendo sido incorporado ao estupro, de forma a manter consonância com a nova legislação16. Este campo, portanto, foi avaliado do período de 2009 a 2014, tendo apresentado baixa incongruência (Mediana = 0,2%) e moderada não conformidade (Mediana = 17,5%); ambos com tendência estável no período (β = -1,60, p = 0,504 e β = -0,11, p =0,453; respectivamente). Calculou-se o percentual de inconsistência anual e a mediana para o período, para cada grupo de campos relacionados. A tendência temporal foi avaliada por regressão linear simples, tendo como variável independente o ano de notificação e como variável dependente o percentual de inconsistência. O aumento ou declínio da proporção de inconsistência no período foi avaliado com base no coeficiente de regressão (β) e respectiva significância estatística (p ≤ 0,05). As análises foram realizadas utilizando o programa estatístico R. ▄ ▄MÉTODO Em relação à notificação de violência interpessoal foram notificados mais de um tipo por vítima (dados não mostrados), embora a recomendação seja registrar apenas o principal14. A inconsistência foi uniforme no período avaliado (p > 0,05). A Tabela 3 apresenta as inconsistências relacionadas a violência interpessoal do tipo psicológica. A proporção de incongruências variou de moderada a alta (Mediana > 0,5%), exceto para os campos agressão por envenenamento e exploração sexual. Os campos agressão por força corporal ou espancamento (β = 0,30; p = 0,001), agressão por enforcamento (β = 1,22; p = 0,033) e agressão por objeto contundente (β = 1,79; p = 0,048) apresentaram aumento estatisticamente significativo da incongruência no período. A mediana da inconsistência por não conformidade variou de 16,6% (estupro) a 29,3% (exploração sexual), todos com declínio no período (p < 0,05). O estudo foi realizado utilizando dados secundários, de acesso público, sem identificação individual das informações. Assim, consoante às recomendações da Resolução do Conselho Nacional de Saúde (CNS) nº 466, de 12 de dezembro de 2012, foram respeitados os princípios éticos de pesquisas que envolvem seres humanos. ▄ ▄RESULTADOS No período avaliado foram notificados 103.841 casos de violência pelos municípios do estado do Rio de Janeiro, sendo 62,5% violência interpessoal e 5,7% autoprovocada, os demais foram classificados como ignorado ou não preenchidos (31,7%). Dados sobre motivo da violência estão disponíveis apenas a partir de 2013. Neste período ocorreram 1.104 notificações de violência interpessoal por motivo de sexismo. Destas, 43 (3,9%) eram vítimas do sexo masculino (dados não mostrados). A incongruência entre a idade e outros campos foi baixa (< 0,2%) em todos os anos (Tabela 1). Já a inconsistência relacionada a não conformidade foi moderada entre “idade” e “pornografia infantil” (Mediana = 29,7%) e baixa entre idade e situação conjugal (Mediana = 5,2%). O campo “trabalho infantil” não tem a opção de categorização “Não se Aplica” e não teve nenhum registro classificado como “Ignorado”. Análise complementar foi realizada considerando a idade de 14 anos ou mais em relação ao trabalho infantil, devido a possibilidade de trabalho como aprendiz15; e também para 18 anos ou mais no que se refere à pornografia infantil, tendo em vista a classificação de maioridade no ECA8. Os resultados, no entanto, foram similares (dados não mostrados). Dados sobre identidade de gênero estão disponíveis a partir de 2014. Neste período ocorreram 198 notificações de violência de travestis e transexuais (Tabela 4). Destas, 53,5% correspondiam a mulheres transexuais, cujo o sexo biológico era feminino, 12,6% a homens transexuais com sexo biológico masculino, e duas notificações de travestis com sexo biológico feminino. A incongruência entre os campos identidade de gênero e sexo biológico foi muito alta (62,7%). Não houve inconsistência entre sexo biológico masculino e gestação, e entre os seguintes procedimentos: coleta de secreção vaginal, contracepção de emergência ou aborto previsto em lei (dados não mostrados). A Tabela 2 apresenta as inconsistências relacionadas às notificações de violência autoprovocada. A proporção de incongruência foi moderada ou alta para maioria dos campos avaliados (> 0,5%), com aumento estatisticamente significativo Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 320 Qualidade das notificações de violência no Rio de Janeiro Tabela 1. Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 321 ▄ ▄RESULTADOS Percentual de inconsistência das notificações de violência do Estado do Rio de Janeiro relacionadas a idade da vítima - 2009 a 2016a Campos de checagem 2009 2010 2011 2012 2013 2014 2015 2016 Total Mediana β p N % N % N % N % N % N % N % N % N % <6 anos e escolaridade Ensino médio incomp.b 0 0,0 0 0,0 0 0,0 0 0,0 0 0,0 1 0,0 0 0,0 0 0,0 1 0,0 0,0 - - Ignoradoc 0 0,0 1 0,1 1 0,1 3 0,1 1 0,0 2 0,1 3 0,1 0 0,0 11 0,1 0,1 0,00 1,000 Total < 6 anos 291 - 847 - 1653 - 2084 - 2329 - 3371 - 2480 - 519 - 13574 - - - - <10 anos e situação conjugal Solteirob 3 0,8 0 0,0 0 0,0 1 0,0 0 0,0 1 0,0 0 0,0 0 0,0 5 0,0 0,0 -5,00 0,134 Ignoradoc 16 4,0 77 6,5 145 6,8 161 5,8 114 3,7 108 2,4 154 4,7 57 8,0 832 4,6 5,2 0,08 0,892 Total < 10 anos 398 - 1180 - 2139 - 2769 - 3082 - 4450 - 3295 - 709 - 18022 - - - - ≥16 anos e trabalho infantil Simb 1 0,1 2 0,1 7 0,1 10 0,1 8 0,1 15 0,1 10 0,1 2 0,0 55 0,1 0,1 -40,00 0,134 Total ≥16 anos 1314 - 3828 - 6452 - 9198 - 13177 - 19848 - 15711 - 4160 - 73688 - - - - ≥20 anos e pornografia infantil Simb 1 0,1 1 0,0 0 0,0 3 0,0 2 0,0 3 0,0 4 0,0 1 0,0 15 0,0 0,0 -40,00 0,134 Não/Ignoradoc 217 20,2 1254 38,5 1985 37,0 3141 41,7 4228 38,9 3780 22,5 1297 9,9 480 13,6 16382 26,6 29,7 -0,11 0,165 Total ≥20 anos 1072 - 3256 - 5367 - 7535 - 10881 - 16828 - 13045 - 3525 - 61509 - - - - aDados de 2014 a 2016 são preliminares, com situação em 03/06/2016, sujeitos a retificação e recebimento de informações de novas notificações; bIncongruência; cNão conformidade Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 321 Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 321 eis Girianelli, Aldo Pacheco Ferreira, Marcos Besserman Vianna, Nair Teles, Regina Maria de Carvalho Erthal, Maria Helena Barros de Oliveira Tabela 2. ▄ ▄RESULTADOS Percentual de inconsistência das notificações de violência autoprovocada do Estado do Rio de Janeiro - 2009 a 2016a Campos de checagem 2009 (N = 96) 2010 (N = 180) 2011 (N = 468) 2012 (N = 742) 2013 (N = 838) 2014 (N = 1240) 2015 (N = 1824) 2016 (N = 568) Total (N = 5956) Mediana β p N % N % N % N % N % N % N % N % N % Violência físicab 51 53,1 111 61,7 261 55,8 506 68,2 603 72,0 812 65,5 801 43,9 195 34,3 3340 56,1 58,7 -0,08 0,291 Violência psicológicab 20 20,8 41 22,8 86 18,4 124 16,7 119 14,2 230 18,5 185 10,1 43 7,6 848 14,2 17,5 -0,42 0,004 Torturab 2 2,1 9 5,0 11 2,4 9 1,2 13 1,6 21 1,7 21 1,2 3 0,5 89 1,5 1,6 -1,23 0,061 Violência sexualb 4 4,2 12 6,7 16 3,4 18 2,4 14 1,7 19 1,5 24 1,3 8 1,4 115 1,9 2,0 -1,07 0,012 Tráfico de seres humanosb 0 0,0 0 0,0 1 0,2 0 0,0 0 0,0 3 0,2 1 0,1 0 0,0 5 0,1 0,0 4,25 0,707 Violência financeirab 1 1,0 2 1,1 4 0,9 1 0,1 3 0,4 4 0,3 7 0,4 1 0,2 23 0,4 0,4 -5,00 0,019 Negligência/abandonob 21 21,9 19 10,6 39 8,3 52 7,0 59 7,0 80 6,5 99 5,4 23 4,0 392 6,6 7,0 -0,35 0,014 Trabalho infantilb 0 0,0 1 0,6 1 0,2 1 0,1 2 0,2 1 0,1 3 0,2 0 0,0 9 0,2 0,1 -4,31 0,416 Intervenção legalb 0 0,0 0 0,0 4 0,9 1 0,1 2 0,2 3 0,2 0 0,0 0 0,0 10 0,2 0,1 -1,53 0,651 Agressão por ameaçab 1 1,0 16 8,9 40 8,5 66 8,9 50 6,0 75 6,0 60 3,3 12 2,1 320 5,4 6,0 -0,22 0,494 Sexo dos prováveis agressoresb 72 75,0 146 81,1 411 87,8 671 90,4 756 90,2 1132 91,3 1710 93,8 524 92,3 5422 91,0 90,3 0,34 0,004 Número agressores envolvidos b 72 75,0 154 85,6 441 94,2 711 95,8 772 92,1 1190 96,0 1743 95,6 550 96,8 5633 94,6 94,9 0,26 0,020 Agressor c/suspeita uso álcoolb 28 29,2 37 20,6 99 21,2 179 24,1 179 21,4 272 21,9 329 18,0 108 19,0 1231 20,7 21,3 -0,50 0,046 Vínculo do agressorb 91 94,8 103 57,2 241 51,5 346 46,6 342 40,8 564 45,5 483 26,5 128 22,5 2298 38,6 46,1 -0,10 0,003 Assédio sexual Simb 1 1,0 2 1,1 4 0,9 3 0,4 3 0,4 4 0,3 4 0,2 1 0,2 22 0,4 0,4 -6,08 0,001 Não/Ignoradoc 14 14,6 40 22,2 89 19,0 205 27,6 107 12,8 97 7,8 20 1,1 7 1,2 579 9,7 13,7 -0,19 0,031 Estupro Simb 3 3,1 9 5,0 11 2,4 12 1,6 9 1,1 13 1,0 10 0,5 7 1,2 74 1,2 1,4 -1,33 0,017 Não/Ignoradoc 10 10,4 35 19,4 82 17,5 196 26,4 101 12,1 88 7,1 14 0,8 1 0,2 527 8,8 11,2 -0,18 0,069 Pornografia infantil Simb 0 0,0 0 0,0 1 0,2 0 0,0 0 0,0 1 0,1 0 0,0 0 0,0 2 0,0 0,0 -3,87 0,782 Não/Ignoradoc 15 15,6 42 23,3 92 19,7 208 28,0 110 13,1 100 8,1 24 1,3 8 1,4 599 10,1 14,4 -0,19 0,025 Exploração sexual Simb 0 0,0 2 1,1 3 0,6 0 0,0 1 0,1 1 0,1 2 0,1 0 0,0 9 0,2 0,1 -2,91 0,092 Não/Ignoradoc 15 15,6 40 22,2 90 19,2 208 28,0 109 13,0 100 8,1 22 1,2 8 1,4 592 9,9 14,3 -0,19 0,027 aDados de 2014 a 2016 são preliminares, com situação em 03/06/2016, sujeitos a retificação e recebimento de informações de novas notificações; bIncongruência; cNão conformidade Cad. ▄ ▄RESULTADOS Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 322 Qualidade das notificações de violência no Rio de Janeiro Tabela 3. Percentual de inconsistência das notificações de violência psicológica do Estado do Rio de Janeiro - 2009 a 2016a Campos de checagem 2009 (N = 165) 2010 (N = 891) 2011 (N = 1987) 2012 (N = 2760) 2013 (N = 3378) 2014 (N = 5800) 2015 (N = 4407) 2016 (N = 1285) Total (N = 20673) Mediana β p N % N % N % N % N % N % N % N % N % Meio de Agressão Forçab 101 61,2 519 58,2 1342 67,5 1911 69,2 2409 71,3 4204 72,5 3421 77,6 1051 81,8 14958 72,4 70,3 0,30 0,001 Enforcamentob 12 7,3 56 6,3 186 9,4 258 9,3 299 8,9 484 8,3 401 9,1 145 11,3 1841 8,9 9,0 1,22 0,033 Obj.contundenteb 9 5,5 38 4,3 130 6,5 188 6,8 229 6,8 372 6,4 279 6,3 96 7,5 1341 6,5 6,5 1,79 0,048 Obj. Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 323 ▄ ▄DISCUSSÃO digitação no caso de notificação de vítima do sexo biológico feminino, e na categoria “homem transexual” para vítima do sexo masculino. Os resultados observados são preocupantes, especialmente em relação a incongruência que inviabiliza a utilização da base de dados. Wyatt & Wyatt17 destacam a necessidade da realização de avaliações ao longo de todo o processo de qualquer sistema de informação. Desde a implantação do Sinan têm sido realizadas sucedidas modificações estruturais para melhorar a adequação da informação disponibilizada, de forma a atender às normas e regulamentos definidos para o Sistema Único de Saúde (SUS), e de prover respostas às demandas da população. Já na migração da base de dados do Sinan DOS para o Sinan Windows foram incluídos procedimentos para identificar as principais inconsistências na base de dados2. A complexidade da ficha de notificação de violência, no entanto, requer um aperfeiçoamento das estratégias de críticas na entrada de dados, de forma a proporcionar informações válidas. É possível que alguns problemas tenham sido ocasionados por erros de digitação e outros pela própria estruturação dos campos na ficha. Assim, no campo tipo de violência, por exemplo, tem a opção de marcar: “sim”, “não” e “ignorado” para cada tipo; o que induz o profissional a registrar mais de um tipo de violência, embora a recomendação seja registrar apenas o tipo principal. No caso de violência psicológica, por exemplo, em que o agressor também utilizou de um instrumento para atingir a vítima apenas a violência física deveria ser registrada; e no caso de estupro a prioridade poderia ser o registro de violência sexual. Não obstante, isto possa ser difícil de definir em alguns casos. Na situação de tráfico de seres humanos com violência sexual, por exemplo, qual seria o tipo de violência principal? Por outro lado, registrar apenas o principal tipo de violência seria a melhor estratégia para conhecer as características deste fenômeno na nossa sociedade? A maioria da incongruência observada poderia ser evitada com programação, considerando a orientação do próprio instrutivo de preenchimento da ficha de notificação14. O campo lesão autoprovocada deveria ser de digitação obrigatória, o que evitaria a perda de um terço da base de dados, como observado neste estudo. Adicionalmente, deveria ter preenchimento automático com “Não se aplica” para todos os campos referentes à violência interpessoal e no campo tipo de violência, apenas a opção “outros” deveria ficar ativa, para atender a recomendação do instrutivo. ▄ ▄RESULTADOS pérfuro-cortanteb 11 6,7 56 6,3 148 7,4 220 8,0 287 8,5 424 7,3 317 7,2 103 8,0 1566 7,6 7,4 1,88 0,153 Subst/obj.quenteb 3 1,8 13 1,5 15 0,8 33 1,2 35 1,0 43 0,7 32 0,7 7 0,5 181 0,9 0,9 -4,88 0,003 Envenenamentob 1 0,6 0 0,0 8 0,4 5 0,2 13 0,4 7 0,1 17 0,4 7 0,5 58 0,3 0,4 1,02 0,841 Arma de fogob 17 10,3 38 4,3 72 3,6 130 4,7 219 6,5 475 8,2 183 4,2 77 6,0 1211 5,9 5,4 -0,20 0,653 Outro meiob 8 4,8 77 8,6 120 6,0 158 5,7 255 7,5 347 6,0 241 5,5 53 4,1 1259 6,1 5,9 -0,65 0,356 Assédio sexual Simb 15 9,1 40 4,5 95 4,8 137 5,0 140 4,1 213 3,7 133 3,0 63 4,9 836 4,0 4,6 -0,87 0,079 Não/Ignoradoc 58 35,2 371 41,6 552 27,8 914 33,1 774 22,9 949 16,4 389 8,8 106 8,2 4113 19,9 25,3 -0,19 0,001 Estupro Simb 46 27,9 146 16,4 229 11,5 354 12,8 469 13,9 521 9,0 395 9,0 125 9,7 2285 11,1 12,2 -0,31 0,018 Não/Ignoradoc 33 20,0 266 29,9 418 21,0 698 25,3 445 13,2 647 11,2 127 2,9 44 3,4 2678 13,0 16,6 -0,21 0,006 Pornografia Infantil Simb 1 0,6 5 0,6 21 1,1 27 1,0 16 0,5 32 0,6 15 0,3 8 0,6 125 0,6 0,6 -2,87 0,086 Não/Ignoradoc 72 43,6 406 45,6 626 31,5 1024 37,1 898 26,6 1134 19,6 507 11,5 161 12,5 4828 23,4 29,0 -0,18 0,001 Exploração sexual Simb 3 1,8 2 0,2 10 0,5 24 0,9 19 0,6 15 0,3 8 0,2 3 0,2 84 0,4 0,4 -2,87 0,086 Não/Ignoradoc 70 42,4 408 45,8 637 32,1 1027 37,2 895 26,5 1150 19,8 514 11,7 166 12,9 2824 13,7 29,3 -0,18 0,001 aDados de 2014 a 2016 são preliminares, com situação em 03/06/2016, sujeitos a retificação e recebimento de informações de novas notificações; bIncongruência; cNão conformidade Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 323 eis Girianelli, Aldo Pacheco Ferreira, Marcos Besserman Vianna, Nair Teles, Regina Maria de Carvalho Erthal, Maria Helena Barros de Oliveira Tabela 4. ▄ ▄RESULTADOS Percentual de inconsistência das notificações de violência do Estado do Rio de Janeiro sofridas por travestis e transexuais por sexo biológico - 2014 a 2016a Campos de checagem 2014 (N = 27) 2015 (N = 121) 2016 (N = 50) Total (N = 198) N % N % N % N % Sexo Feminino e Travesti 0 0,0 2 1,7 0 0,0 2 1,0 Sexo Feminino e Mulher transexual 16 59,3 65 53,7 25 50,0 106 53,5 Sexo Masculino e Homem transexual 3 11,1 17 14,0 5 10,0 25 12,6 Total 19 70,4 84 69,4 30 60,0 133 67,2 aDados de 2014 a 2016 são preliminares, com situação em 03/06/2016, sujeitos a retificação e recebimento de informações de novas notificações tência das notificações de violência do Estado do Rio de Janeiro sofridas por travestis e transexuais por sexo ▄ ▄DISCUSSÃO Ainda são poucos os estudos sobre consistência das bases de dados. Em estudo de revisão dos artigos que avaliaram a qualidade das informações dos SIS18 foram analisadas as dimensões: acessibilidade, clareza metodológica, cobertura, completitude, confiabilidade, consistência, não-duplicidade, oportunidade e validade. Dos 78 artigos incluídos no estudo apenas um avaliou consistência19, mas não se referia à consistência interna entre os campos do sistema de informação. Este estudo estimou inconsistência na taxa de mortalidade infantil dos estados, capitais e macrorregiões do país; com base na comparação das estimativas oficiais fornecidas pelo IBGE e dados oriundos das bases do sistema de informação de mortalidade (SIM) e de nascidos vivos (Sinasc).i Os erros também poderiam ser minimizados se a categoria “Não se Aplica” fosse incluída nos campos: tipo de violência, ciclo de vida do provável autor da violência, número e sexo dos agressores envolvidos, vínculo de parentesco e violência relacionada ao trabalho; e com preenchimento automático nos casos de tentativa de suicídio e autoagressão. A mesma estratégia poderia ser implantada nos campos escolaridade para vítimas menores de seis anos, situação conjugal para menores de dez anos, trabalho infantil para maiores de quinze anos, pornografia infantil para maiores de dezenove anos, e para os campos restritos ao sexo biológico feminino (gestante, coleta de secreção vaginal, aborto, etc.) em que a vítima é do sexo masculino. No campo identidade de gênero, as categorias “travesti” e “mulher transexual” não deveriam ficar ativas para Em levantamento realizado, foram identificados apenas dois estudos que avaliaram consistência, e a propósito nos dados da ficha de notificação de violência20,21. Um estudo analisou as notificações de Recife20 no período de 2009 a 2012, tendo encontrado uma média de 1% de inconsistência entre os campos avaliados, sendo que a maior proporção (10,1%) ocorreu entre os campos violência psicológica e meio de agressão. O outro restringiu a avaliação às notificações de violência sexual contra Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 324 Qualidade das notificações de violência no Rio de Janeiro Os problemas encontrados sinalizam a necessidade de treinamento dos profissionais que atuam no nível local em relação ao instrutivo de preenchimento da ficha de notificação, bem como, dos conceitos que são abordados para propiciar uma classificação adequada. Ademais, os gestores municipais precisam criar estratégias para revisar os dados, identificar as inconsistências e corrigi-las antes de encaminhar os dados para o nível estadual. ▄ ▄REFERÊNCIAS 10. Brasil. Lei federal nº 10.778, de 24 de novembro de 2003. Estabelece a notificação compulsória, no território nacional, do caso de violência contra a mulher que for atendida em serviços de saúde públicos ou privados. Diário Oficial da União, Brasília, 25 de novembro de 2003; Seção 1:11. 1. Lima CRA, Leal CD, Dias EP, Gonzalez FL, Santos HL, Silva MEM, et al. Departamento de Informática do SUS – DATASUS: a experiência de disseminação de informações em saúde. In: Ministério da Saúde. A experiência brasileira em sistemas de informação em saúde: produção e disseminação de informações sobre saúde no Brasil. Brasília: Organização Pan-Americana da Saúde, Fundação Oswaldo Cruz; 2009. p. 109-28. (Série B Textos Básicos de Saúde, vol. 1). 11. Brasil. Ministério da Saúde. Portaria nº 1876, de 14 de agosto de 2006. Institui Diretrizes Nacionais para Prevenção do Suicídio, a ser implantadas em todas as unidades federadas, respeitadas as competências das três esferas de gestão. Diário Oficial da União, Brasília, 15 de agosto de 2006; Seção 1:65. 2. Souza WV, Domingues CMAS. Notificação compulsória de doenças e agravos no Brasil: um breve histórico sobre a criação do Sistema de Informação de Agravos de Notificação – SINAN. In: Ministério da Saúde. A experiência brasileira em sistemas de informação em saúde: produção e disseminação de informações sobre saúde no Brasil. Brasília: Organização Pan-Americana da Saúde, Fundação Oswaldo Cruz; 2009. p. 39-48. (Série B Textos Básicos de Saúde, vol. 1). 2. Souza WV, Domingues CMAS. Notificação compulsória de doenças e agravos no Brasil: um breve histórico sobre a criação do Sistema de Informação de Agravos de Notificação – SINAN. In: Ministério da Saúde. A experiência brasileira em sistemas de informação em saúde: produção e disseminação de informações sobre saúde no Brasil. Brasília: Organização Pan-Americana da Saúde, Fundação Oswaldo Cruz; 2009. p. 39-48. (Série B Textos Básicos de Saúde, vol. 1). 12. Paim I, Nehmy RMQ, Guimarães C. Problematização do conceito de “qualidade” da informação. Perspect Cienc Inf. 1996;1:113-22. 13. Campbell SE, Campbell MK, Grimshaw JM, Walker AE. A systematic review of discharge coding accuracy. J Public Health Med. 2001;23(3):205-11. http://dx.doi.org/10.1093/pubmed/23.3.205. PMid:11585193. 3. Brasil. Secretaria de Vigilância em Saúde. Departamento de Vigilância Epidemiológica. Guia de vigilância epidemiológica. 7. ed. Brasília: Ministério da Saúde; 2009. 3. Brasil. Secretaria de Vigilância em Saúde. Departamento de Vigilância Epidemiológica. Guia de vigilância epidemiológica. 7. ed. Brasília: Ministério da Saúde; 2009. 14. Brasil. Secretaria de Vigilância em Saúde. ▄ ▄REFERÊNCIAS Departamento de Vigilância de Doenças e Agravos Não Transmissíveis e Promoção da Saúde. Viva: instrutivo de notificação de violência interpessoal e autoprovocada. 2. ed. Brasília: Ministério da Saúde; 2016. 4. Brasil. Ministério da Saúde. Portaria nº 104, de 25 de janeiro de 2011. Define as terminologias adotadas em legislação nacional, conforme disposto no Regulamento Sanitário Internacional 2005 (RSI 2005), a relação de doenças, agravos e eventos em saúde pública de notificação compulsória em todo o território nacional e estabelecer fluxo, critérios, responsabilidades e atribuições aos profissionais e serviços de saúde. Diário Oficial da União, Brasília, 26 de janeiro de 2011; Seção 1:37. 4. Brasil. Ministério da Saúde. Portaria nº 104, de 25 de janeiro de 2011. Define as terminologias adotadas em legislação nacional, conforme disposto no Regulamento Sanitário Internacional 2005 (RSI 2005), a relação de doenças, agravos e eventos em saúde pública de notificação compulsória em todo o território nacional e estabelecer fluxo, critérios, responsabilidades e atribuições aos profissionais e serviços de saúde. Diário Oficial da União, Brasília, 26 de janeiro de 2011; Seção 1:37. 15. Brasil. Lei federal nº 10.097, de 19 de agosto de 2000. Altera dispositivos da Consolidação das Leis do Trabalho - CLT, aprovada pelo Decreto-Lei no 5.452, de 1o de maio de 1943. Diário Oficial da União, Brasília, 20 de dezembro de 2000; Seção 1:2. 16. Brasil. Lei federal nº 12.015, de 7 de agosto de 2009. Altera o Título VI da Parte Especial do Decreto-Lei no 2.848, de 7 de dezembro de 1940 - Código Penal, e o art. 1º da Lei no 8.072, de 25 de julho de 1990, que dispõe sobre os crimes hediondos, nos termos do inciso XLIII do art. 5º. da Constituição Federal e revoga a Lei no 2.252, de 1º. De julho de 1954, que trata de corrupção de menores. Diário Oficial da União, Brasília, 10 de agosto de 2009; Seção 1:2. 5. Brasil. Secretaria de Vigilância em Saúde. Coordenação-Geral de Desenvolvimento da Epidemiologia em Serviços. Guia de vigilância em saúde. vol. 3. Brasília: Ministério da Saúde; 2017. 5. Brasil. Secretaria de Vigilância em Saúde. Coordenação-Geral de Desenvolvimento da Epidemiologia em Serviços. Guia de vigilância em saúde. vol. 3. Brasília: Ministério da Saúde; 2017. 6. Brasil. Ministério da Saúde. Portaria nº 1271, de 6 de junho de 2014. ▄ ▄DISCUSSÃO As informações inadequadas poderão influenciar as políticas de saúde e dificultar o enfrentamento das violências na sociedade. a mulher em Santa Catarina21 no período de 2008 a 2013, em que a média de inconsistência dos campos analisados foi de 1,1%. O denominador utilizado em ambos os estudos, no entanto, foi o total de fichas de notificação, ao invés do total de fichas com as categorias dos campos envolvidos na análise, o que pode ter subestimado este indicador. O total de fichas só deveria ser utilizado como denominador para descrever o percentual de fichas que ficaram comprometidas para análise em função dos erros encontrados. Infelizmente esta informação não está disponível nos estudos supracitados e não pode ser avaliada neste estudo, porque foram utilizados dados agregados. 10. Brasil. Lei federal nº 10.778, de 24 de novembro de 2003. Estabelece a notificação compulsória, no território nacional, do caso de violência contra a mulher que for atendida em serviços de saúde públicos ou privados. Diário Oficial da União, Brasília, 25 de novembro de 2003; Seção 1:11. 20. Abath MB, Lima MLLT, Lima OS, Silva MCM, Lima MLC. Avaliação da completitude, da consistência e da duplicidade de registros de violências do Sinan em Recife, Pernambuco, 2009-2012. Epidemiol Serv Saude. 2014;23(1):131-42. http://dx.doi.org/10.5123/S1679-49742014000100013. 21. Delziovo CR, Bolsoni CC, Lindner SR, Coelho EBS. Qualidade dos registros de violência sexual contra a mulher no Sistema de Informação de Agravos Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 326 ▄ ▄REFERÊNCIAS Define a lista Nacional de Notificação Compulsória de doenças, agravos e eventos de saúde pública nos serviços de saúde públicos e privados em todo território nacional, nos termos do anexo, e dá outras providências. Diário Oficial da União, Brasília, 9 de junho de 2014; Seção 1:67. 17. Wyatt JC, Wyatt SM. When and how to evaluate health information systems? Int J Med Inform. 2003;69(2-3):251-9. http://dx.doi.org/10.1016/ S1386-5056(02)00108-9. PMid:12810128. 7. Brasil. Secretaria de Vigilância em Saúde. Departamento de Vigilância de Doenças e Agravos Não Transmissíveis e Promoção da Saúde. Instrutivo de notificação de violência interpessoal e autoprovocada. Brasília: Ministério da Saúde; 2014. 18. Lima CR, Schramm JM, Coeli CM, Silva ME. Revisão das dimensões de qualidade dos dados e métodos aplicados na avaliação dos sistemas de informação em saúde. Cad Saude Publica. 2009;25(10):2095-109. http:// dx.doi.org/10.1590/S0102-311X2009001000002. PMid:19851611. 8. Brasil. Estatuto da criança e do adolescente: lei federal nº 8.069, de 13 de julho de 1990. Rio de Janeiro: Imprensa Oficial; 2002. 19. Szwarcwald CL, Leal MC, Andrade CL, Souza PR Jr. Estimação da mortalidade infantil no Brasil: o que dizem as informações sobre óbitos e nascimentos do Ministério da Saúde? Cad Saude Publica. 2002;18(6):1725-36. http:// dx.doi.org/10.1590/S0102-311X2002000600027. PMid:12488900. 9. Brasil. Estatuto do idoso: lei federal nº 10.741, de 01 de outubro de 2003. Brasília: Secretaria Especial dos Direitos Humanos; 2004. Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 325 Vania Reis Girianelli, Aldo Pacheco Ferreira, Marcos Besserman Vianna, Nair Teles, Regina Maria de Carvalho Erthal, Maria Helena Barro de Notificação (Sinan) em Santa Catarina, 2008-2013. Epidemiol Serv Saude. 2018;27(1):e20171493. http://dx.doi.org/10.5123/S1679-49742018000100003. PMid:29412348. Recebido em: Abr. 05, 2018 Aprovado em: Ago. 01, 2018 Recebido em: Abr. 05, 2018 Aprovado em: Ago. 01, 2018 Cad. Saúde Colet., 2018, Rio de Janeiro, 26 (3): 318-326 326
https://openalex.org/W2985448638	https://repositorioaberto.uab.pt/bitstream/10400.2/9230/1/2019_Discriminative%20characteristics%20of%20marginalized%20NPS%20users.pdf	English	null	Discriminative Characteristics of Marginalised Novel Psychoactive Users: a Transnational Study	International journal of mental health and addiction	2,019	cc-by	10,180	International Journal of Mental Health and Addiction https://doi.org/10.1007/s11469-019-00128-8 International Journal of Mental Health and Addiction https://doi.org/10.1007/s11469-019-00128-8 ORIGINAL ARTICLE * Zsolt Demetrovics demetrovics.zsolt@ppk.elte.hu Katalin Felvinczi1 & Annemieke Benschop2 & Róbert Urbán1 & Marie Claire Van Hout3 & Katarzyna Dąbrowska4 & Evelyn Hearne5 & Susana Henriques6 & Zsuzsa Kaló1 & Gerrit Kamphausen7 & Joana Paula Silva6 & Łukasz Wieczorek4 & Bernd Werse7 & Michal Bujalski4 & Zsolt Demetrovics1 & Dirk Korf2 Katalin Felvinczi1 & Annemieke Benschop2 & Róbert Urbán1 & Marie Claire Van Hout3 & Katarzyna Dąbrowska4 & Evelyn Hearne5 & Susana Henriques6 & Zsuzsa Kaló1 & Gerrit Kamphausen7 & Joana Paula Silva6 & Łukasz Wieczorek4 & Bernd Werse7 & Michal Bujalski4 & Zsolt Demetrovics1 & Dirk Korf2 Katalin Felvinczi1 & Annemieke Benschop2 & Róbert Urbán1 & Marie Claire Van Hout3 & Katarzyna Dąbrowska4 & Evelyn Hearne5 & Susana Henriques6 & Zsuzsa Kaló1 & Gerrit Kamphausen7 & Joana Paula Silva6 & Łukasz Wieczorek4 & Bernd Werse7 & Michal Bujalski4 & Zsolt Demetrovics1 & Dirk Korf2 # The Author(s) 2019 Extended author information available on the last page of the article * Zsolt Demetrovics demetrovics.zsolt@ppk.elte.hu Extended author information available on the last page of the article * Zsolt Demetrovics Abstract New psychoactive substances (NPS) continue to be considered as a major public health concern in many European countries. The study was implemented within the framework of a transnational project of six European countries (Germany, Hungary, Ireland, Netherlands, Poland, Portugal). Our aim here is to report on the distinct and differentiating characteristics of marginalised NPS users. Three subgroups of a total of 3023 adult NPS users (socially marginalised, night life, online community) were examined regarding their socio- demographic characteristics, substance use, and external motives towards NPS use. Poland and Hungary reported higher rates of NPS use in comparison to traditional controlled drugs. The external/contextual motives did not play a central role in the background of NPS use, the least important motives were alleged legality and non-detectability of these substances. Marginalised (defined as those accessing low threshold harm reduction services) users’ substance use patterns are different from the other two groups in terms of showing more intense and riskier drug use. The most important variables which contributed to be categorised as a marginalised NPS user were lower level education, being older, having an unfavourable labour market position and using drugs intravenously. Contextual motives did not play a decisive role in being categorised as a marginalised user when drug use pattern was controlled. These identified discriminative features of marginalised drug users should inform policy makers to develop and implement tailor-made interventions targeting this user group to successfully tackle the elevated public health concerns associated with NPS use. Keywords NPS use . User groups . Contextual motivation . Public health concern . Drug policy Extended author information available on the last page of the article International Journal of Mental Health and Addiction New psychoactive substances (NPS) are a major public health and policy concern in several parts of the world. NPS are defined as “synthetic or naturally occurring sub- stances that are not controlled under international law, and often produced with the intention of mimicking the effects of controlled drugs” (EMCDDA 2014, p.27). In general, NPS represent a multitude of synthetic and natural compounds marketed as alternatives to conventional illicit drugs (i.e. cannabis, heroin, cocaine, speed, ecstasy), the most popular being cathinone derivatives (i.e. mephedrone, 4-MEC, MDPV), other amphetamine-type substances, and synthetic cannabinoids, with the phenomenon ob- served since 2007 (Caudevilla 2016; EMCDDA 2016). Abstract The legislative procedures related to NPS are very much varied in the different EU member states, covering, and initiating a wide range of measures to tackle the problem. Neicun et al. (2019) concluded that in most countries the general psychoactive substance-related legislation is in place to establish a temporary or permanent ban on new psychoactive substances; the actual legal consequences related to personal use (possession, cultivation, purchasing) of these substances is depending on the general rigour or strictness of the penal code of the given country. The number of compounds labelled as NPS and observed through the Early Warning System (EWS) of the EMCDDA has grown from 13 (in 2004) to 650 (in 2016). General population-based prevalence estimates are rare and inconsistent with regard to trend monitoring of NPS use (Korf et al. 2019). According to the Eurobarometer (telephone surveys on drugs among young people aged 15–24 years in Europe), lifetime prevalence of NPS use increased from 5 to 8% between 2011 and 2014 (Eurobarometer 2011 and 2014). The most recent ESPAD study in 2015 reported 4% life time and 3% last year prevalence of NPS use among the 15–16-year-old European school population (ESPAD group 2016), with large differences among the different countries. In England and Wales, the 2010/2011 British Crime Survey identified a 0.6% lifetime ketamine use and 1.4% lifetime mephedrone use (Smith and Flatley 2011). The 2015 National Survey on Addiction Problems in Hungary found a 1.9%, 1.3% and 0.6% lifetime prevalence rate for synthetic cannabinoids, designer stimulants and mephedrone, respectively (Paksi et al. 2016). A national prevalence estimates among high-risk drug users in Belgium showed that the average reported use was 26%, in case of synthetic cannabinoid receptor agonists (SCRA) 19%, whilst for mephedrone 12.5% (European Monitoring Centre for Drugs and Drug Addiction 2017a). In the Czech Republic, one-third of the high-risk users had experience (LTP) with NPS (European Monitoring Centre for Drugs and Drug Addiction 2017b). Surveys targeting special (vulnerable) populations of NPS use are rare and struggle with methodological challenges; the available data suggest that these groups are more prone to NPS use, and show higher prevalence rates across all NPS types (MacLeod et al. 2016). Studies exploring the knowledge related to NPS in different countries and age groups revealed that the relevant knowledge is unevenly distributed among young respondents, with young people from rural areas showing less knowledge. Abstract These studies also suggested that there is a strong association between NPS use and the risk of binge drinking (Martinotti et al., 2015). The NPS situation has become even more challenging due to the emerging synthetic opioid crisis, which seems to be less alerting in most European countries than in the USA (Schifano et al. 2019); however, opiate use and related death rates are also increasing in the EU countries. The synthetic opioids— mostly fentanyl—play a significant role in those seeking specialised treatment in Europe (EMCDDA 2018). International Journal of Mental Health and Addiction Little knowledge is available about the specific motivations for NPS use and more in-depth research is needed to better understand these reasons (Moore et al. 2013; Korf et al. 2019). There are studies suggesting that the main motivations towards NPS use are curiosity, the enjoyable effects, or enhancement of social situations, similar to other conventional/classical drugs (Corazza et al. 2014; Measham et al. 2010; Werse and Morgenstern 2012; Winstock et al. 2015). Studies which have assessed the motivational background of NPS use have applied the four-factor motivation scale of Cooper (1994) which is used to measure alcohol consumption motivations, and was later further developed to measure cannabis (Simons et al. 1998) and amphetamine use related motivations (Thurn et al. 2017). These motivation scales are based on the Cox and Klinger’s Motivational Model (Cox and Klinger 1988) and concentrate on internal motivations such as coping, curiosity, conformity and enhancement. The internal mo- tivation of different user groups and the psychometric properties of a new motivation scale appropriate for cross national study of NPS motivation were also analysed based on the dataset of this research. Apart from identifying the psychometric properties and usability of the scale in a transnational setting and in a special population, it was found that marginalised users showed a higher endorsement in coping motives. The reported results also suggest that marginalisation can be considered as a cross-cultural factor behind instrumental substance use (Benschop et al. 2019 submitted). Studies assessing contextual and/or external (pragmatic) motivations towards NPS use, revealed that temporary legal status, easy availability, low price, high perceived or expected quality/purity of these substances might play an important role in predicting use of NPS (e.g., Werse and Morgenstern 2012; Barratt et al. 2013; Soussan et al. 2018; Sutherland et al., 2017). The aim of the present article is (i) to present some short term (last month, last year) frequencies of substance use; (ii) to identify the role of external/contextual motivation factors in NPS use of different user groups; and most importantly (iii) to identify a set of variables (discriminating features) which might contribute to being categorised as marginalised NPS user. Sample In most cases the socially marginalised users were recruited from the capital of the participat- ing country. The German marginalised sample consisted of, almost exclusively, persons from the Munich scene, which is probably the only urban hard drugs users’ setting with substantial (and mostly stimulant) NPS use in Germany. 2. Users in nightlife are recreational drug users who frequent clubs, raves and/or festi- vals. They were mainly recruited face-to-face on-site at clubs, raves and festivals. Users in night life mostly self-completed either a pen-and-paper or online question- naire to which they were referred by a flyer containing a link and an individual code; 3. Users in online communities are users who are very active on the internet, and actively participate in drug forums. They were recruited by posting messages on drug-related social media and internet forums (i.e. eve-rave.ch/forum; www.legal-high-inhaltsstoffe.de; www.daath.hu; www.drugs-forum.com; www.legalhighsforum.com). Users in online communities were only given access to the online questionnaire. Sample The work was undertaken as part of a large transnational interdisciplinary research project (“New Psychoactive Substances: transnational project on different user groups, user characteristics, extent and patterns of use, market dynamics, and best practices in prevention” [NPS-t]) spanning six European countries (Germany, Hungary, Ireland, Netherlands, Poland and Portugal). The data collection was carried out on a conve- nience sample of three different user groups supposed to have diverging consumption habits, motivations and socioeconomic background in both countries. Despite the high number of respondents (N = 3023), none of our findings can be generalised for the countries involved or for Europe as such. The survey was conducted between April and November 2016. The different user groups reached have an uneven representation in the sample; marginalised users proved to be hard to reach, mostly in Germany, the Netherlands and Portugal. International Journal of Mental Health and Addiction Three inclusion criteria were set as (1) recent NPS use (at least once in the past 12 months); (2) being resident of one of the participating countries; and (3) an age of 18 years or older. To reach the most heterogeneous sample of NPS users as possible, three subgroups were aimed to examine (for more details see Van Hout et al. 2018; Korf et al. 2019): 1. Socially marginalised users were defined by their (known and supposed) drug using habits on one hand and by their availability (low threshold, harm reduction services) and also by the venue of substance use, on the other. Marginalised users were reported as using opioids, (crack) cocaine and or (meth) amphetamine—often and/ or frequently intravenously or through smoking. They were recruited and interviewed face-to-face by trained fieldworkers or care professionals in the street, or through care and treatment facilities (e.g., drug services, shelters) and through snowball sampling; In most cases the socially marginalised users were recruited from the capital of the participat- ing country. The German marginalised sample consisted of, almost exclusively, persons from the Munich scene, which is probably the only urban hard drugs users’ setting with substantial (and mostly stimulant) NPS use in Germany. In most cases the socially marginalised users were recruited from the capital of the participat- ing country. The German marginalised sample consisted of, almost exclusively, persons from the Munich scene, which is probably the only urban hard drugs users’ setting with substantial (and mostly stimulant) NPS use in Germany. 1 The current paper is not presenting findings on all motivation related items; it is focusing exclusively on the newly added 6 items. The motivation scale without the newly added items (its psychometric properties and cross- cultural usability) was analysed in another article (Benschop et al. 2019, submitted). Measures The questionnaire contained items on demographics, routes of administration of sub- stances, motives and frequencies of NPS and other drug use, social and health problems related to it, modes of obtaining NPS, and possible perceived ways of tackling NPS problems. Seven categories of NPS were initially included in the study based on their epidemiological and clinical relevance: i.e., ‘herbal blends (e.g., ‘Spice’)’; ‘synthetic cannabinoids (obtained pure)’; ‘branded stimulants (e.g., “bath salts”)’; ‘stimulants/ empathogens/nootropics (obtained pure, e.g., mephedrone, MDPV, a-PVP)’; ‘psyche- delics (e.g., NBOMe-x, 2C-x’); ‘dissociatives (e.g., methoxetamine)’; and ‘other’. However, survey responses indicated that participants were generally unable to either properly categorise the NPS they had used or differentiate between certain categories (i.e. between herbal blends and synthetic cannabinoids obtained pure and similarly between branded stimulants and stimulants obtained pure) (Van Hout et al. 2018; Korf et al. 2019). Therefore, these uncertain categories were merged into the following categories: synthetic cannabinoids (including herbal blends names/ingredients); NPS International Journal of Mental Health and Addiction stimulants (including bath salts and ingredients). In case of NPS dissociatives and NPS psychedelics no merging of the categories was necessary. These NPS substance cate- gories were used during the process of statistical analysis and throughout the presen- tation of the results. The questionnaire also contained motivations related items1 to investigate what were the most important reasons to use NPS. These questions had to be answered on a 5-point Likert scale (1 = very unimportant; 5 = very important). The items of the questionnaire were based on the Marijuana Motivation Scale (MMM, Simons et al. 1998) and its further developed amended and psychometrically tested version (Benschop et al. 2015). Six items reflecting external/contextual motives unique for NPS use and not covered by the revised MMM were added: poor quality of other (traditional) drugs, price, alleged legality, expecting different or new experiences (regarding drug effects), non-detectability and low availability of other (traditional) drugs. Statistical Analysis During the statistical analysis, SPSS 25.0 programme was used. Descriptive statistics are presented on substance use frequencies and on external/contextual motives. Prin- cipal component analysis was carried out on the external motivation items to assess if these items construct a coherent scale. Variance analysis was carried out to assess the differences regarding external/contextual motives in different user groups and/or coun- tries, and finally, binary logistic regression models were developed to assess the possible factors contributing to be categorised as a marginalised user. 2 We deliberately chose last month frequency of use to present the differences among user groups and countries. The inclusion criteria for all the respondents were NPS use at least once in the preceding 12 months; consequently, the measured last year frequency of use was considerably high in all user groups. According to the recently introduced new “High Risk Drug Use Indicator” by the EMCDDA (2013) this “…..indicator area focuses on “High-risk drug use”, “recurrent drug use that is causing actual harms (negative consequences) to the person (including dependence, but also other health, psychological or social problems) or is placing the person at a high probability/risk of suffering such harms. High risk drug use is measured as the use of psychoactive substances by high risk pattern (e.g. intensively) and/or by high risk routes of administration….” EMCDDA, 2013, p.3.). As the main objective of this paper is to describe the discriminating features of marginalised users we wanted to capture the most visible elements of high-risk drug use, among them short-term (last month) use (as recurrent and intensified drug use) and the most harmful modes of administering substances (injecting). Descriptive Statistics—Sample Characteristics The number of respondents reached in the different countries and user groups varied significantly across the participating countries (see Table 1). The highest number of marginalised users was recruited in Hungary, whilst the highest number of online users was reached in the Netherlands. The most balanced sample (in number of respondents belonging to the different user groups) was achieved in Poland. As there was no way to carry out a probabilistic and random sample selection in these special populations none of the samples were representative of the participating countries. The sample was very much varied as far as the socio-demographic characteristics of the respondents are concerned. Most of them were male, and lived in larger cities, the most noticeable differences can be seen in the level of education (the lowest level education was observed among marginalised users and the highest in the on-line community) and in the employment conditions; most marginalised users were unemployed whilst night life and online community representatives were either students or full-time workers (for further details see Korf et al. 2019). International Journal of Mental Health and Addiction Table 1 Composition of the sample according to countries and user groups (number and percentage of respondents) Marginalised Nightlife Online Total Germany 23 (32.47%) 98 (14.78%) 542 (81.75%) 663 Hungary 101 (37%) 15 (5.51%) 156 (57.35%) 272 Ireland 48 (77.42%) 3 (4.84%) 11(17.74%) 62 The Netherlands 1(0.08%) 189 (17.04%) 1000 (90.17%) 1190 Poland 86 (17.43%) 172 (28.86%) 338 (56.71%) 596 Portugal 7 (2.92%) 170 (70.83%) 63 (26.25%) 240 TOTAL 266 (8.80%) 647 (21.40%) 2.110 (69.80%) 3.023 Short Term (Last Month2) Frequency of Substance Use in Different User Groups According to last month frequency of substance use, all user groups, regardless of their country of residence, are also using controlled substances; apart from Hungary and Poland, these frequencies are the highest in all user groups. Due to the fact that the different user groups were reached in an uneven proportion in the different countries, the following results would describe the short-term frequencies of substance use in different user groups. To be able to show the problematic nature of substance use, frequencies of intravenous drug use were also included into the analysis. Users in Online Community In the online NPS user group, after the most prevalent controlled substances, NPS stimulants rank second and NPS psychedelics the third place (43% and 17%, respective- ly). Synthetic cannabinoids are the fourth most popular NPS substance used in this group during the month preceding the data collection. Last month frequency of intravenous drug use is low, 1.20% in this user group (Table 2). Short-term frequencies of substance use show that NPS users do not form an independent group of substance users, traditional controlled substances still play a significant role in the poly-substance use patterns of NPS users. The last month frequencies of substance use, especially when we compare them with data stemming from general adult population surveys of the last 3 years, show that recent NPS users in the current sample are very intensive substance users. The frequencies of substance use estimates suggest that the most popular new psychoactive substances (synthetic cannabinoids, NPS stimulants) are 2–3 times more fre- quently used monthly among recent NPS consumers (the frequencies are between 33.80 and 45.50%) than any controlled substances in the general population of the same countries. For the most popular substances (cannabis and MDMA) the highest prevalence was 15.70% and 9.20% respectively (EMCDDA 2018). However, not just NPS are used heavily among recent NPS users, but the last month frequencies of substance use for all controlled substances is four to five times higher among recent NPS users than in the general population, as it is presented in the European Drug Report (EMCDDA 2018). Table 2 Last month frequencies of substance use (%) according to user groups User groups Marginalised N = 266 Night life N = 647 Internet N = 2110 χ2 (p) Substance categories Synthetic cannabinoids, N(%) 90 (33.8)a 78 (12.1)b 243 (11.5)b 101.8* NPS stimulants N(%) 121 (45.50)a 162 (25.00)b 914 (43.30)a 73.4* NPS psychedelics N(%) 15 (5.60)a 130 (20.10)b 366 (17.30)b 29.0*** NPS dissociatives N(%) 16 (6.00)a 21 (3.20)a 118 (5.60)a 6.1* All controlled substances N(%) 174 (66.40)a 533 (84.50)b 1712 (81.60)a 41.5* Intravenous drug use N(%) 133 (50.00)a 9 (1.40)b 26 (1.20)b 1084.1* Synthetic cannabinoids included synthetic cannabinoids pure, herbal brands; NPS stimulants included branded stimulants (e.g. “bath salts”), empathogens, nootropics, (obtained pure, e.g.. mephedrone, MDPV, a-PVP); NPS Psychedelics included e.g. NBOMe-x. 2C-x′; NPS dissociatives included, e.g. methoxetamine; controlled substances included cannabis, amphetamine, xtc, cocaine, crack-cocaine, heroin, magic mushroom, LSD, metamphetamine, ketamin, GHB. Marginalised Users Last month frequency of controlled substance use is lower among marginalised users than in the other two user groups (66.4% vs. above 80%). The most popular substances among new psychoactive substances are the NPS stimulants and the synthetic cannabinoids. The highest frequency of intravenous drug use was observed among this user group, half of them applied this route of administration in the month preceding the data collection. NPS psychedelic use is very rare among marginalised users, less than 6% of them used this type of substance in the last month. International Journal of Mental Health and Addiction Users in Night Life Controlled substances are the most prevalent substances in the night life user group (84.5%) as well, the second most prevalent substances in this user group are NPS stimulants and NPS psychedelics (25% and 20.1%, respectively). Synthetic cannabi- noids are also popular in this user group as every 9th respondents used this substance during the month before the data collection was carried out. Almost no intravenous drug use can be observed in this user group (1.40%). Perceived Context of NPS Use among Different User Groups The focus of our current analysis regarding motives was to find out if the external/pragmatic items added to the motivation scale can deepen our understanding regarding the motivation towards NPS use. The mean scores of the sample show that these external/contextual motives were rated 3.1 or less on the five-grade Likert scale. Items related to the legal status and to the perceived non- detectability of NPS were given ratings lower than 2. There are considerable differences among user groups; in general, marginalised users rate these items higher (consider them more important) than the other user groups with the only exception of “Expecting different or new experiences (regarding drug effects)”. This item is rated highest among night life users (see Table 3). Even though the analysis found relatively low ratings of external/contextual motives towards NPS use, during the next step, for methodological clarity, we studied if the newly included items of the motivation scale construct a coherent factor to measure external/contextual motivations towards NPS use. The principal component analysis on these items extracted only one component with the following items: ‘alleged legality’, ‘price’, ‘low availability of other (traditional) drugs’, ‘non-detectability’ and ‘poor quality of other (traditional) drugs’ whilst ‘expecting different or new experiences (regarding drug effects)’ did not become part of this factor. The identified factor was able to explain more than 44% of the variances. The component loadings ranged between 0.60 and 0.73. This factor was given the contextual motivation factor label. The Cronbach Alpha for these items was 0.682, which is on the threshold of acceptability. As the main focus of the present analysis is to identify the discriminative features of marginalised users compared to other user groups, we analysed the variances of the external/ contextual motivation factor scores in marginalised and non-marginalised users across the different countries. To be able to do so, we applied a re-coding during which we kept the marginalised group as it was (N = 265), and the night life and online community groups were merged into one variable labelled as non-marginalised (N = 1568). The Netherlands was left out from this analysis as only one marginalised respondent was interviewed in this country. As the figure below (Fig. Users in Online Community p < 0.05; *p < 0.001. Different subscript letters indicate a significant (p ≤ 0.05) differences in proportions Table 2 Last month frequencies of substance use (%) according to user groups Synthetic cannabinoids included synthetic cannabinoids pure, herbal brands; NPS stimulants included branded stimulants (e.g. “bath salts”), empathogens, nootropics, (obtained pure, e.g.. mephedrone, MDPV, a-PVP); NPS Psychedelics included e.g. NBOMe-x. 2C-x′; NPS dissociatives included, e.g. methoxetamine; controlled substances included cannabis, amphetamine, xtc, cocaine, crack-cocaine, heroin, magic mushroom, LSD, metamphetamine, ketamin, GHB. p < 0.05; **p < 0.001. Different subscript letters indicate a significant (p ≤ 0.05) differences in proportions Synthetic cannabinoids included synthetic cannabinoids pure, herbal brands; NPS stimulants included branded stimulants (e.g. “bath salts”), empathogens, nootropics, (obtained pure, e.g.. mephedrone, MDPV, a-PVP); NPS Psychedelics included e.g. NBOMe-x. 2C-x′; NPS dissociatives included, e.g. methoxetamine; controlled substances included cannabis, amphetamine, xtc, cocaine, crack-cocaine, heroin, magic mushroom, LSD, metamphetamine, ketamin, GHB. p < 0.05; **p < 0.001. Different subscript letters indicate a significant (p ≤ 0.05) differences in proportions International Journal of Mental Health and Addiction Perceived Context of NPS Use among Different User Groups 1) shows, contextual/external motivation factor scores are higher among marginalised users in all countries included into the analysis, whilst in two countries Table 3 The variance of contextual motives among user groups Motives User groups Statistics Marginalised N = 266 Night life N = 647 Internet N = 2110 F Alleged legality, mean (SD) 1.79 (1.26)a 1.36 (0.84)b 1.57 (1.16)c 14.93* Poor quality of other traditional drugs, Mean (SD) 2.45 (1.58)a 1.48 (0.94)b 1.78 (1.26)c 53.34* Price, mean (SD) 2.46 (1.64)a 1.53 (1.08)b 1.81 (1.24)c 47.60* Expecting new experiences, mean (SD) 2.55 (1.51)a 3.08 (1.54)b 2.79 (1.53)a 13.2* Non-detectability, mean (SD) 1.94 (1.42)a 1.31 (0.94)b 1.50 (1.12)c 28.70* Lack of traditional drugs, mean (SD) 2.35 (1.52)a 1.44 (0.92)b 1.77 (1.31)c 47.05* Contextual motive scale, mean (SD) 2.23 (0.96)a 1.49 (0.70)b 1.91 (0.91)c 75.720*** ANOVA p < 0.001 The Games-Howell test was used as post hoc test. Different subscript letters indicate a significant difference (p ≤ 0.05) in the parameters Contextual motive scale results are presented without the Netherlands, the N for the user groups is marginalised, 265; night life, 458; Internet 1110 Table 3 The variance of contextual motives among user groups Table 3 The variance of contextual motives among user groups International Journal of Mental Health and Addiction 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 Non marginalised Marginalised Germany Hungary Ireland Portugal Poland Means Fig. 1 The variance of contextual motivation according to user groups and countries (N = 1801). Note: F value for countries, 5.811; for user groups, 16.343*; for countryuser group interaction, 1.612, * p < 0.001 Fig. 1 The variance of contextual motivation according to user groups and countries (N = 1801). Note: F value for countries, 5.811; for user groups, 16.343*; for countryuser group interaction, 1.612, *** p < 0.001 (Hungary and Poland), the differences in the means between marginalised and non- marginalised user groups are much smaller than in the others. These comparisons show that contextual motivations presumed to be NPS specific play a different role depending on the country of residence of the respondents. Discriminating Features of Marginalised NPS Use—Binary Logistic Regression In the next phase, we studied the factors that might describe the discriminating features of marginalised NPS users. For this purpose, two binary logistic regression models were created; in each case, the dependent variable was if the respondent was categorised as marginalised vs. non-marginalised user. The Netherlands was left out from this analysis as the total number of marginalised users reached during the interview was only one person. The total number of respondents included into this analysis was 1833. In the first model, we included only socio- demographic characteristics (education, current status of employment, gender, age, size of the settlement and living conditions). The low level of accomplished education, living on benefits, being unemployed, being older, living with friends, using homeless-type accommodation arrangements and living in bigger settlements significantly increased the probability of being a marginalised user (see Table 4). Country of residence seems to be associated with significant and high odds ratios for being categorised as a marginalised user, though these findings might just be the consequence of sample attainment. As the number of marginalised users reached during the data collection was very much varied we carried out a country based binary logistic regression analysis as well during which the same set of variables were used to find out the country specific characteristics of the marginalised user group. Level of education played a role in being categorised as a marginalised user in the case of Hungary and Germany. In Hungary, less education elevated the risk of being part of the marginalised user group whilst in Germany any accomplished education increased this risk (very probably due to the high number of participants with higher education attainment in the sample). The employment status proved to be an active ingredient of being categorised as a marginalised user in Poland. Living on benefits or being unemployed elevated this risk 32–9.2 times higher than being a student. Age also had a different role depending on the country of International Journal of Mental Health and Addiction Table 4 Binary logistic regression analysis of socio-demographic factors for being categorised as a marginalised user Variables OR (95% C.I.) p Country of residence Hungary 46.71 (16.80–129.84) < 0.001 Ireland 72.00 (16.78–308.92) < 0.001 Portugal 4.56 (1.38–14.99) 0.012 Poland 19.09 (7.56–48.21) < 0.001 Germany Ref. Discriminating Features of Marginalised NPS Use—Binary Logistic Regression Level of education None 26.26 (1.83–377.20) 0.016 Primary school 9.72 (1.17–80.88) 0.035 Secondary school 1.91 (0.24–15.32) 0.541 College 0.49 (0.06–4.09) 0.508 MA or higher Ref. Current status of employment Full-time worker 0.84 (0.25–2.75) 0.768 Part-time worker 1.70 (0.59–4.90) 0.326 Self-employed 1.16 (0.32–4.17) 0.825 Benefits 12.47 (3.88–40.08) < 0.001 Retired 2.61 (0.19–35.28) 0.471 Home duties 8.14 (0.51–129.14) 0.137 Unemployed 4.05 (1.53–10.73) 0.004 Other 0.69 (0.02–25.91) 0.841 Student Ref. Age categories 18–24 Ref. 25–34 3.32 (1.64–6.72) < 0.001 35–44 7.96 (3.19–19.88) < 0.001 45– 16.56 (4.41–62.14) < 0.001 Living conditions Own home Ref. With parents/family 1.80 (0.72–4.62) 0.201 With friends/in friends’ home 13.15 (3.80–45.53) < 0.001 Residential care 64.34 (16.01–258.58) < 0.001 Rent apartment or room 1.48 (0.59–3.70) 0.403 Homeless accommodation/hostel 81.60 (18.94–351.66) < 0.001 Other 31.36 (6.29–156.43) < 0.001 Size of the settlement Below 50,000 Ref. Between 50 and 100,000 4.30 (1.46–12.64) 0.008 More than 100,000 8.43 (3.49–20.35) < 0.001 Ref., Reference category Binary logistic regression analysis (FSTEP), 8 iterations All variables remained in the equation, though just a few of them contributed significantly to the categorisation All variables remained in the equation, though just a few of them contributed significantly to the categorisat residence of the respondent. In case of Hungary and Poland, being older than 24 highly elevates the risk of being part of the marginalised group. Living with friends just in Germany, residential care in Germany and Poland and homeless accommodation in Germany, Hungary and Poland contributed to the marginalised categorisation. Size of the settlement where the respondent was living did not add to this categorisation. In the case of Portugal and Ireland, none of the variables revealed any patterns (see detailed result in Appendix Table 6). Another binary logistic regression model was developed on substance use–related behav- iour (Table 5). We included NPS and controlled substance use frequencies (life time, last year, last month), venue of substance use, problems (physical, social and mental health) experienced after using NPS, parallel usage of different substances, and contextual motivation factor. The model shows that characteristics of problem drug use (EMCDDA, 2009) are clearly associated with being a marginalised respondent. Those who used heroin or methamphetamine during International Journal of Mental Health and Addiction Table 5 Binary logistic regression analysis of substance use related factors for being categorised as a marginalised user Name of the variable OR (95% C.I. ) p. Ref., reference category Binary logistic regression analysis (FSTEP), 7 iterations Synthetic cannabinoids included synthetic cannabinoids pure. Herbal brands; NPS stimulants included branded stimulants (e.g.: “bath salts”), empathogens, nootropics (obtained pure, e.g.: mephedrone. MDPV, a-PVP), NPS psychedelics included, e.g. NBOMe-x. 2C-x′, NPS dissociatives included, e.g. methoxetamine; controlled substances included cannabis, amphetamine, xtc, cocaine, crack-cocaine, heroin, magic mushroom, LSD, metamphetamine, ketamin, GHB Variables not remained in the equation during the FSTEP procedure: Substance use related variables: synthetic cannabinoids last year consumption; stimulants/empathogens/ nootropics pure last year and last month consumption; dissociative last year and last month consumption, psychedelics last month consumption, cannabis las year consumption, cocaine last year, last month consumption, heroine last year, last month consumption, crack last year consumption, LSD last month consumption, metamphetamine last month consumption, NPS use together with another NPS Venue of the substance use related items: work place, special places created for drug users Contextual motive scale their life time or in the last year have a 2.1 to 3.1 times bigger chance to be categorised as marginalised than those who did not. In case of those who had ever or during the last year injected any kind of substances, the odds ratios are considerably elevated. Frequencies of current substance use reveal that using synthetic cannabinoids or crack also highly increases the odds ratios for being categorised as marginalised user, whilst using cannabis or amphet- amine during the last month decreased the chance for belonging to this category. If someone is using substances in residential care, the chance for being categorised as a marginalised user is 5.4 times higher, whilst using NPS and alcohol or controlled substances together decreases the likelihood of being categorised as a marginalised user. their life time or in the last year have a 2.1 to 3.1 times bigger chance to be categorised as marginalised than those who did not. In case of those who had ever or during the last year injected any kind of substances, the odds ratios are considerably elevated. Frequencies of current substance use reveal that using synthetic cannabinoids or crack also highly increases the odds ratios for being categorised as marginalised user, whilst using cannabis or amphet- amine during the last month decreased the chance for belonging to this category. Discriminating Features of Marginalised NPS Use—Binary Logistic Regression Last year/life time substance use Last year NPS psychedelic use 0.25 (0.13–0.47) < 0.001 Last year ecstasy use 0.60 (0.34–1.03) 0.066 Life time heroin use 2.06 (1.10—3.85) 0.023 Last year methamphetamine use 3.09 (1.70–5.59) < 0.001 Last year LSD use 0.41 (0.19–0.85) 0.017 Life time intravenous drug use 8.32 (4.28–16.20) < 0.001 Last month substance use Last month synthetic cannabinoids use (incl. all herbal blends) 2.28 (1.38–3.77) 0.001 Last month cannabis use 0.37 (0.23–0.59) < 0.001 Last month amphetamine use 0.40 (0.21–0.77) 0.006 Last month crack use 6.93 (2.18–22.05) 0.001 Last month intravenous drug use 5.22 (2.52–10.81) < 0.001 Problems and risk behaviour Perceived social problems 0.48 (0.30–0.76) 0.002 NPS consumption together with controlled substances 0.54 (0.34–0.85) 0.008 NPS consumption together with alcohol 0.49 (0.31–0.77) 0.002 Venue of consumption At own home 0.24 (0.15–0.39) < 0.001 At a friend’s house 0.57 (0.35–0.91) 0.018 Night life 0.57 (0. 34–0.95) 0.029 In residential care 5.42 (1.62–18.09) 0.006 Discussion The study has provided unique insight into NPS and controlled drug using patterns of different user groups and identified a series of demographic and external contextual factors potentially contributing to being categorised as marginalised user in six European countries. Based on frequencies, and mostly on short-term frequencies of substance use (last month) we can say that among recent NPS users, much higher frequencies of substance use can be observed than in adult population surveys, suggesting that we studied a very specific population. The special nature of the population we studied was illustrated in the results chapter, based on comparing our own data and the ones presented in the 2018 European Drug Report (EMCDDA 2018). Based on this comparison, it can be assumed that those three user groups which were investigated in the context of a large European transnational project are much more intense, and in some cases, much more problematic users than those representing the adult population in the relevant countries. The impression of intense substance use in these groups might also be the consequence of poly drug using tendencies of NPS users; NPS use very frequently happens in the context of polysubstance use (European Monitoring Centre for Drugs and Drug Addiction 2017b). The problematic nature of substance use is supported by the frequent intravenous substance use especially among marginalised users. Another warning sign is that the respondents were not necessarily able to say what kind of substance they were using, and they were not able to reflect on the active ingredients of the substances (see further details in Van Hout et al. 2018; Korf et al. 2019). Epidemiological research so far has not yet revealed the discriminating features of different NPS user groups as large-scale representative studies cannot reflect on the specificities of special groups. There are plenty of studies which describe the socioeconomic properties of substance users (including legal and illegal substances); these studies consistently claim that disadvantageous socioeconomic situation significantly contributes to elevated risk of heavy/ problematic substance use (see e.g. Janicijevic et al. 2017). In the light of the few adult population surveys implemented in the last few years, it can be assumed that NPS play a substantial role in the poly-consumption patterns of drug users (i.e. Paksi et al. 2016, 2017). According to our analysis and in accordance with other studies (i.e. Van Hout et al. 2018), the substance-related preferences of different user groups vary. Ref., reference category If someone is using substances in residential care, the chance for being categorised as a marginalised user is 5.4 times higher, whilst using NPS and alcohol or controlled substances together decreases the likelihood of being categorised as a marginalised user. As we found considerable differences in the sociodemographic variables according to countries of residence, we analysed the country-specific contributing factors of substance International Journal of Mental Health and Addiction use (the same set of variables as described above) in being categorised as a marginalised user. (see Appendix, Table 7). The analysis revealed that last year and last month use of heroin and intravenous drug use (in Germany and Poland), amphetamine, methamphetamine use (in Poland) and synthetic cannabinoid and crack use (in Hungary), highly elevated the risk of being categorised as a marginalised user. The parallel use of different new substances (NPS) contributed to being categorised as a marginalised user in Germany (OR = 6.53), whilst the venue of substance use found to be influential in the case of Poland where usage in residential care substantially increased the odds ratios of this type of categorisation. Contextual motives did not contribute to being categorised as a marginalised user among German respondents. In the case of Portugal and Ireland none of the variables revealed any patterns, probably due to the low sample size reached in these countries. Discussion In all user groups (marginalised users, night life and online), the most frequently used substances are the controlled ones, with NPS being just part of poly-substance using repertoires. Based on our data and in accordance with other research (Potter and Chattwin 2018), it can be said that NPS use does not constitute an independent category of drug use; NPS users are still using controlled substances, though International Journal of Mental Health and Addiction marginalised users are using significantly less of them than the other user groups. In the case of marginalised users, regardless of their country of residence, the rates for current use show that NPS play a relatively bigger role in their drug using preferences than traditional controlled substances, which is the reverse in the case of other user groups (night life and internet community). The most popular NPS type substances among marginalised users are synthetic cannabinoids, and NPS stimulants. In some countries, it is worth noting, as qualitative research also revealed (Kaló et al. 2017), that some individuals in disadvantageous social-economic situations start drug-using careers or pathways with NPS. The early onset of psychoactive substance use and especially the preference towards NPS, contributes substantially to the development of particularly destructive forms of drug use in socially marginalised groups. Our analysis also revealed that the newly added items to the motivation scale are organised in one factor named contextual factor. The analysis also revealed that the perceived (reported) importance of these motives do not play a decisive role in drug using preferences of recent NPS users. We can conclude that, if at all, particularly low availability and poor quality of controlled substances as well as the relatively low price of NPS may play a more important role in the motivation towards these substances in case of marginalised users compared with other recent NPS user groups. The respondents from all user groups rated the contextual motivation items relatively low, the least important motives are the alleged legality and non- detectability of these substances. The relative unimportance of these items might be the consequence of the legal status of different new psychoactive substances in the different countries. Another possible explanation of this finding might also be that all recent NPS users are (intensively) using controlled substances; consequently, the temporary legal status of any given NPS is not a significant issue to articulate their choices for consumption. Discussion This finding might be highly relevant in the drug policy/legislation debates in connection to the current regulatory efforts. Some of the external/contextual motivation items received higher ratings from marginalised respondents with significant differences depending on country of residence. Data describing the drug use patterns of the three diverse NPS user groups and the regression model, including a wide range of socio-demographic variables (age, level of education, labour market position, recent use of NPS and controlled substances, route of administration), strongly suggest that marginalised users are the most at-risk population with lower education and with the most risky drug using patterns. As our selection criteria regarding marginalised users was very much concentrating on the venue of substance use (streets, public places) and/or being clients of low threshold/harm reduction services and their networks, it is very much reassuring that our analysis revealed that the marginalised users reached have indeed a well identifiable socio-demographic profile, describing their elevated vulnerability. These findings not just mirror the sample selection criteria, but it draws a more refined picture regarding the socio-demographic factors and the discriminating substance use habits of marginalised users; outlining the profile of high-risk problematic drug use. The binary logistic regression model, which included all respondents from the participating countries (apart from the Netherlands) is conclusive regarding the main sociodemographic characteristics and substance use patterns of the marginalised user group. More detailed country-based analysis suggests that variables supposed to be contributing to the marginalised—non-marginalised categorisation of the respondents in different countries are very much varied, at some points, the findings are even contradictory. Therefore, it is suggested that further epidemiological studies should be carried out to obtain reliable data from different countries on the relative weight of NPS use in different user groups and on other factors contributing to enhance our user group specific understanding of NPS use and, in International Journal of Mental Health and Addiction general, problematic drug use behaviour. Further research is especially needed as our data cannot be considered as representative for the participating countries or for the chosen user groups. Contextual motives, as the logistic regression model shows, do not contribute to be categorised as a marginalised user when other variables were controlled. Discussion The intense drug use of marginalised NPS users should call the attention of policy makers to strengthen tailor-made harm reduction services and targeted prevention activities in these groups, to properly respond to the public health and legal concerns in some European countries where NPS use is especially prevalent among low social status marginalised users. Conclusions This study contributes to the better understanding of NPS use in special populations, especially in the marginalised user group, by describing the discriminating features of substance use, its setting and the sociodemographic properties of respondents in this group. This statement might remain to be true even though the data collection was carried out on a convenience sample (with different but standardised recruitment strategies in each user group) and the number of respondents in the different user groups was very much uneven, and we have no sufficient reliable information on the prevalence of NPS use in the general population. Our data can inform decision-makers about the necessity of enhancing complex interventions targeting high risk marginalised users and also the scientific community that large-scale epidemiological surveys in special populations should be carried out to further develop our understanding regarding NPS use in certain populations. Funding Information This study was supported by the European Union (New Psychoactive Substances: transnational project on different user groups, user characteristics, extent and patterns of use, market dynamics, and best practices in prevention [HOME/2014/JDRU/AG/DRUG/7077]), the Hungarian National Research, Development and Innovation Office (Grant number: KKP126835), and the Hungarian Ministry of Human Capacities (ELTE Institutional Excellence Program, 783-3/2018/FEKUTSRAT). The study was also supported for the realisation of this international co-financed science project in 2016-2017 by the Polish Ministry of Science and Higher Education. The funding institutions had no role in the study design or the collection, analysis and interpretation of the data, writing the manuscript, or the decision to submit the paper for publication. Compliance with Ethical Standards Conflict of Interest The authors (Katalin Felvinczi, Annemieke Benschop, Róbert Urbán, Marie Claire Van Hout, Katarzyna Dąbrowska, Evelyn Hearne, Susana Henriques, Zsuzsa Kaló, Gerrit Kamphausen, Joana Paula Silva, Łukasz Wieczorek, Bernd Werse, Michal Bujalski, Zsolt Demetrovics, Dirk Korf) declare that they do not have any conflicts of interest that could constitute a real, potential or apparent issue with respect to their involvement in the publication. The authors also declare that they do not have any financial or other relations (e.g. directorship, consultancy or speaker fee) with companies, trade associations, unions or groups (including civic associations and public interest groups) that may affect the results or conclusions in the study. Sources of funding are acknowledged. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, dis- tribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. International Journal of Mental Health and Addiction Table 6 Binary logistic regression analysis of socio-demographic factors for being categorised as a marginalised user according to country of residence Country name of the variable GER HU IR PL PT OR [95% CI] Level of education None 0.087 [0.003–2.814] x x x x Primary school 0.046* [0.003–0.675] 193.911* [5.079–7402.937] x x x Secondary school 0.015* [0.001–0.266] 34.166* [1.051–1110.678] x x x College 0.028* [0.001–0.527] 1.430 [0.05–40.92] x x x MA/MSc or higher Ref. Ref. Ref. Ref. Ref. Current status of employment Full-time worker x x x 0.531 [0.055–5.148] x Part-time worker x x x 1.984 [0.246–15.396] x Self-employed x x x 1.402 [0.097–20.204] x Benefits x x x 32.098* [3.784–272.12] x Retired x x Home duties x x Unemployed x x x 9.221* [1.813–46.902] x Other x x 1 x x Student Ref. Ref. Ref. Ref. Ref. Age 18–24 Ref. Ref. Ref. Ref. Ref. 25–34 x 5.691* [1.021–31.703] x 2.223 [0.608–8.129] x 35–44 x 54.236* [5.316–553.321] x 34.739** [4.982–242.241]x x 45+ x 24.787 [0.892–689.086] x 33.168* [2.16–509.312] x Gender Female Ref. Ref. Ref. Ref. Ref. Compliance with Ethical Standards Male 9.479* [1.872–48.009] Living conditions With parents/family 0.584 [0.031–11.13] 0.679 [0.096–4.82] x 1.592 [0.281–9.037] x With friends/in friends’ home 19.933* [1.247–318.649] 5.758 [0.342–96.836] x 19.891* [1.515–261.155] x Residential care 106.556* [4.904–2315.114] 48.227 [0.607–3832.513] x 142.338 [13.485–1502.466] x Rent apartment or room 2.690 [0.278–31.567] 0.487 [0.054–4.383] x 1.287 [0.222–7.48] x tional Journal of Mental Health and Addiction International Journal of Mental Health and Addiction Table 6 (continued) Country name of the variable GER HU IR PL PT Homeless accommodation/hostel 169.212 [11.502–2489.414] 93.947* [4.723–1868.553] x x x other 47.130* [1.734–1281.187] 17.095* [1.046 × 279.35] x 99.617* [4.503–2203.68] x Own home Ref. Ref. Ref. Ref. Ref. Size of the settlement Below 50.000 Ref. Ref. Ref. Ref. Ref. Between 50 × 100.000 x x x 7.273* [1.43–36.988] x More than 100.000 x x x 2.848 [0.616–13.174] x p < 0.05; p < 0.001; x the variable is not in the equation or the OR not within the confidence interval International Journal of Mental Health and Addicti Table 6 (continued) Country name of the variable GER HU IR PL PT Homeless accommodation/hostel 169.212* [11.502–2489.414] 93.947* [4.723–1868.553] x x x other 47.130* [1.734–1281.187] 17.095* [1.046 × 279.35] x 99.617* [4.503–2203.68] x Own home Ref. Ref. Ref. Ref. Ref. Size of the settlement Below 50.000 Ref. Ref. Ref. Ref. Ref. Between 50 × 100.000 x x x 7.273* [1.43–36.988] x More than 100.000 x x x 2.848 [0.616–13.174] x p < 0.05; p < 0.001; x the variable is not in the equation or the OR not within the confidence interval Table 6 (continued) Country name of the variable GER HU IR PL PT Homeless accommodation/hostel 169.212* [11.502–2489.414] 93.947* [4.723–1868.553] x x x other 47.130* [1.734–1281.187] 17.095* [1.046 × 279.35] x 99.617* [4.503–2203.68] x Own home Ref. Ref. Ref. Ref. Ref. Size of the settlement Below 50.000 Ref. Ref. Ref. Ref. Ref. Compliance with Ethical Standards Between 50 × 100.000 x x x 7.273* [1.43–36.988] x More than 100.000 x x x 2.848 [0.616–13.174] x p < 0.05; p < 0.001; x the variable is not in the equation or the OR not within the confidence interval International Journal of Mental Health and Addiction Table 7 Binary logistic regression analysis of substance use related factors for being categorised as a marginalised user according to country of residence Country name of the variable GER HU IR PL PT OR [95% CI] Frequency of life time/last year substance use Cannabis x x x 0.179 [0.048–0.661] x NPS psychedelics x 0.055* [0.008–0.4] x x x Ecstasy x 0.247* [0.079–0.766] x 0.245* [0.078–0.772] x Amphetamine x x x 14.902** [4.504–49.301] x Cocaine x x 0.079* [0.016–0.403] x Heroin 45.241* [4.874–419.894] x x x x Methamphetamine x x x 4.721* [1.641–13.587] x Intravenous drug use x 87.058 [24.871–304.73] x 28.784 [8.943–92.64] x Synthetic cannabinoids (incl. All herbal blends) x x x 6.533* [2.268–18.819] x Frequency of last month substance use Synthetic cannabinoids (incl. All herbal blends) x 4.938* [1.661–14.684] x x x NPS psychedelics x x x 0.017* [0.001–0.278] x cannabis x x x 0.176* [0.062–0.501] x ecstasy xx x x x amphetamine 0.218* [0.054–0.886] x x 0.09* [0.019–0.433] x cocaine x x x x x crack x 66.594* [2.125–2086.848] x x x NPS dissociatives x x x 0.106 [0.007–1.56] x intravenous drug use 18.733 [5.008–70.073] x x 80.204 [12.109–531.231] x PROBLEMS AND RISK BEHAVIOUR mental and physical side effects related to drug use x x x 0.393 [0.164–0.941] 0.064* NPS use together with alcohol x 0.2* [0.068–0.588] x x x NPS use together with other NPS 6.532* [1.38–30.923] x x x x NPS use together with controlled substances x 0.144* [0.045–0.458] x x x Venue of consumption At own home 0.168* [0.04–0.703] 0.202* [0.065–0.627] x 0.12 [0.045 × 0.316] x In residential care x x x 209.463 [19.483–2251.944] x p < 0.05; *p < 0.001; x the variable is not in the equation or the OR not within the confidence interval International Journal of Mental Health and Addiction References Barratt, M. J., Cakic, V., & Lenton, S. (2013). Patterns of synthetic cannabinoid use in Australia. Drug and Alcohol Review, 32, 141–146. https://doi.org/10.1111/j.1465-3362.2012.00519.x. Benschop, A., Liebregts, N., van der Pol, P., Schaap, R., Buisman, R., van Laar, M., van den Brink, W., de Graaf, R., & Korf, D. J. (2015). Reliability and validity of the marijuana motives measure among young adult frequent cannabis users and associations with cannabis dependence. Addictive Behaviors, 40, 91–95. Benschop A., Bujalski M, Dabrowska K., Demetrovics Z., Egger D., Felvinczi K., Henriques S., Kalo Z., Kamphausen G., Korf D.J., Nabben T., Silva J.P., Van Hout M.C., Werse B., Wells J., Wieczorek L., Wouters M. (2017). New psychoactive substances: transnational project on different user groups, user characteristics, extent and patterns of use, market dynamics, and best practices in prevention (NPS-transnational project; HOME/2014/JDRU/AG/DRUG/7077). Final Report. Amsterdam: Bonger Institute of Criminology, University of Amsterdam. Benschop A, Urban R, Fövény M, Dąbrowska K, Van Hout MC, Dąbrowska K, Felvinczi K, Hearne E, Henriques S, Kaló Zs, Kamphausen G, Silva JP, Wieczorek L, Werse B, Bujalski M, Korf D, Demetrovics Zs, (2019) Why do people use new psychoactive substances? Development of a new measurement tool in six European countries. Submitted for publication. Caudevilla F (2016) The emergence of deep web marketplaces: a health perspective. The internet and drug markets. (European Monitoring Centre for Drugs and Drug Addiction: Insights 21) (pp. 69–75). Luxembourg: Publications Office of the European Union. Cooper, M. (1994). Motivations for alcohol use among adolescents: development and validation of a four-factor model. Psychological Assessment, 6(2), 117–128. Corazza, O., Simonato, P., Corkery, J., Trincas, G., & Schifano, F. (2014). “Legal highs”: safe and legal “heavens”? A study on the diffusion, knowledge and risk awareness of novel psychoactive drugs among students in the UK. Rivista di Psichiatria, 49(2), 89–94. Cox, W. M., & Klinger, E. (1988). A motivational model of alcohol use. Journal of Abnormal Psychology, 97(2), 168–180. Eurobarometer. (2011). Youth attitudes on drugs. Analytical report. Flash EB Series #330. Available at: http:// ec.europa.eu/public_opinion/archives/flash_arch_en.htm. Eurobarometer. (2014). Flash Eurobarometer 401. Young people and drugs. Available at: http://ec.europa. eu/public_opinion/flash/fl_401_en.pdf European Monitoring Centre for Drugs and Drug Addiction. (2014). European drug report 2014: trends a developments. Luxembourg: Publications Office of the European Union. European Monitoring Centre for Drugs and Drug Addiction. (2017a). European drug report 2017: trends a developments. Luxembourg: Publications Office of the European Union. European Monitoring Centre for Drugs and Drug Addiction. References (2017b). High-risk drug use and new psychoactive substances. Luxembourg: EMCDDA Rapid Communication, Publications Office of the European Union. European Monitoring Centre for Drugs and Drug Addiction. (2018). European drug report 2018: trends and developments. Luxembourg: Publications Office of the European Union. Janicijevic, K. M., Kocic, S. S., Radevic, S. R., Jovanovic, M. R., & Radovanovic, S. M. (June 2017). Socioeconomic factors associated with psychoactive substance abuse by adolescents in Serbia. Frontiers in Phamacology, 8. Kaló, Z., Paksi, B., Felvinczi, K., Demetrovics, Z. (2017). New psychoactive substances in Hungary: user groups and characteristics, patterns, motives, and problems in Hungary. In: 20th conference of the European Association of Substance Abuse Research. Venue and date of the conference: Nunspeet, Hollandia, 2017.05.18 21.p. 40. Korf, D., Benschop, A., Werse, B., Kamphausen, G., Felvinczi, K., Dabrowska, K., Hernriques, S., Nabben, T., Wieczorek, L., Bujalski, M., Zs, K., Hearne, E., & Van Hout, M. C. (2019). How and where to find NPS users: a comparison of sampling methods in a cross-national survey among current users of new psycho- active substances. International Journal of Mental Health and Addiction. https://doi.org/10.1007/s11469- 019-0052. MacLeod, K., Pickering, Gannon, M., Greenwood, S., Liddell, D., Smith, A., Johnstone, L., Burton, G.: (2016) Understanding the patterns of use, motives, and harms of new psychoactive substances in Scotland. Final Report to the Scottish Government, November 2016. p Measham, F., Moore, K., Newcombe, R., & Welch, Z. (2010). Tweaking, bombing, dabbing and stockpiling: the emergence of mephedrone and the perversity of prohibition. Drugs and Alcohol Today, 10, 14–21. https://doi.org/10.5042/daat.2010.0123. p g Moore, K., Dargan, P. I., Wood, D. M., & Measham, F. (2013). Do novel psychoactive substances displace established club drugs, supplement them or act as drugs of initiation? The relationship between mephedrone, ecstasy and cocaine. European Addiction Research, 19(5), 276–282. International Journal of Mental Health and Addiction Neicun, J., Steenhuizen, M., van Kessel, R., Yang, J. C., Negri, A., Czabanowska, K., Corazza, O., & Roman- Urrestarazu, A. (2019). Mapping novel psychoactive substances policy in the EU: the case of Portugal, the Netherlands, Czech Republic, Poland, the United Kingdom and Sweden. PLoS One, 14(6), e0218011. https://doi.org/10.1371/journal.pone.0218011. Paksi, B., Demetrovics, Zs, Kaló, Zs, Felvinczi, K. (2017) Nagy kitettségű, speciális UPSZ használó csoportok szerhasználati szokásai In: [s.n.] (szerk.) Személyes tér - közös világ: A Magyar Pszichológiai Társaság XXVI. Országos Tudományos Nagygyűlése : Kivonatkötet. 347 p. Konferencia helye, ideje: Szeged, Magyarország, 2017.06.01-2017.06.03. Budapest: Magyar Pszichológiai Társaság, 2017. p. 195. References 1 p. (ISBN:978–615–80241-4-3). Paksi, B. (2017). ÚPSZ-használattal kapcsolatos epidemiológiai tapasztalatok az Országos Lakossági Adatfelvétel Addiktológiai Problémákról (OLAAP 2015) című kutatás alapján. In: Felvinczi Katalin (szerk.) Változó képletek.: ÚJ (abb) szerek: kihívások, mintázatak. 158 p. Budapest: L’Harmattan Kiadó, 2017b. pp. 45–68. (ISBN:978-963-414-120-4). ( ) Paksi, B., Magi, A., Felvinczi, K., Demetrovics, Zs. (2016). The prevalence of new psychoactive substances in Hungary – based on a general population survey dealing with addiction related problems (OLAAP 2015). IV. International Conference on Novel Psychoactive Substances (NPS). Budapest, 30–31 May 2016. Potter, G. R., & Chattwin, C. (2018). Not particularly special: critiquing ‘NPS’ as a category of drugs. Drugs: Education, Prevention and Policy, 25(4), 2018–2336. https://doi.org/10.1080/09687637.2017.1411885. Schifano, F., Chiappini, S., Corkery, J. M., & Guirguis, A. (2019). Assessing the 2004–2018 fentanyl misusing issues reported to an international range of adverse reporting systems. Frontiers in Pharmacology, 10, 46. https://doi.org/10.3389/fphar.2019.00046. Simons, J. S., Correia, C. J., Carey, K. B., & Borsari, B. E. (1998). Validating a five-factor marijuana motives measure: relations with use, problems, and alcohol motives. Journal of Counseling Psychology, 45(3), 265– 273. Smith, K., & Flatley, J. (2011). Drug misuse declared: Findings from the 2010/2011 British Crime Survey England and Wales. London: Home Office. Soussan, Ch, Andersson, M., Kjellgren, A. (2018). The diverse reasons for using novel psychoactive substances - a qualitative study of the users’ own perspectives. International Journal of Drug Policy, Volume 52, February 2018, p. 71–78. p S., & Kjellgren, A. (2016). The users of novel psychoactive substances: online survey about their characteristics, attitudes and motivations. International Journal of Drug Policy, 32(2016), 77–84. Sutherland, R., Peacock, A., Whittaker, E., Roxburgh, A., Lenton, S., Matthews, A., Butler, K., Nelson, M., Burns, L., & Bruno, R. (2016b). New psychoactive substance use among regular psychostimulant users in Australia, 2010–2015. Drug and Alcohol Dependence, 161(2016), 110–118. nki, D., & Vicente, J. (2013). PDU (problem drug use) revision summary. European Monitoring Centre for Drugs and Drug Addiction. Luxembourg: Publications Office of the European Union. Sutherland, R., Peacock, A., Whittaker, E., Roxburgh, A., Lenton, S., Matthews, A., & Bruno, R. (2016a). New psychoactive substance use among regular psychosti-mulant users in Australia, 2010–2015. Drug and Alcohol Dependence, 161, 110–118. https://doi.org/10.1016/j.drugalcdep.2016.01.024. p p g j g The ESPAD Group. (2016). 2015 ESPAD Report. Lisbon: EMCDDA. The ESPAD Group. (2016). 2015 ESPAD Report. Lisbon: EMCDDA. Thurn, D., Kuntsche, E., Weber, J. A., & Wolstein, J. (2017). References Development and validation of the amphetamine- type stimulants motive questionnaire in a clinical population. Frontiers in Psychiatry, 8(183). https://doi. org/10.3389/fpsyt.2017.00183. Thurn, D., Kuntsche, E., Weber, J. A., & Wolstein, J. (2017). Development and validation of the amphetamine- type stimulants motive questionnaire in a clinical population. Frontiers in Psychiatry, 8(183). https://doi. org/10.3389/fpsyt.2017.00183. Van Hout, M. C., Benschop, A., Bujalski, M., Zs, D., Dąbrowska, K., Felvinczi, K., Hearne, E., Henriques, S., Zs, K., Kamphausen, G., Korf, D., Nabbe, T., Silva, J., Wieczorek, L., & Werse, B. (2018). Health and social problems associated with recent novel psychoactive substance (NPS) use amongst marginalised, nightlife and online users in six European countries. International Journal of Mental Health and Addiction., 16(2), 480–495. Werse, B., & Morgenstern, C. (2012). How to handle legal highs? Findings from a German online survey and considerations on drug policy issues. Drug and Alcohol Today, 12, 222–231. https://doi.org/10.1108 /17459261211286636. Winstock, A. R., Lawn, W., Deluca, P., & Borschmann, R. (2015). Methoxetamine: an early report on the motivations for use, effect profile and prevalence of use in a UK clubbing sample. Drug and Alcohol Review., 35, 212–217. https://doi.org/10.1111/dar.12259. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. International Journal of Mental Health and Addiction Affiliations Katalin Felvinczi1 & Annemieke Benschop2 & Róbert Urbán1 & Marie Claire Van Hout3 & Katarzyna Dąbrowska4 & Evelyn Hearne5 & Susana Henriques6 & Zsuzsa Kaló1 & Gerrit Kamphausen7 & Joana Paula Silva6 & Łukasz Wieczorek4 & Bernd Werse7 & Michal Bujalski4 & Zsolt Demetrovics1 & Dirk Korf2 Katalin Felvinczi1 & Annemieke Benschop2 & Róbert Urbán1 & Marie Claire Van Hout3 & Katarzyna Dąbrowska4 & Evelyn Hearne5 & Susana Henriques6 & Zsuzsa Kaló1 & Gerrit Kamphausen7 & Joana Paula Silva6 & Łukasz Wieczorek4 & Bernd Werse7 & Michal Bujalski4 & Zsolt Demetrovics1 & Dirk Korf2 1 Institute of Psychology, ELTE Eötvös Loránd University, Izabella utca 46, Budapest H-1064, Hungary 2 Bonger Institute, University of Amsterdam, Amsterdam, Netherlands 3 Public Health Institute, Liverpool John Moores University, Liverpool, UK 4 Institute of Psychiatry and Neurology, Warsaw, Poland 5 School of Health Sciences, Waterford Institute of Technology, Waterford, Ireland 6 Centre for Research and Studies in Sociology, University Institute of Lisbon, Lisbon, Portugal 7 Goethe-Universität, Frankfurt, Germany Katalin Felvinczi1 & Annemieke Benschop2 & Róbert Urbán1 & Marie Claire Van Hout3 & Katarzyna Dąbrowska4 & Evelyn Hearne5 & Susana Henriques6 & Zsuzsa Kaló1 & Gerrit Kamphausen7 & Joana Paula Silva6 & Łukasz Wieczorek4 & Bernd Werse7 & Michal Bujalski4 & Zsolt Demetrovics1 & Dirk Korf2 1 Institute of Psychology, ELTE Eötvös Loránd University, Izabella utca 46, Budapest H-1064, Hungary 2 Bonger Institute, University of Amsterdam, Amsterdam, Netherlands 3 Public Health Institute, Liverpool John Moores University, Liverpool, UK 4 Institute of Psychiatry and Neurology, Warsaw, Poland 5 School of Health Sciences, Waterford Institute of Technology, Waterford, Ireland 6 Centre for Research and Studies in Sociology, University Institute of Lisbon, Lisbon, Portugal 7 Goethe-Universität, Frankfurt, Germany 7 Goethe-Universität, Frankfurt, Germany
https://openalex.org/W3133603749	https://estudogeral.uc.pt/bitstream/10316/96258/1/Nos%20e%20os%20outros_as%20migra%c3%a7oes%20no%20Portugal%20contemporaneo.pdf	Portuguese	null	Nós e os outros: as migrações no Portugal contemporâneo	Língua lugar	2,021	cc-by	7,670	Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Faculdade de Economia da Universidade de Coimbra e Centro de Estudos Sociais • pedro.gois@uc.pt DOI https://doi.org/10.34913/ journals/lingualugar.2020.e417 25 Há 50 anos Portugal era um país de emigração que tinha alguns imigrantes. Após o anunciado fim da emigração constatamos que atravessámos vários ciclos de emigração e retorno, mas que nunca os fluxos de saída deixaram de ter consequências sociais e sociológicas. Afinal a emigração é mais estrutural do que pensáramos. Hoje é um país de migrações. Entre o retorno ou repatriamento de muitos nacionais portugueses e o acolhi- mento de centenas de milhares de estrangeiros a demografia nacional ganhou diversidade e complexidade. Sem a imigração seríamos menos, mais pobres e mais velhos. A dinâmica e diversidade das origens dos migrantes para Portugal, mas, também, a geografia múltipla dos destinos dos emigrantes portugueses representa um sinal de alteração do posicio- namento do país no sistema migratório global. Numa tradicional lógica evolucionista, Portugal (ainda) não é um centro, mas (já) não é periferia (ou talvez o seja para alguns migrantes). Numa nova lógica de sistemas paralelos talvez as migrações portuguesas possam coexistir num país de emigração e de imigração. Sinal claro deste estatuto é, por exemplo, o facto da lei de nacionalidade ter evoluído, ao sabor de ideologias mais ou menos inclusivas e alargado o número de cidadãos que hoje fazem parte da comunidade nacional. Somos mais do que a soma dos que vivem em Portugal com os que dele partiram. Uma sociedade em movimento, aberta, plena de dinâmicas migratórias que permite antever um futuro cheio de desafios de integração e de gestão da diversidade. Uma sociedade dialogante com o mundo e em interação constante é, seguramente, uma sociedade em crescimento. Palavras-chave: Emigração; imigração; neo-semi-periferia; “comunidade nacional inclusiva”; nação-plural. Revista Língua−lugar N.02 dezembro 2020 26 Il y a 50 ans, le Portugal était un pays d'émigration qui ne comptait que quelques migrants. Après la soi-disant fin de l’émigration, il y a une trentaine d’années, on en éprouve encore les conséquences sociales et sociologiques. Après tout, la migration est plus structurelle qu'on ne l'a cru. Aujourd'hui, le Portugal est un pays de migrations, au pluriel. Entre le retour ou le rapatriement de nombreux Portugais et l'accueil de centaines de milliers d'étrangers, la démographie nationale a gagné en diversité et en complexité. Sans immigration, nous serions moins nombreux, plus pauvres et plus vieux. Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Revista Língua−lugar N.02 dezembro 2020 Nós e os outros: as migrações no Portugal contemporâneo No decurso das últimas décadas, Portugal, a exemplo de outros países da Europa do Sul, designadamente a Espanha, a Itália ou a Grécia, vem realizando uma transição de uma variável demográfica fundamental: as migrações, num percurso de tradicional "país de emigração" a "país de imigração" (Esteves, 1991; Baganha, 1996). Embora os indicadores demográficos não sejam permanentes nem definitivos parece claro que a atratividade do país para migrantes estrangeiros tem sido coerente e estável ao longo das últimas três décadas, isto é, pelo menos desde o final dos anos 80 do século XX. 27 A transição entre emigração e imigração, porém, não ocorreu de forma brusca ou definitiva, sendo percetível através da evolução de múltiplos indicadores que não apenas indicadores demográficos. Por exemplo, a dependência das remessas financeiras dos emigrantes, típica dos anos 60 e 70, foi ultrapassada por novas formas de remessas, como as da União Europeia (UE), o Investimento Direto Estrangeiro (IDE) ou as expor- tações de bens transacionáveis, sem que as remessas tenham deixado de representar um peso importante na economia, mormente na economia regional ou local (Mateus, 2013; Silva, 2006). Entre os anos 80 do século passado e a atualidade as remessas financeiras permaneceram elevadas e importantes em termos absolutos, mas, de certa forma, perderam a sua importância relativa na economia nacional (Ferreira, Lahr, Ramos e Castro, 2020). Para os autores que trabalharam, à época, estas questões, Portugal era, no final dos anos 1990, um autodefinido “país de imigração” imaginando-se a si próprio como um país que realizara a transição entre a emigração (que caraterizava um estado de subdesenvolvimento estrutural) e um país de imigração, um país do centro que abandonava a periferia e se tornava central (Baganha, 1991; 1992). Embora esta transição não tenha sido inteiramente realizada, por motivos que explicaremos adiante, é aceite de forma consensual que Portugal se encontra, ao longo dos últimos 30 anos, num momento de transformação quanto à sua posição no sistema migratório global (de Haas et al., 2019; Haas, Miller e Castles, 2020). Nós e os outros: as migrações no Portugal contemporâneo La dynamique et la diversité des origines des migrants au Portugal, mais aussi la géographie des destinations des émigrés portugais représente un signe d'un changement dans la position du pays dans le système migratoire mondial. Dans une logique évolutive traditionnelle, le Portugal n'est pas (encore) un centre, mais n’est pas (ou n'est plus) une périphérie. Dans une nouvelle logique de systèmes parallèles, peut-être, les migrations portugaises pourront coexister dans un pays de migration et d'immigration. Un signe clair de ce statut est, par exemple, le fait que la loi sur la nationalité a évolué pour devenir vraiment inclusive et progres- siste. Nous, les Portugais, sommes une communauté imaginée représen- tant beaucoup plus que la somme de ceux qui vivent au Portugal et de ceux qui en sont partis. Une société en mouvement, ouverte, pleine de dynamiques migratoires, qui permet toujours un avenir plein de défis d'in- tégration et de gestion de la diversité. Une société en dialogue avec le monde et en interaction permanente avec lui. Mots-clefs: Émigration; immigration; néo-semi-périphérie; «communauté nationale inclusive»; nation plurielle. Os fluxos migratórios, internos ou internacionais, são uma das principais forças de transformação social em todo o mundo capazes de engran- decer diversas questões de ordem económica, jurídica, política, social ou cultural (Baganha, Marques e Góis, 2010; Castles, 2005). Também em Portugal, vistas retrospetivamente, é possível afirmar que as migrações ajudaram a construir o país tal como o conhecemos hoje. Na verdade, é virtualmente impossível pensar um Portugal contemporâneo sem abordar, em conjunto, a emigração e a imigração e a forma como estas dimensões demográficas e sociológicas moldaram a sociedade portuguesa ao longo do último século (Góis e Marques, 2018). Se hoje somos uma sociedade diversa e feita de culturas múltiplas, devemos às migrações muita dessa variedade. Revista Língua−lugar N.02 dezembro 2020 27 No decurso das últimas décadas, Portugal, a exemplo de outros países da Europa do Sul, designadamente a Espanha, a Itália ou a Grécia, vem realizando uma transição de uma variável demográfica fundamental: as migrações, num percurso de tradicional "país de emigração" a "país de imigração" (Esteves, 1991; Baganha, 1996). Embora os indicadores demográficos não sejam permanentes nem definitivos parece claro que a atratividade do país para migrantes estrangeiros tem sido coerente e estável ao longo das últimas três décadas, isto é, pelo menos desde o final dos anos 80 do século XX. Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Neste contexto, Portugal foi sendo, ao longo das últimas décadas, encarado e/ou conceptualizado de diversas formas por diferentes autores: como espaço de transição, espaço de circulação ou plataforma de redis- tribuição de diferentes fluxos migratórios, como semiperiferia no sistema migratório global, como país de desenvolvimento médio numa rota de passagem entre origem e destinos diversos; como porta de entrada na UE para imigrantes de países terceiros (Góis e Marques, 2018). Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Revista Língua−lugar N.02 dezembro 2020 28 Complementarmente, Portugal tem sido destino final de diversos fluxos migratórios, alguns já tradicionais (como os imigrantes dos países africanos de língua oficial portuguesa [PALOP], Índia ou China), outros mais recentes como os imigrantes magrebinos, paquistaneses, nepaleses ou do Leste da Europa (ex. Rússia, Moldávia ou Ucrânia) (Baganha, Marques e Góis, 2005). Simultaneamente, subsiste um Portugal de emigração que, baseado numa história migratória centenária, se reconstrói geração após geração encontrando fora do país espaço para uma empregabili- dade, para uma carreira profissional ou, simplesmente, para uma fuga à pobreza que o país não sabe enquadrar (Peixoto et al., 2019). Complementarmente, Portugal tem sido destino final de diversos fluxos migratórios, alguns já tradicionais (como os imigrantes dos países africanos de língua oficial portuguesa [PALOP], Índia ou China), outros mais recentes como os imigrantes magrebinos, paquistaneses, nepaleses ou do Leste da Europa (ex. Rússia, Moldávia ou Ucrânia) (Baganha, Marques e Góis, 2005). Simultaneamente, subsiste um Portugal de emigração que, baseado numa história migratória centenária, se reconstrói geração após geração encontrando fora do país espaço para uma empregabili- dade, para uma carreira profissional ou, simplesmente, para uma fuga à pobreza que o país não sabe enquadrar (Peixoto et al., 2019). Nos últimos 50 anos Portugal assistiu: i) a diferentes ciclos de emigração e retorno; ii) ao repatriamento dos portugueses residentes nas ex-colónias; iii) a uma migração de refugiados da guerra colonial; iv) a diferentes ciclos e tipos de imigração laboral (de trabalhadores qualificados e de trabalhadores indiferenciados), de reagrupa- mento familiar e de inativos (estudantes ou reformados). ii) ao repatriamento dos portugueses residentes nas ex-colónias; Uma análise histórica da evolução destes fluxos migratórios ajuda-nos a compreender esta realidade e a caraterizar a forma como as migrações em Portugal moldaram a sociedade portuguesa. Comecemos por enunciar a dimensão emigratória portuguesa. Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros. Os últimos 50 anos de migrações portuguesas na Europa A história contemporânea portuguesa é marcada por dois grandes ciclos migratórios: o ciclo transatlântico e o ciclo intraeuropeu. O ciclo tran- satlântico desenvolveu-se ao longo do século XIX, atingiu o seu apogeu na viragem do século XIX-XX e, no que concerne à emissão massiva de emigrantes, entrou em declínio no período entre as duas grandes guerras. Estima-se que durante este primeiro ciclo mais de cinquenta milhões de europeus, dos quais aproximadamente dois milhões de portu- gueses, tenham abandonado os seus países de origem para se fixarem nas Américas. As causas para a decadência deste ciclo migratório são múltiplas. Destacamos, o recrudescimento dos sentimentos nacionalistas nos Estados Unidos, nomeadamente no início do século XX, e a eclosão e desenvolvimento da Grande Depressão de 1929. Estes eventos concer- tados levaram à implementação de políticas antimigratórias nos principais países de imigração do Continente Americano, que irão pôr fim ao ciclo das Grandes Migrações Transatlânticas (Baganha, 1990) e abrir espaço para o posterior desenvolvimento de um ciclo migratório europeu. Revista Língua−lugar N.02 dezembro 2020 29 Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Migrações após 1945 A necessidade de reconstruir uma Europa devastada pela guerra criou, no entretanto, novas possibilidades, dando início ao segundo grande ciclo migratório europeu, cuja primeira vaga perdurará até à crise petrolífera de meados dos anos 70 e cuja principal característica foi a sua dimensão intraeuropeia. Também aqui há causas diversas. Durante este segundo ciclo migratório o mercado de trabalho dos países europeus industriali- zados segmentou-se, sendo a procura no mercado secundário progres- sivamente satisfeita por mão-de-obra estrangeira originária na sua esmagadora maioria dos países da Europa do Sul. As políticas restritivas de imigração implementadas na sequência da crise petrolífera de 1973-74 e da recessão económica que se lhe seguiu, marcam o início de uma nova fase nos processos migratórios europeus, mas, até àquela data, as políticas de imigração podem caracterizar-se como políticas de "porta aberta" (Devoto e González-Bernaldo, 2001). Durante a fase de crescimento sustentado do pós-guerra, a Europa indus- trializada levou a cabo uma política de recrutamento de trabalhadores estrangeiros. Incentivou a vinda de vários milhões de imigrantes e de seus familiares cuja fixação foi facilitada pelas necessidades de mão-de-obra existentes, pelas possibilidades de mobilidade económica e social que daí advinham para os nacionais, bem como e sobretudo, pela convicção generalizada de que esta situação era temporária e poderia ser facilmente invertida, uma vez resolvidos os desequilíbrios conjunturais do mercado de trabalho, ou logo que os imigrantes, amealhadas as poupanças necessárias ou confrontados com situações de desemprego, retornassem aos seus países de origem. A ideia da temporalidade das migrações era defendida como a solução para necessidades conjunturais de mão-de- -obra e o sistema, sob o título de trabalhadores convidados, continha em si uma ideia nacionalista de mercado de trabalho (King, 1995). Portugal, só se encontra substancialmente envolvido no ciclo migra- tório intraeuropeu a partir da década de 60 do século XX e, mais concretamente, após a celebração dos acordos bilaterais entre o governo de Salazar e alguns governos europeus para fornecimento de mão-de-obra nacional, que ocorreram com a França (em 1963), com a Holanda (em 1963) e com a República Federal da Alemanha (em 1964). A emigração para França de trabalhadores do Sul da Europa é um bom exemplo das similitudes e diferenças (ou até de complementaridade) que caracterizam os emigrantes destes países. Entre 1950 e 1959 os emigrantes italianos representavam mais de metade do total de afluxo de Revista Língua−lugar N.02 dezembro 2020 30 mão-de-obra estrangeira. Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Revista Língua−lugar N.02 dezembro 2020 Migrações após 1945 Em 1960, o número de entrada de espanhóis em França estava já equiparado ao número de italianos, contribuindo cada uma destas nações com 30 mil migrantes num total de 73 mil entradas. Entre 1961 e 1965 os espanhóis substituíram os italianos como principal mão-de-obra estrangeira em França. Nos anos seguintes, de 1966 a 1972, Portugal substituiu a Espanha, como o principal fornecedor de mão-de- -obra imigrante à França. Esta contribuição foi particularmente significa- tiva nos anos de 1970 e 1971 em França (Antunes, 1970; 1973). Num total de 255 mil imigrantes em 1970 e de 218 mil em 1971, Portugal contribuiu com 53% (136 mil) e 51% (111 mil migrantes) respetivamente (Góis e Marques, 2018). Um outro facto importante tem a ver com a relevância da componente de emigração clandestina, as saídas do país de forma oculta e não declarada no total da emigração. O número de saídas totais (isto é, incluindo as estimativas feitas para saídas de clandestinos) entre 1931-1974 foi sensivel- mente de dois milhões de partidas. A média anual de partidas legais foi de 31 mil emigrantes e a média anual de partidas clandestinas de 14 mil emigrantes. A distribuição das saídas de clandestinos foi, contudo, particularmente assimétrica uma vez que mais de três quartos dessas partidas ocorreu entre 1964 e 1973 refletindo as alterações sociopolíticas em Portugal com o desenvolvimento da guerra colonial e da imposição de um serviço militar obrigatório (SMO) para todos os jovens do sexo masculino maiores de idade. A fuga a este SMO foi feita, para muitos destes jovens, através da estratégia migratória tendo sido França o destino principal. Votar com os pés foi, também neste caso, uma realidade soci- ológica bem destacada. Como se pode ver pela análise do Quadro 1 [ver página seguinte] a opção pela Europa ganha alguma relevância a partir de 1957 e torna-se dominante para a maior parte dos emigrantes portugueses a partir de 1962. Desde o final dos anos 50 até 1974 a atração exercida pela França e, em menor grau, pela Alemanha, domina completamente a emigração portu- guesa. Assim, do milhão e quatrocentas mil saídas verificadas entre 1962 e 1974, 75% dirigiram-se para estes dois países, respetivamente 62% para a França e 13% para a Alemanha. Esta viragem para as migrações intraeu- ropeias não mais abandonou a história da emigração portuguesa. Migrações após 1945 Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020 31 14.143 28.104 41.518 32.159 29.943 18.486 16.814 19.931 19.829 16.400 12.451 16.073 13.555 11.281 4.929 3.051 2.607 3.271 3.512 2.537 1.669 1.200 1.158 890 729 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 938 676 582 1.455 1.918 1.328 1.503 1.628 1.596 4.569 5.679 3.370 2.425 2.922 1.601 1.852 13.357 11.516 10.841 13.111 9.726 8.839 7.574 8.160 9.540 1.612 4.158 1.619 3.961 4.895 2.635 2.739 3.424 4.770 5.197 6.795 6.615 6.833 6.502 6.529 6.983 6.845 7.403 11.650 21.491 33.341 46.544 39.026 40.234 28.690 26.072 32.150 29.207 29.780 28.513 27.499 24.376 22.420 17.232 17.557 33.266 28.584 27.014 27.383 22.659 21.962 20.122 22.091 25.822 319 418 650 690 747 1.336 1.851 4.640 6.264 4.838 6.434 10.492 16.798 29.843 51.668 60.267 63.611 59.597 58.741 110.614 135.667 110.820 68.692 63.942 37.727 1 2 4 4 6 5 2 6 54 277 1.393 2.118 4.771 12.197 11.250 4.070 8.435 15.406 22.915 24.273 24.946 38.444 13.352 81 254 209 246 205 121 167 99 127 130 158 304 435 837 1.905 1.467 3.868 2.461 2.037 2.269 1.964 1.418 1.785 5.255 3.958 401 674 863 936 956 1.457 2.024 4.744 6.393 4.974 6.646 11.073 18.626 32.798 58.344 73.931 78.729 66.128 69.213 128.289 160.546 136.511 95.423 107.641 55037 21.892 34.015 47.407 39.962 41.190 30.147 28.096 36.894 35.600 34.754 35.159 38.572 43.002 55.218 75.576 91.488 111.995 94.712 96.227 115.672 183.205 158.473 115.545 129732 80859 1.83% 1.98% 1.82% 2.34% 2.32% 4.83% 7.20% 12.86% 17.96% 14.31% 18.90% 28.71% 43.31% 59.40% 77.20% 80.81% 70.30% 69.82% 71.93% 82.41% 87.63% 86.14% 82.59% 82.97% 68.07% Brasil EUA Canadá Total América França Alemanha Outros Europeus Total Europa Total Europa (%) Quadro 1. ano 31 Quadro 1. Emigração Portuguesa por destinos 1950-1974. Fonte: Adaptado de Baganha, 1994. Fonte: Adaptado de Baganha, 1994. Após o período de relativa estagnação que se seguiu à crise económica de 1973 e à revolução de 25 de Abril de 1974, a mobilidade externa da população portuguesa conheceu um ressurgimento a partir de meados dos anos 80 que se intensificou durante as décadas posteriores em ciclos sucessivos de retração e expansão (Marques e Góis, 2013). Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Migrações após 1945 Após o período de relativa estagnação que se seguiu à crise económica de 1973 e à revolução de 25 de Abril de 1974, a mobilidade externa da população portuguesa conheceu um ressurgimento a partir de meados dos anos 80 que se intensificou durante as décadas posteriores em ciclos sucessivos de retração e expansão (Marques e Góis, 2013). Neste novo período de expansão dos anos 80 já serão outros os destinos prevalecentes, designadamente a Suíça, a Alemanha (sobretudo após 1989) e o Luxemburgo. Os dados apresentados no quadro seguinte sintetizam a evolução da emigração portuguesa para destinos sele- cionados da Europa a partir de 1990. É possível notar que, em termos Revista Língua−lugar N.02 dezembro 2020 32 agregados, a entrada de portugueses regista durante a segunda metade da década de 1990 uma diminuição face ao quinquénio anterior, influ- enciada, sobretudo, pela desaceleração económica verificada em alguns dos principais destinos da emigração portuguesa (em particular na Suíça). Este período de retração prolonga-se, com diferentes inten- sidades e com diferentes ritmos de recuperação até ao início do século XXI. As saídas permanentes de emigrantes para países como a França ou Suíça das décadas anteriores foram-se permutando em saídas de tempo- ralidades mais curtas, circulares ou consecutivas para um maior conjunto de países (Peixoto et al., 2019). A partir do ano 2000, a evolução é generi- camente positiva e com uma nova fase de intensificação da emigração após a primeira década do século XXI. Alemanha Espanha Luxemburgo Reino Unido Suíça França Bélgica Total 1990-1995 98561 - 16924 - 95656 17566 11717 240424 1996-2000 102914 7394 10181 5286 28095 31965 7436 193271 2001-2005 34657 34598 16220 50474 52244 21128 8578 217899 2006-2010 29074 82110 20401 59030 72235 - 13042 275892 2011-2015 50692 31488 22117 129761 81406 36850 17742 370056 Quadro 2. Evolução da emigração para países Europeus selecionados (1990-2015) Góis e Marques, 2018. Resumindo. Quando analisamos as principais alterações dos fluxos migratórios ao longo das últimas décadas são várias as características que se destacam. Em primeiro lugar, desde há muito que a emigração portuguesa é, sobretudo, uma emigração laboral que busca em destinos bem definidos um alargamento do mercado de trabalho português, onde os rendimentos são baixos e a estrutura de carreiras é muito deficiente. Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Migrações após 1945 Em segundo lugar, atualmente, e, diferentemente do verificado noutros períodos emigratórios, caracterizados pela existência de um país de destino dominante em cada um dos períodos (que marcaram inclusive a sua denominação: o "ciclo brasileiro", o "ciclo francês"), a emigração Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020 33 portuguesa que se desenvolve a partir do ano 2000 apresenta uma geografia de destinos mais diversificada (Malheiros, Marques e Góis, 2016). Em terceiro lugar, sobretudo a partir de 2005, a emigração para os países da Europa Central e do Norte (responsável por, sensivelmente, dois terços das saídas totais), é complementada por fluxos migratórios que se dirigem para o hemisfério Sul para países como Angola, Brasil ou Moçambique. A emigração portuguesa para estes destinos assinala uma modificação na posição de Portugal no designado sistema migratório lusófono (Baganha, 2009; Góis e Marques, 2009; Peixoto, 2004) e em geral, um reposicio- namento no sistema migratório europeu (Góis e Marques, 2020) ou no sistema migratório global. Nós e os outros: as migrações no Portugal contemporâneo Revista Língua−lugar N.02 dezembro 2020 Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis De país de emigração a país de imigração e vice-versa A discussão e o equilíbrio entre a perenidade e temporalidade dos fluxos migratórios e a permanência de elevado número de portugueses no exterior é uma lição que a história nos tem ajudado a construir. Fernando Luís Machado (1999), por exemplo, afirmava no final dos anos 1990 que Portugal não devia ser visto como um país só de imigração uma vez que, numa escala muito mais reduzida e em moldes diferentes dos do passado, a emigração continuava a ter uma expressão não negligenciável (Machado, 1999). Félix Neto reforçava esta ideia afirmando que “Portugal é hoje (1993) simultaneamente um país de emigração e de imigração” (Neto, 1993). Outros autores afirmam que convém não esquecer que, do ponto de vista quantitativo, se bem que as estatísticas disponíveis rela- tivamente aos fluxos migratórios nem sempre deem conta da realidade, Portugal tem muito mais emigrantes que imigrantes e, neste sentido, continua a ser um país de emigração (Rocha-Trindade, 2000). A transição em curso, de "país de emigração" para "país de imigração" é, no que respeita à imigração, o resultado cumulativo de: i) repatriamento dos portugueses residentes nas ex-colónias e uma migração de refugiados da guerra colonial em meados dos anos 70; ii) diferentes ciclos e tipos de imigração (trabalhadores, reagrupamento familiar e de inativos (estudantes ou reformados) sobretudo a partir de final dos anos 80. Por muito que a imigração tenha contribuído para um reposicionamento de Portugal no sistema migratório global, a estruturalidade da emigração foi sempre mais forte revelando um posicionamento semiperiférico da sociedade e economia portuguesas (Góis e Marques, 2009). Este debate faz hoje parte da sociedade portuguesa e os seus ecos encontram-se bem presentes na estrutura da governação política destes fenómenos. 34 Deixando de ser países de emigração, Espanha, Itália, Grécia e Portugal passam a integrar a Europa da imigração, formando uma "nova porta de entrada", transposta por imigrantes africanos, sul-americanos, do médio oriente e asiáticos, que tanto podem procurar fixar-se no país de chegada, como usá-lo enquanto passagem para outros destinos (Peixoto et al., 2016). Por outro lado, a geografia das origens da imigração é também de destacar. O facto de três países da Europa do Sul (Itália, Espanha e Portugal), não sendo clássicos recetores de imigrantes, aparecerem entre os quatro primeiros lugares da lista dos países com maior número de imigrantes do hemisfério sul é um dado revelador das mudanças em curso no sistema migratório europeu (Machado, 1997, p. Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Revista Língua−lugar N.02 dezembro 2020 De país de emigração a país de imigração e vice-versa 37) e do novo posicionamento estratégico destes países no sistema migratório global. O facto de a imigração laboral em Portugal se cingir até final dos anos 1990 quase só aos países lusófonos sugeria que o país ocupava então, não um lugar de primeira linha nas rotas migratórias que se dirigem à União Europeia como um todo, mas um nicho que só era procurado pelos imigrantes dos palop e do Brasil e relativamente ao qual não se revelava atrativo para outros imigrantes. A imigração portuguesa parecia assim ter um carácter quase "doméstico" ou "caseiro", tudo se passando entre uma mesma família internacional de países (Machado, 1997, p. 39). A partir dos anos 2000, Portugal integra-se por inteiro no sistema global das migrações e vai competir pelos trabalhadores migrantes como qualquer outro país. A sua posição semiperiférica assume, deste modo, novos contornos, não abandonando a lógica de plataforma redistributiva ou de destino final para os migrantes dos países com os quais detinha relações coloniais (ex. PALOP ou Brasil), torna-se destino de atração para migrantes fora da sua lógica tradicional de influência (ex. imigrantes da Europa de Leste) e vê, deste modo, crescer quer o volume total de imigrantes quer o número de nacionalidades que compõem a imigração em Portugal. Muitos destes migrantes escolhem Portugal como porta de entrada na EU, outros, porém, escolhem Portugal como destino final porque conjunturalmente é o que melhor corresponde às suas expecta- tivas (ex. existência de trabalho, possibilidade de legalização, facilidade de permanecer ilegal, etc.). Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020 Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Revista Língua−lugar N.02 dezembro 2020 Evolução recente da população estrangeira em Portugal 35 Os dados do X Recenseamento Geral da População realizado em 1960 indicam que em 1960 residiam no país 29.579 cidadãos de nacionalidade estrangeira, sendo a maioria destes de origem Europeia (67%) e cerca de 40% espanhóis (sobretudo galegos). Dos restantes estrangeiros, 22% eram de nacionalidade brasileira e, somente, 1,5% provinham de um país africano (Esteves, 1991, p. 161). A revolução de 25 de Abril de 1974 que pôs fim à ditadura, os subsequentes processos de descolonização e a inde- pendência dos atuais países africanos de língua portuguesa marcaram uma alteração profunda dos movimentos migratórios com destino a Portugal, assistindo-se ao regresso de centenas de milhar de cidadãos daqueles territórios, originários da então metrópole e dos seus descen- dentes ali nascidos (Pires, 1984). Entre 1975 e 1980 a população estrangeira passou de 32.000 para 58.000, passando a ser constituída maioritariamente por cidadãos de origem africana (48%), grande parte dos quais (98%) provenientes das antigas colónias ultramarinas portuguesas em África (Angola, Cabo Verde, Guiné- -Bissau, Moçambique e São Tomé e Príncipe). O número de estrangeiros a residir em território nacional correspondia, em 1981, a apenas 0,6% do total da população residente (Baganha e Góis, 1998/1999). Durante os anos 80 nota-se um novo aumento significativo na população estrangeira a residir em Portugal, ultrapassando, no final da década, pela primeira vez a centena de milhar (ou 1% da população). O aumento da população estrangeira tornou-se ainda mais intenso no decurso dos anos 90 em que a proporção de estrangeiros na população total passou de 1,1%, em 1990, para 1,9%, em 1999. Esta evolução foi particu- larmente sentida entre a população proveniente do continente africano e do continente europeu que, em 1999, representavam 76,6% do total de imigrantes presentes em território nacional. O somatório de todos os imigrantes provenientes de um país de língua portuguesa mostra que este grupo de países representava, em 1999, aproximadamente 55% da população estrangeira a residir legalmente em Portugal, o que constituía um importante indicador do significado do passado colonial português e da manutenção de contactos sociais e culturais entre estes países e Portugal na constituição e consolidação deste fluxo migratório (Baganha e Góis, 1998/1999). Evolução recente da população estrangeira em Portugal Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020 1980 1985 1990 1995 1996 1997 1998 1999 2000 2001() 2002() 2003() 2004() 2005() 2006() 2007() 2008() 2009(*) 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Total de 58.091 79.594 107.767 168.316 172.912 175.263 178.137 190.896 207.607 350.503 286.601 259.794 264.880 296.000 409.185 435.736 440.277 454.191 445.262 436.822 417.042 398.268 390.113 383.759 392.969 416.682 477.472 588.976 Imigrantes 27.748 34.978 45.255 79.231 81.176 81.717 83.065 89.516 98.754 126.702 122.352 120.694 123.149 153.334 149.982 147.959 127.476 121.852 108.671 105.340 102.389 100.845 98.948 93.543 88.157 85.887 89.771 102.617 África 9.405 19.555 26.369 36.720 36.516 35.847 34.826 35.936 37.590 39.018 40.535 42.509 44.879 47.624 58.708 72.387 115.549 124.667 127.872 120.172 114.540 100.296 94.392 89.728 89.462 94.108 117.965 168.271 América 1.153 2.564 4.154 6.730 7.140 7.192 7.419 7.871 8.721 20.963 13.043 11.839 12.331 12.418 22.418 24.269 28.588 30.277 31.252 33.141 35.246 37.805 42.492 44.969 48.563 53.552 66.941 87.196 Ásia 17.706 22.060 31.410 44.867 47.315 49.747 52.060 56.731 61.709 138.061 98.106 81.113 83.656 83.940 153.307 179.040 167.790 176.561 176.911 177.608 164.335 158.992 153.936 155.137 166.414 182.694 202.298 230.285 Europa 260 438 579 768 765 760 767 789 803 931 850 844 836 838 878 876 356 341 321 316 320 324 339 334 360 - - - Outro Quadro 3. População estrangeira em Portugal por Continente de Origem, 1980-2019 Fonte: 1980-1995: INE, Estatísticas Demográficas e SEF, Estatísticas, cit. in Baganha, 1996 1996-2001: INE, Estatísticas Demográficas, 1996-2001 2001-2012: SEF, Estatísticas [http://sefstat.sef.pt/relatorios.aspx] 2013-2016: INE, INE, População estrangeira com estatuto legal de residente 2017-2019: INE \| SEF/MAI - População Estrangeira com Estatuto Legal de Residente via PORDATA Notas: () Inclui Autorizações de Residência e Autorizações de Permanência () Inclui Autorizações de Residência, Prorrogações de Autorizações de Permanência e Prorrogações de Vistos de Longa Duração () Inclui Autorizações de Residência e Prorrogações de Vistos de Longa Duração 36 Quadro 3. População estrangeira em Portugal por Continente de Origem, 1980-2019 Fonte: 1980-1995: INE, Estatísticas Demográficas e SEF, Estatísticas, cit. Evolução recente da população estrangeira em Portugal in Baganha, 1996 1996-2001: INE, Estatísticas Demográficas, 1996-2001 2001-2012: SEF, Estatísticas [http://sefstat.sef.pt/relatorios.aspx] 2013-2016: INE, INE, População estrangeira com estatuto legal de residente 2017-2019: INE \| SEF/MAI - População Estrangeira com Estatuto Legal de Residente via PORDATA Notas: () Inclui Autorizações de Residência e Autorizações de Permanência () Inclui Autorizações de Residência, Prorrogações de Autorizações de Permanência e Prorrogações de Vistos de Longa Duração () Inclui Autorizações de Residência e Prorrogações de Vistos de Longa Duração Quadro 3. População estrangeira em Portugal por Continente de Origem, 1980-2019 Fonte: 1980-1995: INE, Estatísticas Demográficas e SEF, Estatísticas, cit. in Baganha, 1996 1996-2001: INE, Estatísticas Demográficas, 1996-2001 2001-2012: SEF, Estatísticas [http://sefstat.sef.pt/relatorios.aspx] 2013-2016: INE, INE, População estrangeira com estatuto legal de residente 2017-2019: INE \| SEF/MAI - População Estrangeira com Estatuto Legal de Residente via PORDATA Notas: () Inclui Autorizações de Residência e Autorizações de Permanência () Inclui Autorizações de Residência, Prorrogações de Autorizações de Permanência e Prorrogações de Vistos de Longa Duração (*) Inclui Autorizações de Residência e Prorrogações de Vistos de Longa Duração Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Nós e os outros: as migrações no Portugal contemporâneo Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020 37 Como revela o quadro anterior, o stock da população estrangeira em Portugal cresceu ininterruptamente desde 1980 até 2001, ainda que, com ritmos de crescimento diferentes ao longo deste período. De facto, depois de um intenso crescimento na segunda metade dos anos 70, (a taxa anual média de crescimento foi de 11,9% entre 1975 e 1981,1 o ritmo de crescimento da fixação de estrangeiros abranda durante a década de 80, e acelera novamente nos anos 90. Apesar do cresci- mento da população estrangeira ser constante durante os anos 80 e 90, em 2000 o número de estrangeiros com residência legal era ainda de pouco mais de 200 mil pessoas, ou seja aproximadamente 2% da população do país.2 Acresce que, na viragem para o século XXI, a imigração de países terceiros para Portugal era maioritariamente (76% em 1999 e 77% em 2000) constituída de imigrantes de países lusófonos, isto é, das ex-colónias portuguesas em África e do Brasil. A percentagem restante encontrava-se dispersa por mais de cem nacionalidades, nenhuma das quais apresentava valores numéricos muito significativos (Serviço de Estrangeiros e Fronteiras (SEF), Estatísticas de 1999, Estatísticas de 2000). 2 Serviço de Estrangeiros e Fronteiras (SEF) (1999); Baganha e Marques (2001) e Serviço de Estrangeiros e Fronteiras (SEF) (Estatísticas de 2000). 1 Estatísticas Demográficas, INE,1995. Nós e os outros: as migrações no Portugal contemporâneo Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Evolução recente da população estrangeira em Portugal Apesar de um crescimento contínuo ao longo de 21 anos (1980-2001), a presença de imigrantes em Portugal era, até 2001, relativamente fraca e a maioria dos movimentos que ocorriam podiam atribuir-se diretamente ao passado colonial, a relações históricas e culturais bem como a relações económicas. 1 Estatísticas Demográf 2 Serviço de Estrangeir (SEF) (1999); Baganha e Serviço de Estrangeiros (Estatísticas de 2000). Com o novo milénio chegaram migrantes inesperados alterando toda a dinâmica imigratória no país e as políticas migratórias portuguesas (Góis e Marques, 2016). A emigração do Leste europeu para Portugal começou em 2000. Tendo em conta os padrões migratórios dos anos 80 e dos anos 90 esta mudança foi súbita e inesperada. Por ter sido, simultaneamente, repentino e intenso, transformou a composição e o panorama nacional da população imigrante (Baganha, Marques e Góis, 2004). Até esta época não existiam laços históricos, culturais ou económicos privilegiados dos países da Europa de Leste a que se pudessem atribuir o súbito e intenso movimento de imigrantes da Europa de Leste para Portugal. O ano 2000 marca o início de um processo de transformação das origens geográficas dos imigrantes que começam a provir maioritariamente da Europa do Leste, em especial da Ucrânia, e de um aumento significativo da imigração com origem no Brasil que daria origem a sucessivas vagas migratórias (Peixoto, Padilla, Marques e Góis, 2015). Elementos distintivos deste fluxo, em particular do de origem Leste Europeia, são a sua forte intensidade e a sua concentração num período temporal relativamente Revista Língua−lugar N.02 dezembro 2020 38 curto (2 a 3 anos, entre 2000 e 2002) com uma progressiva retração da vaga migratória e a permanência em Portugal de um número assinalável de imigrantes com esta origem. Trata-se de um fluxo que só se tornou estatisticamente importante após a concessão, ao abrigo do artigo 55.º do Decreto-Lei n.º 4/2001 de 10 de Janeiro, de 126.901 autorizações de permanência a trabalhadores imigrantes que se encontravam ilegal- mente no país. Em 2003, os imigrantes provenientes da Ucrânia passam a constituir o grupo mais numeroso, seguido pelos brasileiros e pelos cabo- -verdianos. Em conjunto estas três nacionalidades representavam nesse ano 52,6% do total de imigrantes de países terceiros a residir legalmente em Portugal. Evolução recente da população estrangeira em Portugal A preponderância destes três grupos nacionais no total da população estrangeira a residir legalmente em Portugal manteve-se até ao presente, alterando-se, somente, a importância relativa de cada uma das naciona- lidades que foi oscilando ao longo da última década e meia. Assim, em 2016, os imigrantes brasileiros representavam 20,4%, os cabo-verdianos 9,2% e os ucranianos 10,6% e do total de 397.731 imigrantes com residência legal em Portugal. Já em 2019, os ucranianos tinham passado para a quinta posição (5%) com os imigrantes com origem no Brasil (25,6%), Cabo Verde (6,3%), Reino Unido (5,8%) e Roménia (5,3%) a ocuparem as quatro primeiras posições (Góis e Marques, 2018). Ressalve-se porém o facto de, em muitos casos esta diminuição (ex. do número de ucranianos ou de cabo-verdianos) não corresponder a um retorno aos países de origem respetivos mas a uma diminuição estatística por via da aquisição de nacionalidade portuguesa a que se segue, num número indetermi- nado de casos, uma re-emigração para outros países europeus (Góis e Marques, 2018). Nós e os outros: as migrações no Portugal contemporâneo Conclusão Atualmente, em resultado de diversos fluxos imigratórios, a composição nacional dos estrangeiros a residir em Portugal é mais diversificada do que durante as décadas anteriores. Integra atualmente, para além de uma proporção significativa de imigrantes das ex-colónias portuguesas, cidadãos da União Europeia ou imigrantes provenientes do Brasil. O número de estrangeiros residentes aumentou ao longo de 21 anos 1980-2001, voltou a aumentar entre 2003 e 2009 ano em que atingiu o seu máximo de sempre. Desde então o número de imigrantes em Portugal reduziu-se continuamente até 2015 ano em que retomou uma trajetória ascendente com um novo máximo em 2019 (588.976 imigrantes legalmente residentes) (SEF, 2020). Este número continuou a subir nos Revista Língua−lugar N.02 dezembro 2020 39 primeiros meses de 2020, mas tem, devido aos efeitos da pandemia Covid-19, uma previsão de decréscimo para os tempos mais próximos. A este grupo de estrangeiros a residir em Portugal há a acrescentar o conjunto de indivíduos que, por diferentes razões, não constam, ou deixaram de constar, das estatísticas portuguesas relativas à população estrangeira. Referimo-nos aos imigrantes que adquiriram nas últimas décadas a nacionalidade portuguesa em resultado de um processo de naturalização, aos descendentes de imigrantes nascidos em território nacional que podem (ou não) deter a nacionalidade portuguesa e, por último, os cidadãos portugueses em que pelo menos um dos seus progen- itores tenha nascido fora do território nacional (ex. em Macau, Angola, Moçambique ou Goa). No período de 39 anos, entre a primeira Lei de Estrangeiros (1981) e a atualidade, ocorreu uma mudança significativa na paisagem imigratória nacional, indiciando, por um lado, uma aceleração e diversificação dos ciclos imigratórios e, por outro, um processo de heterogeneização da população estrangeira residente em várias das suas características princi- pais. Em resultado de diversos fluxos imigratórios, a composição nacional dos estrangeiros a residir em Portugal é atualmente mais diversificada do que foi durante as décadas anteriores. Por seu turno, a emigração continua bem presente na demografia portu- guesa. Os destinos migratórios alteraram-se e diversificaram-se sendo hoje predominantemente destinos europeus embora com a presença de importantes destinos fora da Europa. A dinâmica migratória não demonstra ainda tendências bem definidas com o saldo migratório (diferença entre o número de imigrantes e de emigrantes) a ser de novo positivo após quase uma década em que foi negativo. Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Revista Língua−lugar N.02 dezembro 2020 Conclusão Com o dealbar de uma nova crise económica em Portugal e o ressurgimento do desem- prego em massa a válvula estrutural que é a emigração tende a reemergir e o país será menos atrativo para a imigração. A paisagem humana em Portugal altera-se com as migrações e delas depende num país que envelhece ao sabor de um tempo em que nascem menos crianças e a esperança média de vida vai evoluindo positivamente. Não seríamos o que somos hoje sem as migrações, não seríamos nós sem os outros. Revista Língua−lugar N.02 dezembro 2020 40 Bibliografia Baganha, M. e Peixoto, J. (1996). "O Estudo das Migrações Nacionais: Ponto de Intersecção Disciplinar". In J. M. Carvalho Ferreira, R. Marques, J. Peixoto e R. Raposo (Ed.), Entre a Economia e a Sociologia (pp. 233-239). Oeiras: Celta. Antunes, M. L. M. (1970). "Vinte anos de emigração portuguesa: alguns comentários". Análise Social, 8 (30-31), 299-385. — (1973). A emigração portuguesa desde 1950: dados e comentários. Lisboa: Universidade Técnica de Lisboa. Instituto Superior de Economia. Gabinete de Investigações Sociais. Castles, S. (2005). Globalização, Transnacionalismo e Novos Fluxos Migratórios. Lisboa: Fim de Século. Baganha, M. (1991). "Uma Imagem Desfocada - a emigração portuguesa e as fontes sobre a emigração". Paper presented at the Análise Social. de Haas, H., Czaika, M., Flahaux, M.-L., Mahendra, E., Natter, K., Vezzoli, S., & Villares-Varela, M. (2019). "International Migration: Trends, Determinants, and Policy Efects". Population and Development Review, 45(4), 885-922. doi:10.1111/padr.12291 — (1992). Principais Características e Tendências da Emigração Portuguesa. Paper presented at the Actas do 2º — (1992). Principais Características e Tendências da Emigração Portuguesa. Paper presented at the Actas do 2º Congresso Português de Sociologia. Devoto, F., e González-Bernaldo, P. (2001). Emigration politique : une perspective comparative: Italiens et Espagnols en Argentine et en France, XIXe-XXe siècles. Paris: Harmattan. Congresso Português de Sociologia. — (2009). "The Lusophone Migratory System: Patterns and Trends". International Migration, 47(3), pp. 5-20. Esteves, M. C. (1991). Portugal: País de Imigração. Lisboa: Instituto de Estudos para o Desenvolvimento. Baganha, M., Marques, J. C. e Góis, P. (2010). Imigração Ucraniana em Portugal e no Sul da Europa: a emergência de uma ou várias comunidades? Lisboa: ACIDI IP. Ferreira, J.-P., Lahr, M., Ramos, P. e Castro, E. (2020). "Accounting for global migrant remittances flows". Economic Systems Research, 32(3), 301-317. doi:10.1080/095353 14.2019.1659756 Baganha, M. e Góis, P. (1998/1999). "Migrações internacionais de e para Portugal: o que sabemos e para onde vamos?" Revista Crítica de Ciências Sociais, 52/53, 229-280. Góis, P. e Marques, J. C. (2009). "Portugal as a semiperipheral country in the global migration system". International Migration, 47(3), 19-50. Baganha, M. e Marques, J. (2001). Imigração e Política: O caso Português. Lisboa: Fundação Luso-Americana. — (2016). "A emigração portuguesa e o sistema migratório lusófono : complexidade e dinâmicas de um país de migrações" In: Observatorio Iberoamericano sobre Movilidad Humana, Migraciones y Desarrollo <http://hdl.handle.net/10400.8/1759>. Baganha, M. I., Marques, J. C., Góis, P. (2004). "Novas migrações, novos desafios: A imigração do Leste Europeu". Bibliografia Revista Crítica de Ciências Sociais (69), pp. 95-115. — (2004). "The unforeseen wave: migration from Eastern Europe to Portugal". In M. — (2004). "The unforeseen wave: migration from Eastern Europe to Portugal". In M. I. Baganha & M. L. Fonseca (Eds.), New Waves: Migration from Eastern to Southern Europe (pp. 23-39). Lisboa: Luso-American Foundation. — (2018). Retrato de um Portugal migrante: a evolução da emigração, da imigração e do seu estudo nos últimos 40 anos. Coimbra, CES. I. Baganha & M. L. Fonseca (Eds.), New Waves: Migration from Eastern to Southern Europe (pp. 23-39). Lisboa: Luso-American Foundation. — (2020). "Portuguese intra-EU migration. The dynamics of an ongoing migration process". Racial and Ethnic Studies, (no prelo). — (2004). "Novas migrações, novos desafios: A imigração do Leste Europeu". Revista Crítica de Ciências Sociais, 69, pp. 95-115. Guibentif, P. (1996). "Le Portugal face à l'immigration". Revue Européenne des Migrations Internationales, 12 (1), 121-139. Guibentif, P. (1996). "Le Portugal face à l'immigration". Revue Européenne des Baganha, M., Marques, J. C. e Fonseca, G. (2000). Is an Ethclass Emerging in G. (2000). Is an Ethclass Emerging in Migrations Internationales, 12 (1), 121-139. Europe? The Portuguese Case. Lisboa: Luso American Development Foundation. American Development Foundation. Nós e os outros: as migrações no Portugal contemporâneo P Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020 41 Neto, F. (1993). Psicologia da migração portuguesa. Lisboa: Universidade Aberta. Haas, H. d., Miller, M. J., e Castles, S. (2020). The age of migration : international population movements in the modern world (Sixth edition / ed., pp. 1 online resource (1 volume)). Peixoto, J. (2004). "Dinâmicas e regimes migratórios: o caso das migrações internacionais em Portugal". Análise Social, XLII(183), 445-469. King, R. (1995). Mass migration in Europe: the legacy and the future. Chichester: Wiley. Peixoto, J., Candeias, P., Ferreira, B., Tiago de Oliveira, I., Azevedo, J., Marques, J. C. e Santana, E. (2019). "New Emigration and Portuguese Society: Transnationalism and Return" In: Springer %U http://hdl.handle. net/10451/39031. Machado, F. L. (1999). "Aspects et spécificités de l'immigration au Portugal". Migrances, 15, 82-93. Malheiros, J., Marques, J. C., e Góis, P. (2016). "Geografias, processos migratórios e dinâmicas sociográficas da emigração contemporânea portuguesa: respondendo a algumas questões…". In J. Peixoto, I. T. d. Oliveira, J. Azevedo, J. C. Marques, P. Góis, J. Malheiros, & P. M. Madeira (Eds.), Regresso ao Futuro: A nova Emigração e a Sociedade Portuguesa (pp. 273-296). Lisboa: Gradiva. — (2013). "Dinâmicas do sistema migratório lusófono: um olhar a partir das migrações portuguesas". In L. Fonseca, P. Góis, J. C. Marques, & J. Peixoto (Eds.), Migrações na Europa e em Portugal. Ensaios de homenagem a Maria Ioannis Baganha (pp. 185-203). Coimbra: Almedina. Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Silva, J. R. (2006). O investimento directo estrangeiro. In A. Romão (Ed.), A Economia Portuguesa - 20 Anos Após a Adesão (pp. 491-518). Coimbra: Almedina. Marques, J. C., e Góis, P. (2012). "A evolução do sistema migratório lusófono. Uma análise a partir da imigração e emigração portuguesa". Revista Internacional em Língua Portuguesa, 24, pp. 213-232. Mateus, A. (2013). 25 anos de Portugal europeu: a economia, a sociedade e os fundos estruturais. Lisboa: Fundação Francisco Manuel dos Santos. Bibliografia Peixoto, J., Padilla, B., Marques, J. C., e Góis, P. (2015). Vagas Atlânticas. Migrações entre Brasil e Portugal no Início do Século XXI. Lisboa: Mundos Sociais. Pires, R. P. (1984). Os Retornados: um estudo sociográfico. Lisboa: Instituto de Estudos para o Desenvolvimento. Rocha-Trindade, M. B. (2000). "História da Imigração em Portugal". Janus 2001 - Anuário de Relações Exteriores, Público/ UAL, 170-173. Marques, J. C., e Góis, P. (2012). "A evolução do sistema migratório lusófono. Uma análise a partir da imigração e emigração portuguesa". Revista Internacional em Língua Portuguesa, 24, pp. 213-232. Marques, J. C., e Góis, P. (2012). "A evolução do sistema migratório lusófono. Uma análise a partir da imigração e emigração portuguesa". Revista Internacional em Língua Portuguesa, 24, pp. 213-232. Serviço de Estrangeiros e Fronteiras (SEF). RIFA, vários anos, Lisboa, SEF. — (2013). "Dinâmicas do sistema migratório lusófono: um olhar a partir das migrações portuguesas". In L. Fonseca, P. Góis, J. C. Marques, & J. Peixoto (Eds.), Migrações na Europa e em Portugal. Ensaios de homenagem a Maria Ioannis Baganha (pp. 185-203). Coimbra: Almedina. Silva, J. R. (2006). O investimento directo estrangeiro. In A. Romão (Ed.), A Economia Portuguesa - 20 Anos Após a Adesão (pp. 491-518). Coimbra: Almedina. Mateus, A. (2013). 25 anos de Portugal europeu: a economia, a sociedade e os fundos estruturais. Lisboa: Fundação Francisco Manuel dos Santos. Nós e os outros: as migrações no Portugal contemporâneo Pedro Góis Nós e os outros: as migrações no Portugal contemporâneo Revista Língua−lugar N.02 dezembro 2020 Revista Língua−lugar N.02 dezembro 2020
https://openalex.org/W2439832344	https://discovery.ucl.ac.uk/id/eprint/1501647/1/journal.pone.0157605.PDF	English	null	Detection of Novel Integrons in the Metagenome of Human Saliva	PloS one	2,016	cc-by	10,031	RESEARCH ARTICLE a11111 Supathep Tansirichaiya1☯‡, Md. Ajijur Rahman1,2☯‡, Agata Antepowicz1¤, Peter Mullany1, Adam P. Roberts1* 1 Department of Microbial Diseases, UCL Eastman Dental Institute, University College London, 256 Gray's Inn Road, London, WC1X 8LD, United Kingdom, 2 Department of Pharmacy, University of Rajshahi, Rajshahi-6205, Bangladesh ☯These authors contributed equally to this work. ¤ Current address: Gene Medicine Research Group, Nuffield Division of Clinical Laboratory Sciences, Radcliffe Department of Medicine, John Radcliffe Hospital, University of Oxford, Oxford, United Kingdom ‡ These authors are joint first authors on this work. * adam roberts@ucl ac uk a1111 OPEN ACCESS OPEN ACCESS Citation: Tansirichaiya S, Rahman M.A, Antepowicz A, Mullany P, Roberts AP (2016) Detection of Novel Integrons in the Metagenome of Human Saliva. PLoS ONE 11(6): e0157605. doi:10.1371/journal. pone.0157605 Editor: Axel Cloeckaert, Institut National de la Recherche Agronomique, FRANCE Received: March 11, 2016 Accepted: June 1, 2016 Published: June 15, 2016 Editor: Axel Cloeckaert, Institut National de la Recherche Agronomique, FRANCE Received: March 11, 2016 Accepted: June 1, 2016 Published: June 15, 2016 Copyright: © 2016 Tansirichaiya et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Abstract Integrons are genetic elements capable of capturing and expressing open reading frames (ORFs) embedded within gene cassettes. They are involved in the dissemination of antibi- otic resistance genes (ARGs) in clinically important pathogens. Although the ARGs are common in the oral cavity the association of integrons and antibiotic resistance has not been reported there. In this work, a PCR-based approach was used to investigate the presence of integrons and associated gene cassettes in human oral metagenomic DNA obtained from both the UK and Bangladesh. We identified a diverse array of gene cassettes containing ORFs predicted to confer antimicrobial resistance and other adaptive traits. The predicted proteins include a putative streptogramin A O-acetyltransferase, a bleomycin binding protein, cof-like hydrolase, competence and motility related proteins. This is the first study detecting integron gene cassettes directly from oral metagenomic DNA samples. The predicted proteins are likely to carry out a multitude of functions; however, the function of the majority is yet unknown. OPEN ACCESS Citation: Tansirichaiya S, Rahman M.A, Antepowicz A, Mullany P, Roberts AP (2016) Detection of Novel Integrons in the Metagenome of Human Saliva. PLoS ONE 11(6): e0157605. doi:10.1371/journal. pone.0157605 Editor: Axel Cloeckaert, Institut National de la Recherche Agronomique, FRANCE Received: March 11, 2016 Accepted: June 1, 2016 Published: June 15, 2016 Detection of Novel Integrons in the Metagenome of Human Saliva Supathep Tansirichaiya1☯‡, Md. Ajijur Rahman1,2☯‡, Agata Antepowicz1¤, Peter Mullany1, Adam P. Roberts1* Introduction Integrons are commonly found in bacterial genomes, especially in most Gram-negative bacte- ria. They are involved in the dissemination and differential expression of genes in the bacterial population [1–3]. They contain two common features, a functional platform and an array of gene cassettes (GCs). The former or the 5’ conserved segment (5’CS) contains the integrase gene, intI, an attI recombination site and the promoter Pc. This platform is used for the captur- ing and expression of the GCs, non-replicative mobile elements which generally couple one or more open reading frames (ORFs) with the cassette-associated recombination attC site. The intI gene encodes a site-specific tyrosine recombinase, IntI which catalyses the integration and excision of the GCs. The expression of integrase genes can be upregulated by the SOS response, Data Availability Statement: All nucleotide sequence files are available from the NCBI database (accession number(s) KT921469 to KT921531). Funding: MAR; Commonwealth Scholarship Commission in the UK. Ref number: BDCA2013-4. 1 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome a bacterial stress response induced by the accumulation of single stranded DNA in a cell, such as transformation, conjugation, starvation and exposure to antibiotics such as quinolones and trimethoprim [4]. The recombination usually occurs between attI, located immediately adjacent to intI, and attC which is present on circular gene cassettes [5–7]. The attC sites contain two inverted regions of homology (L’-R’ and R”-L”), which flank the central region containing a highly vari- able sequence. The size of attC can be between 57 to 141 bp [8]. Even though the sequences of attC sites are not conserved, they show a palindromic organization which is essential for the formation of the correct hairpin structure, which is the recognition site of the integrase during integron GC recombination reactions [9]. The GCs normally do not have a promoter. The Pc promoter is usually required for the transcription of GC ORFs, therefore the first GC following Pc often has the higher levels of expression relative to downstream GC located ORFs [10]. Integrons have the potential to drive bacterial evolution and adaptation by differential expression of ORFs within GCs. One of the most clinically significant adaptive traits is antibi- otic resistance [11]. The first integrons were identified by their association with antibiotic resis- tance genes (ARGs) [12]. Introduction Among hundreds of classes of integrons, class 1 integrons are the most commonly associated with multiple ARGs in clinical strains. More than 130 different GCs carried by integrons were predicted to confer resistance to a variety of classes of antibiotics such as aminoglycosides, beta-lactams, chloramphenicol, trimethoprim, and streptothricin [13]. Gene cassettes are abundant and disseminated widely in diverse environments. Different isolates of the same bacterial species can have different GC arrays [14]. The predicted protein functions of ORFs within GCs are varied and include, in addition to antibiotic resistance, virulence, and secondary metabolism, which are likely to be niche-specific [1, 2]. However, metagenomic analyses of the integron cassette gene pool from several studies revealed that vast majority of GCs were novel [15–17]. Due to the fact that, in many environments, less than 1% of the bacterial population is cul- turable [18], one of the approaches to investigate the GCs in the entire bacterial community is the PCR-based amplification of GCs using metagenomic DNA as a template [19]. Several stud- ies on the diversity of GCs in different environments have been performed with this approach such as soil, seawater, marine sediment and deep sea vents [16, 17, 19, 20]. The human oral cavity is one of the most complex microbial ecosystems in the human body. More than 700 bacterial species have been detected from the oral cavity, [21, 22]. Many ARGs have been detected and discovered in the oral cavity, including tetracycline resistance genes tet(Q), tet(W), tet(M), tet(37) and tet(32); erythromycin resistance genes, ermB and mef, and kanamycin resistance gene, aphA-3 [23–25]. Recent genetic analysis of the oral meta- genome showed that 2.8% of the predicted genes had the potential to encode proteins with antibiotic and toxin resistance [26]. However, very few studies investigating integrons in human oral cavity have been performed. There are two major reports on integrons in the human oral cavity; one describing an unusual or reverse integron, an integron with the inte- grase gene oriented in the same direction as a gene cassette array, in Treponema denticola ATCC35405 by using whole genome sequencing analysis, and the in silico analysis of an inte- gron associated with Treponema species by using metagenomic datasets of the Human Micro- biome Project [14, 27]. The presence of other integrons in other oral bacterial species remains to be determined. PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Saliva sample collection and ethical approval Saliva samples were collected from 11 and 10 healthy volunteers (both male and female with age between 21 and 65) from UK and Bangladesh, respectively. The UK samples were col- lected from the staff and international postgraduate students from the UCL Eastman Dental Institute and represent various ethnic and cultural backgrounds including Asian, Australian, European, African and Middle-Eastern, some of which had moved to the UK in the past few months. Therefore, the UK samples represent an international metagenome. The Bangladeshi samples were collected from the staff, undergraduate and post-graduate students of Depart- ment of Pharmacy of Rajshahi University all of which were Bangladeshi. All of the volunteers read and gave written consents before sample collection. None of the volunteers had received antibiotic treatment for 3 months before the sample collection day. Ethical approvals for the analysis of pooled saliva as part of this project were obtained from University College London (UCL) Ethics Committee (project number 5017/001) and the Institutional Animal, Medical Ethics, Biosafety and Biosecurity Committee (IAMEBBC) for Experimentations on Animal, Human, Microbes and Living Natural Sources, University of Rajshahi (project number 54/ 320/IAMEBBC/IBSC). Both ethics committees approved the consent procedures for the sam- ple collection and processing. For the UK samples, 2 ml of saliva were collected in a sterile plastic tube and processed immediately. The samples from Bangladesh were collected and transported using Norgen’s Saliva DNA Collection and Preservation Device, (Norgen, Can- ada) following the manufacturer’s guidelines, and transported to UK for analysis. All samples were anonymised. Extraction of oral metagenomic DNA The freshly collected UK saliva samples were pooled together into a sterile plastic tube in a class I microbiological safety cabinet. The pooled saliva sample was then divided into 1.5ml aliquots and centrifuged at 20238 g for 1 min. The UK oral metagenomic DNA was then extracted by using the Puregene DNA extraction kit (Qiagen, UK), following the Gram-positive bacteria and yeasts protocol with the modification in final step, which the DNA pellets were dissolved in 400μL molecular grade water at room temperature, instead of 100μL. The Bangladeshi oral metagenomic DNA was extracted from the Norgen’s Saliva DNA stor- age buffer using ethanol precipitation technique according to manufacturer’s protocol. The preservative buffer of Norgen devices is designed for rapid cellular lysis and subsequent preser- vation of DNA from fresh saliva samples. Prior to DNA isolation, the storage devices were incubated for 1h at 50°C and mixed by inversion and gentle shaking for 10 seconds. DNA was then extracted from 500 μL of the pooled saliva in preservative buffer by taking 50 μL aliquots from 10 saliva samples. Introduction Despite the oral microbiota being recognised as a potential source of ARGs and the oral environment providing conducive conditions for the transfer of ARGs between a range of spe- cies [25], no in depth studies have been carried out to detect integrons and GCs within the oral microbiota. In this study, we have investigated the presence of integrons and associated GCs in PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 2 / 20 Integrons from the Human Saliva Metagenome the human oral metagenomic DNA from two countries, the UK and Bangladesh using a PCR approach. Different sets of primers targeting different regions of integrons were used for PCR amplification, in which multiple GCs were identified and predicted to encode various proteins including some likely to confer antibiotic resistance. PCR amplification The list of primers and their sequences are shown in S1 Table and the target sites for the prim- ers are shown in Fig 1. The typical PCR was prepared as follows; 50 μL reaction containing 15μl of 2x BioMix Red (Bioline, UK), 0.2 μM of each 10 μM primer, 50–100 ng of DNA 3 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome Fig 1. The primer binding sites of the published and newly designed primers. Primers were indicated as black arrows on A.) the class 1 integrons and B.) the unusual integron structure of T. denticola. The open arrowed boxes represent ORFs, pointing in the probable direction of transcription. The genes in 5’and 3’ conserved segment (CS), the open reading frame (ORF), the recombination site attI and attC are shown in grey, blue, green, yellow and orange respectively. doi:10.1371/journal.pone.0157605.g001 Fig 1. The primer binding sites of the published and newly designed primers. Primers were indicated as black arrows on A.) the class 1 integrons and B.) the unusual integron structure of T. denticola. The open arrowed boxes represent ORFs, pointing in the probable direction of transcription. The genes in 5’and 3’ conserved segment (CS), the open reading frame (ORF), the recombination site attI and attC are shown in grey, blue, green, yellow and orange respectively. Fig 1. The primer binding sites of the published and newly designed primers. Primers were indicated as black arrows on A.) the class 1 integrons and B.) the unusual integron structure of T. denticola. The open arrowed boxes represent ORFs, pointing in the probable direction of transcription. The genes in 5’and 3’ conserved segment (CS), the open reading frame (ORF), the recombination site attI and attC are shown in grey, blue, green, yellow and orange respectively. template, and molecular grade water (Sigma, UK) up to 30 μL. The standard PCR was carried out with (i) an initial denaturation: 94°C for 5 minutes, (ii) denaturation step: 94°C for 1 min- ute, (iii) annealing step: 50–65°C depending on the primers for 30 seconds, (iv) elongation step: 72°C for 30 seconds to 3 minutes depending on the size of expected products, repeated step (ii)-(iv) for 35 cycles and (v) final elongation step 72°C for 10 min. template, and molecular grade water (Sigma, UK) up to 30 μL. Ligation and transformation Purified PCR products were ligated into pGEM-T Easy vector (Promega, UK). The ligation mixtures were transformed into Escherichia coli α-Select Silver Efficiency competent cells (Bio- line, UK) by heat shock at 42°C for 40 s, and plated on Luria-Bertani (LB) agar with ampicillin (100 μg/mL) as a selective marker for the plasmids and 40 μg/ml X-Gal plus 0.4 mM IPTG for the blue-white colony screening. PCR purification and gel extraction The PCR products were subjected to electrophoresis on 1% agarose gel stained with 1:10,000 dilution of GelRed nucleic acid stain (Biotium, UK). The products were then purified by using either QIAquick PCR Purification Kit (Qiagen, UK) or QIAquick Gel Extraction Kit (Qiagen, UK), depending on the amplification results and the target amplicons, according to the manu- facturer’s instructions. PCR amplification The standard PCR was carried out with (i) an initial denaturation: 94°C for 5 minutes, (ii) denaturation step: 94°C for 1 min- ute, (iii) annealing step: 50–65°C depending on the primers for 30 seconds, (iv) elongation step: 72°C for 30 seconds to 3 minutes depending on the size of expected products, repeated step (ii)-(iv) for 35 cycles and (v) final elongation step 72°C for 10 min. PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 4 / 20 Integrons from the Human Saliva Metagenome Plasmid isolation and sequencing White colonies were subcultured in 5 mL of LB broth with ampicillin (100μg/mL) and incu- bated overnight. Plasmids were isolated by using QIAprep Spin Miniprep Kit (Qiagen, UK) fol- lowing the manufacturer’s instructions. The presence of the insert in a plasmid was verified by a 10μl DNA digestion reaction, containing 0.5 μL EcoRI restriction enzyme (20 units/μL, New England Biolabs, UK), 1μL 10x EcoRI buffer, 100–500 ng of DNA and molecular grade water (Sigma, UK) up to 10 μL. The reactions were incubated at 37°C for 1 hour and electrophoresed on 1% agarose gel. Recovery and characterization of PCR products containing intI and the first gene cassette Initially, we used previously published primers that had been used to successfully amplify gene cassettes from a range of environments (Fig 1, S1 and S2 Tables). Unexpectedly none of these primers produced amplicons having the structural features of a gene cassette [17] when oral metagenomic DNA isolated from the UK and Bangladesh was used as a template (see materials and methods). As Treponema denticola integrons are the only ones that have been described in the oral microbiota [27], new primers were designed based on this integron. The PCR were performed by using the intI-based primer TDIF (designed based on the conserved amino acid sequence SSQNQAL of IntI of the Treponema denticola integron) coupled with the attC-based primer MARS2. Resulting amplicons were cloned into pGEM-T Easy vector and a total of 17 clones were randomly selected from both cohorts and the inserts within the plasmids were sequenced. All of these contained the basic features of an integron. Within the amplicons, a major part of intI (768 bp), the full length attI site and a putative integron promoter, Pc were detected. A total of 5 different amplicons containing 5 different GCs including one empty GC with no identifiable ORF were found (Fig 2). The putative ORFs detected on the GCs had a size range of 258 to 777 bp (Table 1). Among the 17clones sequenced from both cohorts, 8 clones (TMB3/5/6/10/11/13/14/16) had a GC having an ORF (768-bp) predicted to encode a protein homologous to a cof-like hydrolase of Treponema putidum. Two of the first gene cassettes with an ORF of 258-bp and 387-bp present on clones TMB1/8/12/15 and TMU18, respectively had no nucleotide sequence similarity to anything in GenBank. However, at the amino acid level the 387-bp ORF on TMU18 showed 100% identity with a hypothetical protein of T. denticola. Another GC detected on clones TMU3/4/11 with an ORF of 777-bp was found to encode a hypothetical protein of Treponema denticola (Table 1). Finally, an empty first GC was found on clone TMB4.All but one ORF detected on the first GCs had putative ribosomal binding sites (RBS) at less than 8-bp upstream of the ORFs. In all first gene cassettes, two putative integrase binding sites (L and R; also termed as S2 and S1, respectively) were detected on the attI sites where the integrase binding sites S1 (R) were found to contain a plausible attI-attC junction (GTT). Sequence analyses DNA sequencing of inserts were performed at the Beckman Coulter Genomics (Beckman Coulter Genomics, UK) with an ABI 3730XL. M13 forward (5’ GTTTTCCCAGTCACGAC 3’) and M13 reverse (5’ GGAAACAGCTATGACCATG 3’) primers were used as the initial primers for sequencing. Additional primers were designed and used for further sequencing for the lon- ger inserts using Primer3 (http://biotools.umassmed.edu/bioapps/primer3_www.cgi). DNA sequences were aligned and manipulated by using BioEdit software version 7.2.0 (http://www.mbio.ncsu.edu/bioedit/bioedit.html). For the inserts which required sequencing with more than one primer, the sequences were assembled using the CAP contig function in the BioEdit program [28]. The sequences were screened for vector contamination by using VecScreen analysis tool (http://www.ncbi.nlm.nih.gov/tools/vecscreen). The primer binding sites were then identified by searching the sequences by eye. The sequences were analysed by the comparison of sequence and translated sequence using the National Centre for Biotechnol- ogy Information (NCBI) tools and databases including BlastN and BlastX [29], ORF finder and Clustal Omega. A sequence obtained using the attC-based primers was considered a putative GCs if (i) it contains both of the primer sequences (designed from conserved nucleotides of attC) (ii) the sites included an integrase-like simple site at each end [10] (iii) the primer sites flank a putative ORF beginning with ATG, TTG or GTG [17]. The sequences which did not contain an ORF, but contained the attC site, were considered as empty GCs. The putative translated sequences were subjected to BlastX searches and matches were considered significant if the e-value was <0.001. 5 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome Nomenclature and accession number of the gene cassettes The gene cassettes (GCs) were named according to the source and the primers. The first two letters indicate the forward and reverse primers used for amplification. The third letter indi- cates the source of oral metagenomic DNA that the GCs amplified from (U for UK; B for Ban- gladesh), which is followed by a numerical code for the number of clone. For example, TMB1 means it is the first GC obtained from Bangladeshi samples by using the primer TDIF and MARS2. The sequences of integron regions, which contained intI, Pc, attI and gene cassettes, were deposited in the DNA database Genbank under accession numbers from KT921469 to KT921473. The accession numbers from KT921474 to KT921509 and from KT921510 to KT921531 represented gene cassette sequences generated by the T. denticola primers from UK and Bangladeshi samples, respectively. PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Recovery and characterization of PCR products containing intI and the first gene cassette The 7 bp core site Rʹʹ (1L) of attC was also detected upstream of the reverse MARS2 primer having the consensus sequence RYY(/R)YAAC (S3 Table). In most cases, the stop codons of the ORFs was located at these Rʹʹ integrase binding sites of attC [8, 30]. 6 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome Table 1. Characterization of all gene cassettes detected in the saliva metagenomic DNA from UK and Bangladesh using TDIF and MARS2 primer combination to detect the first gene cassette. BlastN (ORF of ﬁrst gene cassette) BlastX (ORF of ﬁrst gene cassette) Gene cassettes /clone code Primer pair Size of insert including the integrase gene and ﬁrst cassette (bp) Orientation type Distance between integrase gene and ORF of ﬁrst gene cassette (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) TMB1/8/ 12/15 TDIF-MARS2 1499 A 271 KT921469 No signiﬁcant similarity - - No signiﬁcant similarity 258 - - Yes - TMB3/5/6/ 10/11/13/ 14/16 TDIF-MARS2 1757 A 198 KT921470 Treponema putidum 98 100 Cof-like hydrolase [Treponema denticola] 768 96 100 Yes WP_016512123 TMB4 TDIF-MARS2 1244 B - KT921471 Treponema pedis 82 46 - - - - - - TMU3/4/11 TDIF-MARS2 1775 A 204 KT921472 Treponema sp. OMZ 838 77 97 Hypothetical protein [Treponema denticola] 777 99 100 Yes WP_002678613.1 TMU18 TDIF-MARS2 1692 A 161 KT921473 No signiﬁcant similarity - - Hypothetical protein [Treponema denticola] 387 100 100 Yes WP_002689012.1 doi:10.1371/journal.pone.0157605.t001 Table 1. Characterization of all gene cassettes detected in the saliva metagenomic DNA from UK and Bangladesh using TDIF and MARS2 primer combination to detect the first gene cassette. BlastN (ORF of ﬁrst gene cassette) BlastX (ORF of ﬁrst gene cassette) Gene cassettes /clone code Primer pair Size of insert including the integrase gene and ﬁrst cassette (bp) Orientation type Distance between integrase gene and ORF of ﬁrst gene cassette (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) TMB1/8/ 12/15 TDIF-MARS2 1499 A 271 KT921469 No signiﬁcant similarity - - No signiﬁcant similarity 258 - - Yes - TMB3/5/6/ 10/11/13/ 14/16 TDIF-MARS2 1757 A 198 KT921470 Treponema putidum 98 100 Cof-like hydrolase [Treponema denticola] 768 96 100 Yes WP_016512123 TMB4 TDIF-MARS2 1244 B - KT921471 Treponema pedis 82 46 - - - - - - TMU3/4/11 TDIF-MARS2 1775 A 204 KT921472 Treponema sp. Integrons from the Human Saliva Metagenome Fig 2 Orientation of ORFs in the GCs recovered from metagenomic DNA of saliva samples Fig 2. Orientation of ORFs in the GCs recovered from metagenomic DNA of saliva samples. Fig 2. Orientation of ORFs in the GCs recovered from metagenomic DNA of saliva samples. PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 7 / 20 Gene Cassettes Amplified Using attC-based primers A library of PCR amplicons obtained using a different set of attC-based primers was con- structed in pGEM-T Easy vector and the inserts from 100 clones were sequenced (Table 2, Fig 1, S1 and S2 Tables). By analysing the sequences with different bioinformatics tools we have detected a total of 58 unique GCs having the features of an integron GC and flanked by the primer binding sites. The size of the cassettes ranged from 425 to 1144 bp. Of the 58 GCs, 12 had no identifiable ORFs and the remaining 46 GCs contained one or more putative ORFs giv- ing a total of 72 different ORFs with a size range between 117 to 894 bp. As the forward and reverse primers were designed based on the consensus Lʹ (2R) and Lʹʹ (2L) core sites, respec- tively, we were able to locate the Rʹ (1R) core sites in all GCs with a consensus GTTRR(Y)R(Y) Y(R) after the forward primer sequence. The complementary Rʹʹ (1L) core sites with a consen- sus R(Y)Y(R)Y(R)YAAC were also detected upstream or as a part of the reverse attC primers which confirms that the putative GCs are not PCR artefacts and is consistent with the attC structure of a GC [31]. The majority of the Rʹ and Rʹʹ core sites (51 out of 58 GCs) exhibited 100% complementarity with each other. In the remaining seven, 6 out of 7-bp were compli- mentary (S4 Table). By analysing the arrangement of genetic features within the GCs we found that they were arranged in seven different ways (Fig 2) as defined by the direction, position and number of ORFs within the GCs. The type C arrangement accounted for the majority; found in 24 cas- settes. The sequences of the clones containing two or more ORFs were examined for the pres- ence of other putative attC sequences in between the ORFs, none of which were found. These observations show that the attC-based primers based on the T. denticola integron are able to amplify GCs from oral metagenomic DNA. From 72 putative ORFs found in all GCs, 63 of them had ribosomal binding sites located upstream of the predicted start codons. As in previ- ous studies the GCs other than the toxin-antitoxin encoding GCs did not contain an identifi- able promoter, thus are likely to be dependent on the Pc of the cassette array for expression [19]. OMZ 838 77 97 Hypothetical protein [Treponema denticola] 777 99 100 Yes WP_002678613.1 TMU18 TDIF-MARS2 1692 A 161 KT921473 No signiﬁcant similarity - - Hypothetical protein [Treponema denticola] 387 100 100 Yes WP_002689012.1 d i 10 1371/j l 0157605 t001 A from UK and Bangladesh using TDIF and MARS2 primer combination to detect the Table 1. Characterization of all gene cassettes detected in the saliva metagenomic DNA from UK and Bangladesh using TDIF and MARS2 primer combination to detect the first gene cassette. BlastN (ORF of ﬁrst gene cassette) BlastX (ORF of ﬁrst gene cassette) Gene cassettes /clone code Primer pair Size of insert including the integrase gene and Orientation type Distance between integrase gene and ORF of ﬁrst gene Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) Table 1. Characterization of all gene cassettes detected in the saliva metagenomic DN first gene cassette PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 8 / 20 Integrons from the Human Saliva Metagenome PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers Out of 72 putative ORFs detected on 58 different GCs amplified by using attC primers, 66 (91.66%) of the predicted proteins had a homologue in GenBank. However, only 24 of the 66 ORFs (36.36%) were found to encode proteins with known function and the remaining 42 matched hypothetical proteins. With regards to sequence similarity of the ORFs with those in GenBank, we found that 45 of the 66 ORFs (68.0%) exhibited >90% amino acid identity. Ten putative ORFs were predicted to encode completely novel proteins (e-value <0.001). The putative ORFs detected on the gene cassettes were predicted to encode proteins of diverse functions including antibiotic resistance, host adaptation to stress and competence (Table 2). Four different putative antibiotic resistance genes were found among the cassette ORFs. BlastX searches showed that the clone MMB22 contained an ORF that encoded a pro- tein with 99% identity to streptogramin A O-acetyltransferase from T. denticola. The single ORF (390-bp) present in the clones SSU3, SSU4 and SSU30 of UK was predicted to encode a glyoxalase/bleomycin antibiotic binding protein. Two ORFs were detected in the clone SSU28 encoding potassium ABC transporter ATPase and multidrug transporter MatE. Proteins related to adaptation to stress include different toxin-antitoxin systems and a twitching motility protein. The clones containing the ORFs encoding toxin-antitoxin system includes SSU27, MMB23, MMB38 which encoded HicA (toxin)- HicB (antitoxin), peptidase (antitoxin)-PemK (toxin) and higA (antitoxin)-higB (toxin), respectively. 9 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome Table 2. Characterization of all gene cassettes detected in the saliva metagenomic DNA from UK and Bangladesh using attC-based primers. BlastN BlastX Gene cassettes/ clone code Primer pair Cassette Size (bp) Orientation type Distance between attC and ORF (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) SSU1/2, MMU15 SUPA3-SUPA4, MARS5-MARS2 799 C 36 KT921474 No signiﬁcant similarity found. - - Hypothetical protein [Rhodonellum psychrophilum] 579 66 75 Yes WP_026333632.1 SSU3/4/30 SUPA3-SUPA4 964 E 41 KT921475 No signiﬁcant similarity found. - - Glyoxalase [Treponema pedis] 390 97 100 Yes WP_009105863.1 Competence protein TfoX [Treponema pedis] 315 83 100 Yes WP_024470244.1 SSU5 SUPA3-SUPA4 832 E 136 KT921476 Treponema sp. Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers 88 27 Hypothetical protein [Treponema putidum] 243 67 100 Yes WP_044978234.1 Twitching motility protein PilT [Treponema putidum] 396 71 100 Yes AIN93467.1 SSU6 SUPA3-SUPA4 1138 E 34 KT921477 No signiﬁcant similarity found. - - No signiﬁcant similarity found. 459 - - Yes - Hypothetical protein [Treponema denticola] 339 75 100 Yes WP_044013590.1 SSU7 SUPA3-SUPA4 491 F 0 KT921478 Treponema denticola 92 40 No signiﬁcant similarity found 162 - - No - No signiﬁcant similarity found 132 - - Yes - SSU8 SUPA3-SUPA4 921 C 41 KT921479 Treponema sp. 98 49 Hypothetical protein [Treponema denticola] 567 98 93.7 Yes WP_002692239.1 SSU9/13/ 14/19/20/ 23 SUPA3-SUPA4 425 H - KT921480 Treponema pedis 78 69 No signiﬁcant similarity found - - - - - SSU10 SUPA3-SUPA4 612 C 29 KT921481 No signiﬁcant similarity found. - - Hypothetical protein [Paenibacillus assamensis] 537 39 91 Yes WP_028595336 SSU11 SUPA3-SUPA4 648 C 4 KT921482 No signiﬁcant similarity found. - - Hypothetical protein [Bradyrhizobium sp. STM 3809] 597 42 89.7 Yes WP_035659994.1 SSU12 SUPA3-SUPA4 736 C 77 KT921483 No signiﬁcant similarity found. - - Hypothetical protein [Treponema vincentii] 582 99 100 Yes WP_016518887.1 (Continued) PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 10 / 2 the saliva metagenomic DNA from UK and Bangladesh using attC-based primers. Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 10 / 20 Integrons from the Human Saliva Metagenome BlastN BlastX Gene cassettes/ clone code Primer pair Cassette Size (bp) Orientation type Distance between attC and ORF (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) SSU15 SUPA3-SUPA4 989 G 105 KT921484 Treponema denticola 90 19 Hypothetical protein [Treponema sp.] 396 40 44.3 Yes WP_044015417.1 Hypothetical protein [Treponema denticola] 228 39 16.4 No WP_010689034.1 Hypothetical protein TPE_0657 [Treponema pedis] 147 96 100 Yes AGT43153.1 SSU16 SUPA3-SUPA4 711 E 127 KT921485 Treponema pedis 90 21 Hypothetical protein [Treponema sp.] 363 33 41.6 Yes WP_044015417.1 Hypothetical protein TPE_0657 [Treponema pedis] 147 96 100 Yes AGT43153.1 SSU17 SUPA3-SUPA4 787 C 30 KT921486 Treponema denticola 98 100 Hypothetical protein [Treponema denticola] 711 97 100 Yes WP_002690335.1 SSU18 SUPA3-SUPA4 853 C 31 KT921487 No signiﬁcant similarity found - - Hypothetical protein (Endonuclease) [Treponema putidum] 765 33 99.2 Yes WP_044978378.1 SSU21 SUPA3-SUPA4 871 D 273 KT921488 Treponema denticola 97 100 Toxin [Treponema denticola] 291 96 100 No WP_010692226.1 Hypothetical protein (antitoxin) [Treponema denticola] 258 100 100 Yes WP_010692225.1 SSU22 SUPA3-SUPA4 1017 E 72 KT921489 No signiﬁcant similarity found - - Prevent-host-death family protein [Treponema medium] 264 94 100 Yes WP_016523165.1 XRE family transcriptional regulator [Treponema vincentii] 441 96 100 Yes WP_006188306.1 SSU24 SUPA3-SUPA4 962 E 23 KT921490 No signiﬁcant similarity found - - Glyoxalase [Treponema pedis] 411 97 100 No WP_024470245.1 Competence protein TfoX [Treponema pedis] 222 88 55.2 No WP_024470244.1 SSU25 SUPA5-SUPA6 789 C 31 KT921491 Treponema putidum 96 99 Hypothetical protein [Treponema denticola] 705 97 100 Yes WP_010697531.1 (C ti d) PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 11 / Table 2. Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers (Continued) PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 11 / 20 Integrons from the Human Saliva Metagenome ( ) BlastN BlastX Gene cassettes/ clone code Primer pair Cassette Size (bp) Orientation type Distance between attC and ORF (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) SSU26 SUPA5-SUPA6 848 C 34 KT921492 Treponema putidum 81 93 Cof-like hydrolase [Treponema denticola] 768 76 100 Yes WP_010693073.1 SSU27 SUPA5-SUPA6 833 G 155 KT921493 Treponema denticola 96 100 No signiﬁcant similarity found 126 - - - - Antitoxin HicB [Treponema denticola] 405 99 100 No WP_002669522.1 Toxin HicA [Treponema denticola] 195 97 100 Yes WP_002669524.1 SSU28 SUPA5-SUPA6 927 E 310 KT921494 Treponema putidum 97 91 Multidrug transporter MatE [Treponema putidum] 336 96 100 Yes WP_044979179.1 mRNA-degrading endonuclease [Treponema denticola] 231 99 100 Yes WP_010694033.1 SSU29 SUPA5-SUPA6 425 H - KT921495 Treponema pedis 78 69 - - - - - - MMU2 MARS5-MARS2 519 H - KT921496 Treponema denticola 81 72 - - - - - - MMU3 MARS5-MARS2 826 C 63 KT921497 No signiﬁcant similarity found - - Hypothetical protein [Morganella morganii] 753 32 97.13 Yes WP_036423270.1 MMU7 MARS5-MARS2 765 C 148 KT921498 Treponema denticola 96 98 Hypothetical protein [Treponema sp.] 390 100 100 Yes WP_002688988.1 MMU9/13 MARS5-MARS2 725 C 127 KT921499 Treponema denticola 90 21 Hypothetical protein [Treponema phagedenis] 474 53 98.8 Yes WP_044634793.1 MMU11 MARS5-MARS2 734 H - KT921500 Treponema pedis 79 51 - - - - - - MMU18 MARS5-MARS2 690 C 174 KT921501 No signiﬁcant similarity found - - Hypothetical protein [Treponema denticola] 459 97 100 Yes EGC77532.1 MMU19 MARS5-MARS2 431 H - KT921502 Treponema pedis 78 63 - - - - - - MMU20 MARS5-MARS2 943 C 72 KT921503 Treponema putidum 84 16 No signiﬁcant similarity found 780 - - No - MMU24 MARS5-MARS2 1041 D 435 KT921505 Treponema sp. OMZ 838 97 62 Plasmid maintenance system killer protein [Treponema sp. Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers OMZ 838] 279 97 100 Yes WP_044013090.1 (Continued) PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 12 / PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 12 / 20 Integrons from the Human Saliva Metagenome BlastN BlastX Gene cassettes/ clone code Primer pair Cassette Size (bp) Orientation type Distance between attC and ORF (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) Hig A family addiction module antidote protein [Treponema socranskii] 315 97 100 Yes WP_016519833.1 MMU25 MARS5-MARS2 1144 G 61 KT921506 Treponema sp. OMZ 838, 79 43 No signiﬁcant similarity found 462 - - Yes - Hypothetical protein [Treponema maltophilum] 213 91 91 Yes WP_016526060.1 Hypothetical protein [Treponema vincentii] 351 77 100 Yes WP_044013590.1 MMU26 MARS5-MARS2 972 C 27 KT921507 Treponema denticola 95 100 Carbon-nitrogen hydrolase [Treponema denticola] 891 97 100 Yes WP_044901990.1 MMU27 MARS5-MARS2 896 E 38 KT921508 No signiﬁcant similarity found - - Hypothetical protein [Treponema denticola] 402 95 88 Yes WP_002687406.1 Hypothetical protein [Treponema denticola] 360 99 100 Yes WP_002673043.1 MMU28 MARS5-MARS2 987 E 320 KT921509 Treponema putidum 89 76 Prevent-host-death protein [Treponema putidum] 282 91 100 Yes WP_044979047.1 Hypothetical protein (Plasmid stabilization protein) [Treponema putidum] 213 85 99.1 Yes AIN94307.1 MMU23 MARS5-MARS2 634 H - KT921504 Treponema pedis 84 60 - - - - - - MMB1/2 MARS5-MARS2 633 H - KT921510 Treponema pedis 80 56 - - - - - - MMB3/9 MARS5-MARS2 846 C 369 KT921511 Treponema sp. 27 86 Twitching motility protein PilT [Treponema putidum] 414 70 89 No AIN93467.1 MMB5/35 MARS5-MARS2 423 H - KT921513 Treponema denticola 84 94 - - - - - - MMB4/12/ 31 MARS5-MARS2 877 D 271 KT921512 Treponema denticola 96 100 Toxin [Treponema denticola] 288 95 100 No WP_01069226.1 (Continued) g g LOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 13 / Table 2. Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers (Continued) PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 13 / 20 Integrons from the Human Saliva Metagenome BlastN BlastX Gene cassettes/ clone code Primer pair Cassette Size (bp) Orientation type Distance between attC and ORF (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) Hypothetical protein [Treponema pedis] 258 99 100 Yes WP_024468276.1 MMB11/ 13/21/25/ 29 MARS5-MARS2 629 C 98 KT921514 Treponema denticola 98 81 Membrane protein [Treponema denticola] 312 94 100 Yes WP_002669840.1 MMB15 MARS5-MARS2 889 C 31 KT921516 No signiﬁcant similarity found - - Hypothetical protein [Mariprofundus ferrooxydans] 804 46 89 Yes WP_009850619 MMB22 MARS5-MARS2 854 C 160 KT921521 Clostridium sp. Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers 74 43 Vat family streptogramin A O- acetyltransferase [Treponema denticola] 645 99 100 Yes WP_00267814.1 MMB14 MARS5-MARS2 862 E 86 KT921515 Treponema putidum 66 97 RelB/DinJ antitoxin [Treponema putidum] 417 97 95 Yes WP_044978223 Hypothetical protein [Treponema vincentii] 300 86 100 Yes WP_016519648.1 MMB17 MARS5-MARS2 838 E 49 KT921517 Treponema denticola 99 100 Hypothetical protein [Treponema denticola] 396 98 100 Yes WP_010957066.1 Hypothetical protein [Treponema putidum] 117 100 100 Yes WP_010957065.1 MMB18 MARS5-MARS2 792 C 26 KT921518 Treponema putidum 98 100 Hypothetical protein [Treponema denticola] 708 97 100 Yes WP_010697531.1 MMB19 MARS5-MARS2 817 C 45 KT921519 Treponema denticola 94 100 Hypothetical protein [Treponema denticola] 708 96 100 Yes WP_010697511.1 MMB20 MARS5-MARS2 531 H - KT921520 Treponema pedis 80 66 - - - - - - MMB23/27 MARS5-MARS2 901 D 283 KT921522 No signiﬁcant similarity found - - Peptidase (Antitoxin) [Treponema denticola] 243 63 100 Yes WP_002667901.1 PemK toxin [Treponema denticola] 330 77 100 Yes WP_002680107.1 MMB28 MARS5-MARS2 543 C 122 KT921523 No signiﬁcant similarity found - - Hypothetical protein [Clostridium drakei] 360 62 98 Yes WP_029160182.1 LOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 14 / PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 14 / 20 Integrons from the Human Saliva Metagenome BlastN BlastX Gene cassettes/ clone code Primer pair Cassette Size (bp) Orientation type Distance between attC and ORF (bp) Accession number Closest homologue Percentage identity (%) Coverage (%) Closest homologue ORF size (bp) Percentage identity (%) Coverage (%) The presence of ribosomal binding site Accession number of the homologous proteins (BlastX) MMB30 MARS5-MARS2 808 C 25 KT921524 No signiﬁcant similarity found - - Hypothetical protein [Treponema medium] 729 92 100 Yes WP_016670455 MMB32 MARS5-MARS2 632 H - KT921525 Treponema pedis 79 56 - - - - - - MMB33 MARS5-MARS2 810 E 80 KT921526 Treponema denticola 99 100 Hypothetical protein [Treponema denticola] 372 98 100 No WP_002689013.1 Hypothetical protein [Treponema sp.] 276 99 100 Yes WP_002689021.1 MMB34 MARS5-MARS2 714 C 120 KT921527 No signiﬁcant similarity found - - Hypothetical protein [Treponema medium] 525 94 99 Yes WP_016522334.1 MMB36 MARS5-MARS2 732 H - KT921528 Treponema pedis 79 48 - - - - - - MMB37 MARS5-MARS2 635 H - KT921529 Treponema denticola 82 60 - - - - - - MMB38 MARS5-MARS2 947 G 3 KT921530 Treponema sp. Discussion The PCR strategies to recover novel integron cassettes from metagenomic DNA using primers targeting the conserved sequence of IntI and attC have been successful in previous studies [15– 17, 19, 33]. However, all of these metagenomic studies were carried out on non-human envi- ronmental samples. Most of the metagenomic studies involving human microbiota, were either sequence-based [34] focusing on the recovery of all genetic features or focusing on a function of interest such as antibiotic resistance [35]. No studies have been reported so far on metagen- omes obtained from human saliva to detect integrons using a PCR approach. We detected mostly Treponema integrons and GCs from metagenomic DNA from human saliva from both Bangladeshi and UK samples, indicating that this methodology is applicable to any oral meta- genomic sample. This study provides an analysis of the diversity of integron GCs amplifiable in saliva meta- genomic DNA. Using novel primer combinations based on the structural features of the reverse integron of T. denticola ATCC 35405 [27], we have uncovered a diverse array of gene cassettes including those in the first position, most of which are novel. Although the chromosomal integron of T. denticola ATCC 35405 is the only integron described from the oral bacteria (it has 45 gene cassettes in the array), in silico analysis of metagenomic data sets from the Human Microbiome Project (HMP) showed that two other Treponema species, including T. vincentii ATCC 35580 and T. phagedenis F0421, have also been found to carry integron GCs [14, 27, 32]. However, the PCR strategies used in this study, recovered novel GCs that were predicted to encode proteins related to those from genera other than Trepo- nema spp. Analyzing the proteins encoded by the GCs amplified from the oral cavity showed several interesting ORFs. GC SSU3 was predicted to encode a protein with 97% amino acid identity to the glyoxalase of Treponema pedis (WP_009105863.1, 100% coverage). It contains the Glo- EDI-BRP-like domain which can be found in metalloproteins including glyoxalase I, type I extradiol dioxygenases and bleomycin sequester proteins. Bleomycin is a glycopeptide antibi- otic, which inhibits the peptidoglycan synthesis in bacteria, and also used as an antitumor drug which bind to DNA and generate free radicals that result in both double-strand and single- strand DNA breaks [36, 37]. Diversity of the functions of putative proteins encoded by ORFs within the GCs detected by attC primers 97 67 No signiﬁcant similarity found 201 - - No - HigA antitoxin [Treponema socranskii] 282 98 100 Yes WP_016519833.1 Plasmid maintenance system killer protein [Treponema medium] 318 97 100 Yes WP_040858928.1 MMB39 MARS5-MARS2 857 E 97 KT921531 No signiﬁcant similarity found - - Hypothetical protein [Treponema medium] 357 96 100 Yes WP_016522532.1 Hypothetical protein [Treponema medium] 330 95 100 Yes WP_016522533.1 doi:10.1371/journal.pone.0157605.t002 Table 2. (Continued) PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 15 / 20 Integrons from the Human Saliva Metagenome Most of the proteins encoded by the ORFs on GCs showed similarity with many proteins in the database, some of which were from Treponema spp. (60 out of 66) mostly from T. denticola (24 out of 60) followed by T. putidum, T. medium, T. vincentii, T. pedis, T. phagedenis, and T. socranskii. This observation supports the previous reports that T. denticola, T vincentii and T. phagedenis carry chromosomal integrons [14, 32]. However, we have also identified 27 ORFs related to other Treponema spp.; T. putidum, T. medium, T. pedis and T. socranskii. Only six ORFs out of 66 were predicted to encode proteins related to non-treponemes including those from Paenibacillus sp., Clostridium sp. and Maripofundus sp. however, the homologies of the ORFs with these species were low (<70%) at the amino acid level. Discussion Another ORF found on GC MMB22 detected in the Bangladeshi sample was predicted to encode streptogramin A O-acetyltransferase which had 77.0% nucleo- tide identity with Clostridium sp. BLN1100 and 99.0% amino acid identity with the streptogra- min A O-acetyltransferase from T. denticola. Streptogramin A O-acetyltransferases mediate resistance to the streptogramin A-B combination by adding acetyl group to streptogramin, which inactivates the drugs [38]. Finally, a cof-like hydrolase gene (a member of haloacid dehalogenase superfamily) was pre- dicted to be within a GC amplified using both GC primers and first gene cassette primers (GC SSU26 and GC TMB3). Cof-like hydrolases are a group of enzymes that inactivate 16 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome halogenated aliphatic hydrocarbons by hydrolysing the carbon-halogen bonds. They are essen- tial for detoxification of many chlorinated compounds [39, 40]. Therefore, a cof-like hydrolase in the oral cavity could play a role in detoxifying or inactivating antimicrobials or other com- pounds with carbon-halogen bonds that are used as antibiotics, pesticides and food preserva- tives such as chloramphenicol, atrazine and brominated vegetable oil, respectively. Another function of predicted GC ORFs was related to the adaptation of bacteria to envi- ronmental stress. For example, the twitching motility PilT protein was predicted to be encoded by the ORF in GC of clone SSU5, MMB3 and MMB9. It has been shown to be involved with type IV fimbria-mediated twitching motility and protease secretion [41]. Twitching motility was also shown to play a key role in the development of biofilm from Pseudomonas aeruginosa [42]. As many oral bacteria can form biofilms on the surfaces in the human oral cavity, having a PilT-encoded GC could help them to develop biofilms and survive environ- mental stress. As in previous metagenomic studies to detect integron GCs [16, 17, 43], ORFs predicted to encode proteins with regulatory functions such as toxin-antitoxin (TA) systems have been detected. Four different TA operons including the HicAB, HigBA, RelBE and MazF were detected on GCs in our study. TA cassettes are usually abundant in chromosomal integrons and are thought to have a role in the stability of the integron GC arrays [27, 44]. All of the detected TA cassettes are the members of type II toxin-antitoxin systems [45]. Discussion The toxins (HicA, HigA, RelE and MazF) work by cleaving mRNA, inhibiting translation and exhibit bac- teriostatic activity, and the antitoxins (HicB, HIgB, RelE, MazE) can inhibit the action of toxin by protein-protein complex formation [46–49]. Among the four detected TA operons, only the HicAB TA system was previously found on the T. denticola integron. The nucleotide sequence of HicA and HicB system found on SSU27 cassette exhibited 97% and 99% nucleotide identity to the corresponding fourth gene cassette of the integron of T. denticola, containing HicA (TDE1838) and HicB (TDE1837) genes[27]. We have detected two HigBA TA systems in our GCs (MMU24 and MMB38), and this system has also been detected on the Vibrio cholerae super integron Several recovered GCs did not contain ORFs. This kind of ORF-less GCs was found both in the first position GC and other GC positions in the integron (clone TMB4, SSU29 and MMU2). Other noncoding cassettes have been previously found in cassette arrays comprising, for example, between 4 and 49% of Vibrio spp. cassette arrays [50]. They have been hypothesised to contain promoters or encode regulatory RNAs [2]. It was previously shown that a Xanthomonas campestris integron GC encoded trans-acting small RNA, which was capable of regulating the virulence in Xanthomonas [51]. This survey on the presence of integrons and associated GCs in salivary metagenomic DNA has resulted in new information regarding the putative functions and diversity of GCs which likely reflects the highly variable physicochemical and stressful environment of the human oral cavity. Supporting Information Supporting Information S1 Table. Primers used in this study. (DOCX) S2 Table. The lists of primer pairs used in this study. (DOCX) S3 Table. Complementary integrase binding site S1 (R) and R” seqauence on the first gene cassettes. (DOCX) References 1. Cambray G, Guerout AM, Mazel D. Integrons. Annual review of genetics. 2010; 44:141–66. Epub 2010/ 08/17. doi: 10.1146/annurev-genet-102209-163504 PMID: 20707672. 1. Cambray G, Guerout AM, Mazel D. Integrons. Annual review of genetics. 2010; 44:141–66. Epub 2010/ 08/17. doi: 10.1146/annurev-genet-102209-163504 PMID: 20707672. 1. Cambray G, Guerout AM, Mazel D. Integrons. Annual review of genetics. 2010; 44:141–66. Epub 2010/ 08/17. doi: 10.1146/annurev-genet-102209-163504 PMID: 20707672. 2. Gillings MR. Integrons: past, present, and future. Microbiology and molecular biology reviews: MMBR. 2014; 78(2):257–77. Epub 2014/05/23. doi: 10.1128/mmbr.00056-13 PMID: 24847022; PubMed Cen- tral PMCID: PMCPMC4054258. 2. Gillings MR. Integrons: past, present, and future. Microbiology and molecular biology reviews: MMBR. 2014; 78(2):257–77. Epub 2014/05/23. doi: 10.1128/mmbr.00056-13 PMID: 24847022; PubMed Cen- tral PMCID: PMCPMC4054258. 3. Michael CA, Gillings MR, Holmes AJ, Hughes L, Andrew NR, Holley MP, et al. Mobile gene cassettes: a fundamental resource for bacterial evolution. The American naturalist. 2004; 164(1):1–12. Epub 2004/07/22. doi: 10.1086/421733 PMID: 15266366. 4. Guerin E, Cambray G, Sanchez-Alberola N, Campoy S, Erill I, Da Re S, et al. The SOS response con- trols integron recombination. Science (New York, NY). 2009; 324(5930):1034. Epub 2009/05/23. doi: 10.1126/science.1172914 PMID: 19460999. 5. Collis CM, Hall RM. Gene cassettes from the insert region of integrons are excised as covalently closed circles. Molecular microbiology. 1992; 6(19):2875–85. Epub 1992/10/01. PMID: 1331702. 6. Partridge SR, Recchia GD, Scaramuzzi C, Collis CM, Stokes HW, Hall RM. Definition of the attI1 site of class 1 integrons. Microbiology (Reading, England). 2000; 146 (Pt 11):2855–64. Epub 2000/11/07. PMID: 11065364. 7. Collis CM, Recchia GD, Kim M-J, Stokes HW, Hall RM. Efficiency of Recombination Reactions Cata- lyzed by Class 1 Integron Integrase IntI1. Journal of bacteriology. 2001; 183(8):2535–42. doi: 10.1128/ JB.183.8.2535-2542.2001 8. Stokes HW, O'Gorman DB, Recchia GD, Parsekhian M, Hall RM. Structure and function of 59-base ele- ment recombination sites associated with mobile gene cassettes. Molecular microbiology. 1997; 26 (4):731–45. doi: 10.1046/j.1365-2958.1997.6091980.x PMID: 9427403 9. Escudero JA, Loot C, Nivina A, Mazel D. The Integron: Adaptation On Demand. Microbiology spectrum. 2015; 3(2). doi: 10.1128/microbiolspec.MDNA3-0019-2014 10. Collis CM, Hall RM. Expression of antibiotic resistance genes in the integrated cassettes of integrons. Antimicrobial agents and chemotherapy. 1995; 39(1):155–62. Epub 1995/01/01. PMID: 7695299; PubMed Central PMCID: PMCPMC162502. 11. Hall RM, Collis CM. Antibiotic resistance in gram-negative bacteria: the role of gene cassettes and inte- grons. Drug resistance updates: reviews and commentaries in antimicrobial and anticancer chemother- apy. 1998; 1(2):109–19. Epub 2006/08/15. PMID: 16904397. 12. S1 Table. Primers used in this study. (DOCX) S2 Table. The lists of primer pairs used in this study. (DOCX) S2 Table. The lists of primer pairs used in this study. (DOCX) S3 Table. Complementary integrase binding site S1 (R) and R” seqauence on the first gene cassettes. (DOCX) S3 Table. Complementary integrase binding site S1 (R) and R” seqauence on the first gene cassettes. ( ) 17 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome S4 Table. Complementary of the cores sites Rʹ (1R) and Rʹʹ (1L) abutting the forward and reverse attC primer sequence on the gene cassette. (DOCX) S4 Table. Complementary of the cores sites Rʹ (1R) and Rʹʹ (1L) abutting the forward and reverse attC primer sequence on the gene cassette. (DOCX) Author Contributions Conceived and designed the experiments: ST MAR APR. Performed the experiments: ST MAR AA. Analyzed the data: ST MAR AA PM APR. Contributed reagents/materials/analysis tools: ST MAR APR. Wrote the paper: ST MAR AA PM APR. Conceived and designed the experiments: ST MAR APR. Performed the experiments: ST MAR AA. Analyzed the data: ST MAR AA PM APR. Contributed reagents/materials/analysis tools: Conceived and designed the experiments: ST MAR APR. Performed the experiments: ST MAR AA. Analyzed the data: ST MAR AA PM APR. Contributed reagents/materials/analysis tools: ST MAR APR. Wrote the paper: ST MAR AA PM APR. Acknowledgments We are grateful to Professor Md. Anwar Ul Islam of University of Rajshahi, Bangladesh for helping us to collect and send the saliva sample from Bangladeshi volunteers. Md. Ajijur Rah- man is supported by the Commonwealth Scholarship Commission in the UK (Ref: BDCA 2013–4). We thank Prof Pål Johnsen of UiT The Arctic University of Norway for helpful com- ments and suggestions on the initial version of the manuscript. PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 References Stokes HW, Hall RM. A novel family of potentially mobile DNA elements encoding site-specific gene- integration functions: integrons. Molecular microbiology. 1989; 3(12):1669–83. Epub 1989/12/01. PMID: 2560119. 13. Partridge SR, Tsafnat G, Coiera E, Iredell JR. Gene cassettes and cassette arrays in mobile resistance integrons. FEMS microbiology reviews. 2009; 33(4):757–84. Epub 2009/05/07. doi: 10.1111/j.1574- 6976.2009.00175.x PMID: 19416365. 18 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 Integrons from the Human Saliva Metagenome 14. Wu YW, Rho M, Doak TG, Ye Y. Oral spirochetes implicated in dental diseases are widespread in nor- mal human subjects and carry extremely diverse integron gene cassettes. Applied and environmental microbiology. 2012; 78(15):5288–96. Epub 2012/05/29. doi: 10.1128/aem.00564-12 PMID: 22635997; PubMed Central PMCID: PMCPMC3416431. 14. Wu YW, Rho M, Doak TG, Ye Y. Oral spirochetes implicated in dental diseases are widespread in nor- mal human subjects and carry extremely diverse integron gene cassettes. Applied and environmental microbiology. 2012; 78(15):5288–96. Epub 2012/05/29. doi: 10.1128/aem.00564-12 PMID: 22635997; PubMed Central PMCID: PMCPMC3416431. 15. Holmes AJ, Gillings MR, Nield BS, Mabbutt BC, Nevalainen KM, Stokes HW. The gene cassette meta- genome is a basic resource for bacterial genome evolution. Environmental microbiology. 2003; 5 (5):383–94. Epub 2003/04/26. PMID: 12713464. 16. Elsaied H, Stokes HW, Kitamura K, Kurusu Y, Kamagata Y, Maruyama A. Marine integrons containing novel integrase genes, attachment sites, attI, and associated gene cassettes in polluted sediments from Suez and Tokyo Bays. The ISME journal. 2011; 5(7):1162–77. Epub 2011/01/21. doi: 10.1038/ ismej.2010.208 PMID: 21248857; PubMed Central PMCID: PMCPMC3146285. 17. Elsaied H, Stokes HW, Nakamura T, Kitamura K, Fuse H, Maruyama A. Novel and diverse integron integrase genes and integron-like gene cassettes are prevalent in deep-sea hydrothermal vents. Envi- ronmental microbiology. 2007; 9(9):2298–312. Epub 2007/08/10. doi: 10.1111/j.1462-2920.2007. 01344.x PMID: 17686026. 18. Amann RI, Ludwig W, Schleifer KH. Phylogenetic identification and in situ detection of individual micro- bial cells without cultivation. Microbiological reviews. 1995; 59(1):143–69. Epub 1995/03/01. PMID: 7535888; PubMed Central PMCID: PMCPMC239358. 19. Stokes HW, Holmes AJ, Nield BS, Holley MP, Nevalainen KM, Mabbutt BC, et al. Gene cassette PCR: sequence-independent recovery of entire genes from environmental DNA. Applied and environmental microbiology. 2001; 67(11):5240–6. Epub 2001/10/27. doi: 10.1128/aem.67.11.5240-5246.2001 PMID: 11679351; PubMed Central PMCID: PMCPMC93296. 20. Koenig JE, Sharp C, Dlutek M, Curtis B, Joss M, Boucher Y, et al. Integron Gene Cassettes and Degra- dation of Compounds Associated with Industrial Waste: The Case of the Sydney Tar Ponds. PloS one. 2009; 4(4):e5276. PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 References doi: 10.1186/1471-2148-13-16 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 19 / 20 Integrons from the Human Saliva Metagenome 33. Nield BS, Holmes AJ, Gillings MR, Recchia GD, Mabbutt BC, Nevalainen KM, et al. Recovery of new integron classes from environmental DNA. FEMS Microbiol Lett. 2001; 195(1):59–65. Epub 2001/02/ 13. PMID: 11166996. 34. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Manichanh C, et al. A human gut microbial gene cata- logue established by metagenomic sequencing. Nature. 2010; 464(7285):59–65. doi: 10.1038/ nature08821 PMID: 20203603; PubMed Central PMCID: PMC3779803. 35. Sommer MO, Dantas G, Church GM. Functional characterization of the antibiotic resistance reservoir in the human microflora. Science. 2009; 325(5944):1128–31. doi: 10.1126/science.1176950 PMID: 19713526. 36. Chen J, Stubbe J. Bleomycins: towards better therapeutics. Nature reviews Cancer. 2005; 5(2):102– 12. Epub 2005/02/03. doi: 10.1038/nrc1547 PMID: 15685195. 37. Kahne D, Leimkuhler C, Lu W, Walsh C. Glycopeptide and lipoglycopeptide antibiotics. Chemical reviews. 2005; 105(2):425–48. Epub 2005/02/11. doi: 10.1021/cr030103a PMID: 15700951. 38. De Meester C, Rondelet J. Microbial acetylation of M factor of virginiamycin. The Journal of antibiotics. 1976; 29(12):1297–305. Epub 1976/12/01. PMID: 13056. 39. Koonin EV, Tatusov RL. Computer analysis of bacterial haloacid dehalogenases defines a large super- family of hydrolases with diverse specificity. Application of an iterative approach to database search. J Mol Biol. 1994; 244(1):125–32. doi: 10.1006/jmbi.1994.1711 PMID: 7966317. 40. Hardman DJ. Biotransformation of halogenated compounds. Crit Rev Biotechnol. 1991; 11(1):1–40. doi: 10.3109/07388559109069182 PMID: 1863980. 41. Han X, Kennan RM, Davies JK, Reddacliff LA, Dhungyel OP, Whittington RJ, et al. Twitching motility is essential for virulence in Dichelobacter nodosus. Journal of bacteriology. 2008; 190(9):3323–35. Epub 2008/03/04. doi: 10.1128/jb.01807-07 PMID: 18310333; PubMed Central PMCID: PMCPMC2347375. 42. Chiang P, Burrows LL. Biofilm formation by hyperpiliated mutants of Pseudomonas aeruginosa. Journal of bacteriology. 2003; 185(7):2374–8. Epub 2003/03/20. PMID: 12644510; PubMed Central PMCID: PMCPMC151504. 43. Koenig JE, Boucher Y, Charlebois RL, Nesbo C, Zhaxybayeva O, Bapteste E, et al. Integron-associ- ated gene cassettes in Halifax Harbour: assessment of a mobile gene pool in marine sediments. Envi- ron Microbiol. 2008; 10(4):1024–38. doi: 10.1111/j.1462-2920.2007.01524.x PMID: 18190517. 44. Szekeres S, Dauti M, Wilde C, Mazel D, Rowe-Magnus DA. Chromosomal toxin–antitoxin loci can diminish large-scale genome reductions in the absence of selection. Molecular microbiology. 2007; 63 (6):1588–605. doi: 10.1111/j.1365-2958.2007.05613.x PMID: 17367382 45. Leplae R, Geeraerts D, Hallez R, Guglielmini J, Dreze P, Van Melderen L. References doi: 10.1371/journal.pone.0005276 PMID: 19390587 21. Aas JA, Paster BJ, Stokes LN, Olsen I, Dewhirst FE. Defining the normal bacterial flora of the oral cav- ity. Journal of clinical microbiology. 2005; 43(11):5721–32. doi: 10.1128/JCM.43.11.5721-5732.2005 PMID: 16272510; PubMed Central PMCID: PMC1287824. 22. Dewhirst FE, Chen T, Izard J, Paster BJ, Tanner AC, Yu WH, et al. The human oral microbiome. Jour- nal of bacteriology. 2010; 192(19):5002–17. Epub 2010/07/27. doi: 10.1128/jb.00542-10 PMID: 20656903; PubMed Central PMCID: PMCPMC2944498. 23. Diaz-Torres ML, Villedieu A, Hunt N, McNab R, Spratt DA, Allan E, et al. Determining the antibiotic resistance potential of the indigenous oral microbiota of humans using a metagenomic approach. FEMS Microbiol Lett. 2006; 258(2):257–62. Epub 2006/04/28. doi: 10.1111/j.1574-6968.2006.00221.x PMID: 16640582. 24. Seville LA, Patterson AJ, Scott KP, Mullany P, Quail MA, Parkhill J, et al. Distribution of tetracycline and erythromycin resistance genes among human oral and fecal metagenomic DNA. Microbial drug resis- tance (Larchmont, NY). 2009; 15(3):159–66. Epub 2009/09/05. doi: 10.1089/mdr.2009.0916 PMID: 19728772. 25. Roberts AP, Mullany P. Oral biofilms: a reservoir of transferable, bacterial, antimicrobial resistance. Expert Rev Anti Infect Ther. 2010; 8(12):1441–50. doi: 10.1586/eri.10.106 PMID: 21133668. 26. Xie G, Chain PS, Lo CC, Liu KL, Gans J, Merritt J, et al. Community and gene composition of a human dental plaque microbiota obtained by metagenomic sequencing. Molecular oral microbiology. 2010; 25 (6):391–405. Epub 2010/11/03. doi: 10.1111/j.2041-1014.2010.00587.x PMID: 21040513; PubMed Central PMCID: PMCPMC2975940. 27. Coleman N, Tetu S, Wilson N, Holmes A. An unusual integron in Treponema denticola. Microbiology (Reading, England). 2004; 150(Pt 11):3524–6. Epub 2004/11/06. doi: 10.1099/mic.0.27569-0 PMID: 15528643. 28. Huang X. A contig assembly program based on sensitive detection of fragment overlaps. Genomics. 1992; 14(1):18–25. Epub 1992/09/01. PMID: 1427824. 29. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. Journal of molecular biology. 1990; 215(3):403–10. Epub 1990/10/05. doi: 10.1016/s0022-2836(05)80360-2 PMID: 2231712. 30. Mazel D. Integrons: agents of bacterial evolution. Nat Rev Microbiol. 2006; 4(8):608–20. Epub 2006/07/ 18. doi: 10.1038/nrmicro1462 PMID: 16845431. 31. Stokes HW, O'Gorman DB, Recchia GD, Parsekhian M, Hall RM. Structure and function of 59-base ele- ment recombination sites associated with mobile gene cassettes. Molecular microbiology. 1997; 26 (4):731–45. PMID: 9427403. 32. Wu Y-W, Doak TG, Ye Y. The gain and loss of chromosomal integron systems in the Treponema spe- cies. BMC evolutionary biology. 2013; 13(1):1–9. PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016 References Diversity of bacterial type II toxin-antitoxin systems: a comprehensive search and functional analysis of novel families. Nucleic acids research. 2011; 39(13):5513–25. Epub 2011/03/23. doi: 10.1093/nar/gkr131 PMID: 21422074; PubMed Central PMCID: PMCPMC3141249. 46. Makarova KS, Grishin NV, Koonin EV. The HicAB cassette, a putative novel, RNA-targeting toxin-anti- toxin system in archaea and bacteria. Bioinformatics. 2006; 22(21):2581–4. doi: 10.1093/ bioinformatics/btl418 PMID: 16895922. 47. Jorgensen MG, Pandey DP, Jaskolska M, Gerdes K. HicA of Escherichia coli defines a novel family of translation-independent mRNA interferases in bacteria and archaea. Journal of bacteriology. 2009; 191 (4):1191–9. Epub 2008/12/09. doi: 10.1128/jb.01013-08 PMID: 19060138; PubMed Central PMCID: PMCPMC2631989. 48. Pandey DP, Gerdes K. Toxin-antitoxin loci are highly abundant in free-living but lost from host-associ- ated prokaryotes. Nucleic acids research. 2005; 33(3):966–76. Epub 2005/02/19. doi: 10.1093/nar/ gki201 PMID: 15718296; PubMed Central PMCID: PMCPMC549392. 49. Christensen-Dalsgaard M, Gerdes K. Two higBA loci in the Vibrio cholerae superintegron encode mRNA cleaving enzymes and can stabilize plasmids. Molecular microbiology. 2006; 62(2):397–411. Epub 2006/10/06. doi: 10.1111/j.1365-2958.2006.05385.x PMID: 17020579. 50. Boucher Y, Nesbo CL, Joss MJ, Robinson A, Mabbutt BC, Gillings MR, et al. Recovery and evolution- ary analysis of complete integron gene cassette arrays from Vibrio. BMC evolutionary biology. 2006; 6:3. Epub 2006/01/19. doi: 10.1186/1471-2148-6-3 PMID: 16417647; PubMed Central PMCID: PMCPMC1382264. 51. Chen X-L, Tang D-J, Jiang R-P, He Y-Q, Jiang B-L, Lu G-T, et al. sRNA-Xcc1, an integron-encoded transposon- and plasmid-transferred trans-acting sRNA, is under the positive control of the key viru- lence regulators HrpG and HrpX of Xanthomonas campestris pathovar campestris. RNA Biology. 2011; 8(6):947–53. doi: 10.4161/rna.8.6.16690 PMID: 21941121 20 / 20 PLOS ONE \| DOI:10.1371/journal.pone.0157605 June 15, 2016
https://openalex.org/W4386781077	https://www.e3s-conferences.org/articles/e3sconf/pdf/2023/63/e3sconf_icobar23_02141.pdf	English	null	The effect of eWOM and e-service quality on purchase intention at Artic.house restaurant	E3S web of conferences	2,023	cc-by	6,457	1 Introduction Advertising, promotion, advancement (promotion), sales, personal selling, and promotional activities publicity such as public relations are type of promotion [4]. The first step in creation of promotion is to identify target groups that focus on advertising. Therefore, this study will emphasis on advertising and public relation that connected with Electronic Word of Mouth (eWOM). eWOM is a positive or negative statement made by potential customers, actual customers, and former customers about a product or company via the internet [5]. eWOM can communicate about new items, recommends new uses for the product, informs about changing products costs, and describes how the product works. eWOM purposively serve to convince an economic propaganda (advertisement) to face competition. Remembering propaganda is engaged in the maturation stage of items, which customers continue to ponder about [5]. A corporation must establish a positive image and handle rumours or other discouraging news in order to build strong connections with diverse groups of people and accomplish targeted exposure [6]. Along with the development of technology and information, tourism industry cannot be separated from its connection with the internet. As such, customer view all aspects of reviews, from e-commerce to food and beverage via the internet. In this sense, the rapid development of social media has forced business actors to use social media application to compare to traditional marketing such as billboards, radio and TV advertisements [1]. Moreover, people critically examine the food item that they order for advantages or disadvantages in terms of presentation, taste, packaging, and the delivery process to customers. So that there is a working atmosphere that is smooth, dynamic, and conducive to minimize the occurrence of errors in preparing and meeting customer needs. Marketing management is the process of developing, conducting, and tracking an organization ‘s marketing strategy [2]. This involves the marketing strategy, techniques, and methods used to generate and fulfil demand from target consumers to increase profitability. Types of marketing are product, place, price, and promotion or can be called 4P [3]. This study focusses on promotion types of marketing strategy since benefiting to increases the speed of product and service acceptance as well as promote the sales of goods and services in imperfect market conditions [2]. In short, an effective promotion strategy provides effectual sales. The effect of eWOM and e-service quality on purchase intention at Artic.house restaurant Maharani Juliette Limwirya1, Kelly Kelly 1, and Indra Kusumawardhana1 1Hotel Management Department, Faculty of Digital Communication and Hotel & Tourism, Bina Nusantara University, Jaka Indonesia 11480 Abstract. This study aims to determine the effect of eWOM and e-service quality on purchase intentions at Arctic House restaurant. Quantitative data is formulated through the SPSS application. This study uses more than two variables, so this study uses multiple correlation analysis using the partial hypothesis test T, F test, and determination test. The population in this study are people who meet the same criteria in the study. This study took a sample of 97 people. This study found that eWOM and e-service quality have a positive effect on purchase intention at Arctic.House Restaurant. This implies that higher service quality and positive reviews from previous customers increase the likelihood that consumers will buy a restaurant's products or services. The results of hypothesis testing show that the independent variables together have an influence of 88.9% on Purchase Intention at Arctic.House Restaurant, this shows that there is a significant influence of the independent variables on the dependent variable. Corresponding author: maharani.limwirya@binus.ac.id © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http ://creativecommons.org/licenses/by/4.0/). s E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 DP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 ses/by/4 0/) y EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 licenses/by/4.0/). 1 Introduction Promotion is practically stated as a communication system or collection of tactics, forms, and shapes that communicate information about products, services, and general customers on the action of businesses and larger environment [4]. In the context of a restaurant, eWOM can play a significant role in shaping its reputation and influencing consumer behaviour. Positive reviews and comments can attract new customers and generate more business, while negative reviews can have the opposite effect. Therefore, it's crucial for restaurant owners and managers to monitor and manage their online reputation [7]. Some strategies for leveraging eWOM to benefit a restaurant include encouraging satisfied customers to leave reviews on popular review websites such as Instagram, Google Reviews, Yummy advisor and applicable application. By actively engaging with Corresponding author: maharani.limwirya@binus.ac.id * E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 field training on. The researcher is already doing an observe for a year, from August 2021 to July 2022. eWOM, restaurants can build a positive reputation, increase customer loyalty, and ultimately drive more business. A lack of understanding and focus on fundamentals of marketing management could have a negative impact on restaurant business in the long run. Additionally, the problem of marketing strategy in a restaurant still overlooked [12–14]. Particularly the discussion on eWOM and e-service quality to influence customer decisions towards purchase intention at Arctic House Restaurant. For example, eWOM at the Arctic House restaurant that collected through Google Reviews, Yummy advisor, pergikuliner.com display as below. Table 1 eWOM for ARTIC.House Table 1 eWOM for ARTIC.House eWOM Positive Negative Google review https://g.co/kgs/ 1. Good Food 2. Varians Menu 3. The place is strategic 1. Lack Of Quality control of food 2. The food is expensive Yummy advisor Arctic House Restoran Kafe terdekat di Jakarta Jakarta (yummyadvisor.id) 1. Good design interior 2. Good food 3. Good beverage 1. The outdoor is too hot 2. The food quality Pergikuliner.com Arctic House, Puri Lengkap: -Menu terbaru, jam buka, & no telepon, alamat dengan peta (pergikuliner.com) 1. Aesthetic place and food 2. Good food 3. Good ambience 1. The quality sometimes offered badly 2. The services not too good Source: Adapted from Google review, yummy advisor, and pergikuliner.com Therefore, this study builds two hypotheses to develop the scientific discussion. The first one is H1 eWOM has a positive and significant effect on purchase intention at Arctic House restaurant. 1 Introduction And the second one H2 e- service quality has a positive and significant effect on purchase intention at Arctic House restaurant. The illustration of research framework display as below. Fig. 1 Hypothesis Framework Purchase Intention Electronic Word Of Mouth E-Service Quality Purchase Intention Electronic Word Of Mouth E-Service Quality Source: Adapted from Google review, yummy advisor, and pergikuliner.com Source: Adapted from Google review, yummy advisor, and pergikuliner.com The table 1 explains that Restaurant Arctic House has received both positive and negative reviews on Google Reviews. From the existing positive comments, customers more often express their opinions about the food, setting and service. Consumers are more likely to complain about high food prices and narrow parking spaces through negative comments. Fig. 1 Hypothesis Framework Fig. 1 Hypothesis Framework 2.1 eWOM (Electronic Word of Mouth) Intensity, the number of opinions or judgments published by consumers on a social media is 2 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 referred to as the intensity of electronic word of mouth (eWOM). The effect of the intensity of reading other consumer reviews on the internet on restaurant business. and this can help consumers choose a product that consumers want and need. It moves efficiently and can also be easier. Privacy/security, Privacy is useful for maintaining the security of consumer data that consumers use to process transactions. Privacy/security, Privacy is useful for maintaining the security of consumer data that consumers use to process transactions. 2. 2. Valence of Opinion, Valence of Opinion refers to how consumers feel toward products, services, and brands, both positive and negative. The difference between expectations and perceived performance or achievements is referred to as customer satisfaction. Opinion valence is classified into two types: negative and positive. x Fulfillment/reliability, Fulfillment is a service that is useful for dealing with a problem when a problem occurs with the company and also dealing with all complaints submitted by consumers in a friendly and kind manner. yp g p 3. Information, the amount of data on social media about products and services is known as content. To ensure client pleasure, the company's products must be of high quality. This is because satisfaction is also affected by the quality of products and services. A business owner has to create an effective communication program aimed at customers to transmit existing knowledge and aims to produce purchases that lead to restaurant business profits in order to promote a product [17]. y x Site Aesthetic, Site aesthetic is a dimension that is tangible or can be said to be visible to the naked eye because site aesthetic is useful for showing a feature available on the website, and a reflection of the image of a company. x Responsiveness, Responsiveness is to measure a timeliness and a good and appropriate response when answering a question related to a product or service. x Ease of use, Ease of use is an interface from a website that is comfortable and also websites or social media can be used comfortably [21]. 2.1 eWOM (Electronic Word of Mouth) On the other hands, eWOM and e-service quality are two important concepts in the realm of e-commerce. E- service quality is a service that has been provided on social media or internet networks to be used as an extension because it can provide the ability to facilitate consumer activities in shopping, making purchases, and also distribution effectively and efficiently [8]. A service known as e-quality of service, is one that is offered on a network of websites as an enhancement of the site's capacity to facilitate activities such as shopping, purchasing, and distribution effectively and efficiently [9]. Since eWOM is an exchange of information communication between old customers and new customers, by using technological developments such as online discussion forums, reviews, websites, and social media networking sites that provide information exchange facilities among customers. communicator. Promotions strategy using eWOM can disseminate information briefly, quickly, and the information can spread more widely than using conventional promotions [15]. In the traditional way, word of mouth communication was done face to face with people you know, but now word of mouth can be done in seconds with a very wide scope, what we share can be read with other people [16]. This paradigm shift in word of mouth through social media is called electronic word of mouth. This information is given to the recipient again without asking or searching but digitalization as a form of very rapid technological development with the main goal of providing convenience and efficiency from various aspects, energy efficiency, procedural costs. This research discusses the topic of eWOM and e- service quality as they in line with the recent conversation about marketing management that evolves from traditional to digital transformation. Therefore, it is crucial to embrace transformative challenges in combination with economic returns, resulting in a new sustainable service era for restaurant business [10], particularly that affect to purchase and re-purchase intention in the restaurant. Artic house Restaurant is newly brand and uses the modern marketing as promotion strategy in their marketing management [11]. Additionally, the researchers have the opportunity to observe their promotion activity when completing the supervised Previous study provides the dimension that used in quantifying the influence of Electronic Word of Mouth (eWOM): 1. Intensity, the number of opinions or judgments published by consumers on a social media is 1. 2.1 eWOM (Electronic Word of Mouth) But this study adopted the eWOM indicator based on the 3 pointers, namely: But this study adopted the eWOM indicator based on the 3 pointers, namely: x Sharing Information, Information sharing is a way for consumers to share information about a business's products or services with other consumers. Examples of social media that consumers use are Line, WhatsApp, and Instagram. 2.3 Purchase Intention Purchase intention can be regarded as a consumer decision regarding preferences for brands in the choice set [22]. Furthermore, intention is an intention that the dimensions of consumer's desire to make a repeat purchase. There are several other meanings, namely: x Trust, Trust is a level of trust that consumers believe in information that consumers get from other consumers through social media. x Intention can be spelled out as a trap or is an intermediary between motivational factors that can influence consumer behaviours. x Information attraction, the attractiveness of the information is how interesting the information is and is in accordance with the evaluation of recipients of information from the internet and social media and is measured according to consumer ratings [18]. x Intention can be defined by how far consumers have the willingness to try and make transactions. x Intention also shows a measurement of consumer will. g y x Intention also shows a measurement of consumer will. x Intention can relate to persistent behaviours [23]. x Intention can relate to persistent behaviours [23]. 2.2 E-Service Quality x Persuasion, which is consumer confidence in an item or service. x Persuasion, which is consumer confidence in an item or service. ݊= ܰ 1 + ܰ(݁)ଶ n = sample size N = population size e = margin of error Result: ݊= ଷହ଻଼ ଵାଷହ଻଼(଴.ଵ)మ (1) ݊= ଷହ଻଼ ଵାଷହ,଻଼ (2) ݊= ଷହ଻଼ ଷ଺,଻଼ (3) ݊= 97,2811310495 (4) x Purchase, namely the purchase or consumer transaction on a good or service [26]. x Purchase, namely the purchase or consumer transaction on a good or service [26]. ݊= ܰ 1 + ܰ(݁)ଶ 2.4 Hypothesis Hypotheses as assumptions or propositions which state the expected relationship between variables. It serves as a guide for research and provides a basis for making predictions and drawing conclusions. The first hypothesis examines the relationship between eWOM and purchase intention. Previous studies have consistently found a positive effect of eWOM on purchase intention. Positive reviews and recommendations shared through online platforms are expected to improve customer perceptions of restaurants and increase their intention to make purchases [30]. The sample is a representation of the size or characteristics of the population [28]. So the total of the sample is 97 sample need. The sample is suitable for use because it is representative of the target population to be studied and conformity with the research objectives. The sample includes respondents who visited Artic.house restaurant from various age groups and consumer backgrounds, which will help in generalizing the research findings to a wider population. This allows researchers to gain a more comprehensive understanding of the effect of eWOM and e-service quality on purchase intentions at Artic.house restaurant. The second hypothesis focuses on the effect of e- service quality on purchase intention. Research has established a strong relationship between service quality and purchase intention. It was hypothesized that a higher level of e-service quality, including factors such as responsiveness, reliability, and ease of use, would have a positive impact on customer intentions to purchase at Artic.house restaurants [31]. 4 Result This study used quantitative research techniques. To arrive at a conclusion, the authors use quantitative research methods based on concrete data. Research data can be in the form of numbers which will eventually be measured using statistics as a calculation test tool the application of the tool called SPSS with 21 version. The research instrument is a tool that can be used to measure natural or social phenomena to be observed by researchers. The tests used are validity test, reliability test, classical assumption test, normalization test, descriptive test, multicollinearity test, heteroscedasticity test, autocorrelation test, hypothesis test, partial test, simultaneous test, coefficient of determination. Data collection techniques are very important because they will always be involved in research. There are various methods of collecting data for research, including primary data and secondary data. 2.2 E-Service Quality An intention is a factor of motivation that can influence a person's behaviours to do something [23]. Thus, purchase intention is the beginning of the consumer's desire to get service in the future, and this can also be regarded as customer shopping behaviours. And if the customer is happy and satisfied with a product, it will be called shopping intention [24]. E-service quality is a service that has been provided on social media or internet networks to be used as an extension because it can provide the ability to facilitate consumer activities in shopping, making purchases, and also distribution effectively and efficiently [19]. Moreover, e-service quality is an important tool for a company where a consumer's needs will be automatically channeled via the internet or social media and in the consumption life cycle [20]. Thus, according to Blut, e-service quality is a transaction from start to finish and also includes information search, privacy policy, website navigation, ordering process, customer service interaction, delivery, return policy, and also a consumer's satisfaction with the product that the customer buys. consumption or order. That is, traditional service quality has a drawback where there are salesmen who have to directly provide a service to consumers [21]. Additionally, e service has several indicators that applied to this study, namely: Therefore, it is necessary to identify the purchase intention of consumers that said purchase intention as a cognitive plan or can be called a consumer desire for a particular item or brand. And this purchase intention can be measured by asking about the probability of buying the advertised product [25]. At last, this study follows Kotler and Armstrong who stated that customer purchase intention has 6 indicators, namely: x Awareness, which is consumer awareness of a product or service. x Knowledge, which is a consumer's knowledge of a product or service. x Interest, namely the interest of a consumer in buying or using goods and services. x Efficiency, Efficiency is a function on a site or social media that can provide correct information, x Efficiency, Efficiency is a function on a site or social media that can provide correct information, x Preference or preferences. 3 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 the total margin of error is 10 percent. The formula is as follow. x Persuasion, which is consumer confidence in an item or service. 4.1 Profile Respondents Primary data is a source of information that is given directly to data collectors. This study will collect data independently, originating from initial data sources or research project locations [27]. The data will be collected by using google from that will be spread from 23 March 2023 until 5 April 2023. Arctic house in the West Jakarta area will be the location of this research. The meaning of the object of research is an attribute or also the value of people, objects and activities that have certain variations that have been determined by researchers so that they can be studied and will eventually be studied. come to a conclusion [28]. 4.1.1 Age Based on the data provided, there are several findings that can be concluded. First, the 13–27-year-old category had the highest number of respondents or visitors, namely 77 people, or around 50% of the total respondents in that population. This shows that the relatively young population constitutes the majority of visitors or respondents in this context. Furthermore, the 28–46-year age category had 71 respondents or visitors, which accounted for around 46.10% of the total respondents. This category also has a significant 4.1.1 Age 4.1.1 Age Operational variables in this study are eWOM (Electronic Word of Mouth) Variables, e-service quality Variables, and purchase intention Variables [19]. Based on the data provided, there are several findings that can be concluded. First, the 13–27-year-old category had the highest number of respondents or visitors, namely 77 people, or around 50% of the total respondents in that population. This shows that the relatively young population constitutes the majority of visitors or respondents in this context. Furthermore, the 28–46-year age category had 71 respondents or visitors, which accounted for around 46.10% of the total respondents. This category also has a significant Populations are topics and objects whose size and other characteristics have been defined by the researcher so that they can be examined and from which conclusions can be drawn. There are two types of populations, namely limited populations and unlimited populations [29]. The population size will be the follower of artic. house Instagram the total of the follower is 3578. With 4 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 did not fall into the previously mentioned job categories. This number accounts for about 1.9% of the total respondents. number, although slightly lower than the 13–27-year age category. However, there are quite striking differences in the number of respondents or visitors in the older age category. The 47–58-year age category only had 6 respondents or visitors, which only covered 3.90% of the total respondents. Meanwhile, there were no respondents or visitors who were in the age range of 59- 77 years, with a percentage of 0%. This data provides an overview of the age composition of visitors or respondents in the population. It can be concluded that the majority of visitors or respondents are in the young age category (13-27 years) and middle age (28-46 years), while the number of older respondents or visitors is very limited. 4.1.2 Gender 4.1.2 Gender Based on the data provided, there are two identified genders, namely male and female. There were 62 respondents or visitors who were male, which accounted for around 40% of the total respondents in the population. In addition, there were 92 respondents or visitors who were women, which covered around 59.70% of the total respondents. This data provides an overview of the gender composition of visitors or respondents in that population. It can be concluded that the population of respondents or visitors is dominated by women, with a higher percentage than men. This information can be useful in understanding the preferences or needs that may differ between men and women in the context of experiences at Arctic House. 4.1.3 Type Of Work Respondents 4.2 Validity Test 4.2 Validity Test Table 2. Output Validity Test Variabel Indikator Hasil Keterangan E-wom Review produk dan pelayanan 0.000 Valid Produk dan pelayanan direview dan dikomentar oleh situs web besar seperti zomato, google review. 0.000 Valid Konsumen akan lebih tertarik setelah melihat review sebelumnya. 0.000 Valid E-service quality Konsumen lebih praktis dalam mencari informasi 0.000 Valid Keamanan data konsumen terjaga pada saat mencari informasi atau melakukan transaksi 0.000 Valid Keluhan konsumen teratasi dengan baik 0.000 Valid Informasi dan situs web restoran terlihat menarik 0.000 Valid Kecepatan tanggapan restoran mengatasi keluhan konsumen 0.000 Valid Kepraktisan konsumen dalam menggunakan atau melakukan transaksi secara online 0.000 Valid Purchase Intention Konsumen menyadari adanya restoran 0.000 Valid Konsumen mengetahui dengan baik tentang restoran 0.000 Valid Konsumen memiliki keinginan membeli 0.000 Valid Konsumen memberikan saran membeli kepada konsumen lain 0.000 Valid Konsumen menjadi langganan restoran 0.000 Valid Konsumen melakukan transaksi 0.000 Valid Table 2. Output Validity Test Table 6. Output Multicollinearity Test Coefficientsa Model Unstandard ized Coefficient s Standa rdized Coeffi cients t Sig. Collinearit y Statistics B Std. Error Beta Toler ance VIF 1 (Constant) .114 .127 .897 .371 Total_EW OM -.009 .014 -.298 -.662 .509 .028 6.084 Total_Eser vice .013 .010 .568 1.261 .209 .028 8.084 4.3 Reability Test Table 3. Output Reability Test Variabel Hasil Keterangan E-wom 0.984 Reliabel E-service quality 0.989 Reliabel Purchase Intention 0.926 Reliabel Table 3. Output Reability Test Based on these results, it can be concluded that the tested data follows a normal distribution. This is indicated by the asymp value. Sig. (2-tailed) which is greater than 0.05, which indicates that there is not enough evidence to reject the hypothesis that the data comes from a normal distribution. Based on the results of this study, the variable E-wom (Electronic Word of Mouth) has a reliability measurement result of 0.984. This figure shows that the measurement for the E-wom variable is reliable or reliable in the context of this study. Furthermore, the E- service quality variable (Electronic Service Quality) has a reliability measurement result of 0.989. This high value indicates that the measurement for the E-service quality variable is also reliable or reliable in this study. The Purchase Intention variable has a reliability measurement result of 0.926. Although slightly lower than the previous variable, this figure still shows that the measurement for the Purchase Intention variable is reliable in this study. Thus, all the variables in this study have high reliability measurement results, indicating that the measurements made on these variables can be considered consistent and reliable. 4.1.3 Type Of Work Respondents 4.1.3 Type Of Work Respondents Based on the data provided, there are several types of work identified. There were 52 respondents or visitors who were students, which accounted for around 33.8% of the total respondents in the population. In addition, there were 19 respondents or visitors who were housewives, which covered around 12.3% of the total respondents. Furthermore, there were 53 respondents or visitors who were private employees, which covered around 34.4% of the total respondents. In addition, there were 27 respondents or visitors who were entrepreneurs, which covered around 17.5% of the total respondents. In addition, there were 3 other respondents or visitors who Based on the results of product moment validity testing, it is known that each item for the relationship with total has a sig.(2-tailed) value of 0.000 <0.05 and the Pearson correlation is positive. So, it can be concluded that the question. Questionnaire used are valid, meaning that they can be used as an accurate data collection tool in research conducted on "The Effect Of 5 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 Normal Parametersa,b Mean .0000000 Std. Deviation .32604300 Most Extreme Differences Absolute .232 Positive .152 Negative -.232 Kolmogorov-Smirnov Z 2.583 Asymp. Sig. (2-tailed) .982 eWOM and e-service Quality on Purchase Intention at Arctic.house Restaurant". eWOM and e-service Quality on Purchase Intention at Arctic.house Restaurant". 4.5.3 Heteroscedasticity Test 4.5.3 Heteroscedasticity Test Fig. 5. Output Heteroscedasticity Test Based on the test results, it was found that there was no heteroscedasticity that occurred in the question items, due to the diversity of univariate data distribution. Based on the results of the descriptive test it can be seen that N or the number of respondents is 167 with the valid level of the respondent's answers being 164, the minimum value in the answer score is 0 and the maximum value of each respondent's answer is 5, in the descriptive test the average answer is equal to 4.00, then with an average number equal to 600, and also a standard deviation of 1,000 and above. Table 4. Output Descriptive Test Table 4. Output Descriptive Test Based on these data, it can be concluded that there is a relationship between the independent variables and the dependent variable. However, the results of the multicollinearity test showed no indication of multicollinearity between the two independent variables, indicated by the VIF value which was <10.00 the threshold value. Descriptive Statistics N Mini mum Maxim um Mean Std. Deviation Total_EWOM 167 0 65 50.87 15.051 Total_Eservice 167 0 90 72.19 21.263 Total_purchas eintention 167 0 10 8.07 2.468 Valid N (listwise) 167 4.5 Classic Assumption Test 4.5.1 Normality Test Table 5. Output Normality Test One-Sample Kolmogorov-Smirnov Test Unstandardized Residual N 165 4.5.1 Normality Test 4.5.1 Normality Test Table 5. Output Normality Test Table 5. Output Normality Test One-Sample Kolmogorov-Smirnov Test Fig. 5. Output Heteroscedasticity Test Based on the test results, it was found that there was no heteroscedasticity that occurred in the question items, due to the diversity of univariate data distribution. Unstandardized Residual 165 6 6 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 4.6 Hypothesis Test 4.6.1 Partial Test (T Test) Table 7. Output Partial Test ANOVAa Model Sum of Squares df Mean Square F Sig. 1 Regre ssion 922.013 2 461.007 848.309 .000b Residu al 89.124 164 .543 Total 1011.138 166 4.6 Hypothesis Test Based on the results of hypothesis testing, it can be interpreted that the independent variable has a significant influence on the dependent variable in this model. This is shown by a very low significance value, showing strong evidence that the independent variable has a significant effect on the dependent variable. All independent variables in this model have a significant influence on the dependent variable. This can be seen from the calculated F value which is greater than the F table value (36.669 > 3.05). Therefore, the independent variables simultaneously have an effect of 88.9% on Purchase Interest at Arctic.House Restaurant. Based on the results of hypothesis testing, it can be interpreted that the independent variable has a significant influence on the dependent variable in this model. This is shown by a very low significance value, showing strong evidence that the independent variable has a significant effect on the dependent variable. All independent variables in this model have a significant influence on the dependent variable. This can be seen from the calculated F value which is greater than the F table value (36.669 > 3.05). Therefore, the independent variables simultaneously have an effect of 88.9% on Purchase Interest at Arctic.House Restaurant. 4.6.1 Partial Test (T Test) Table 7. Output Partial Test p ANOVAa Model Sum of Squares df Mean Square F Sig. 1 Regre ssion 922.013 2 461.007 848.309 .000b Residu al 89.124 164 .543 Total 1011.138 166 In addition, the results of the coefficient of determination test show that the Adjusted R Square value is 0.899. This indicates that the independent variables together have an influence of 88.9% on the dependent variable. This conclusion indicates that the independent variable has a large influence on the dependent variable. Table 9. Output Coefficient of Determination Model Summaryb Model R R Square Adjusted R Square Std. Error of the Estimate 1 .948a .899 .874 .460 Although the results of this study indicate a significant effect of the independent variables on Purchase Intention, there are other factors that may contribute to customer buying behavior. Therefore, it is advisable to carry out further research to identify additional factors that may influence Purchase Intention and develop more effective strategies to improve restaurant business performance. Although the results of this study indicate a significant effect of the independent variables on Purchase Intention, there are other factors that may contribute to customer buying behavior. Therefore, it is advisable to carry out further research to identify additional factors that may influence Purchase Intention and develop more effective strategies to improve restaurant business performance. Based on the results of the coefficient of determination test, the Adjusted R Square value is 0.899, this indicates that all independent variables simultaneously have an effect of 88.9% on Purchase Intention at Arctic.House Restaurant. This means that the independent variables in the study have a large influence on the dependent variable. Arctic.House Restaurant Company needs to carry out further research to understand the factors that influence customer purchase intentions online. In this 4.5.1 Normality Test Thus, it can be interpreted that all the independent variables in this study have a significant influence on Purchase Interest at Arctic.House Restaurant. The partial T test also involves an analysis of the residuals from the model. The sum of squares for the model remnants is 89.124 with a df of 164. The mean square for the remnants is calculated by dividing the sum of squares by the df, which in this case is 0.543. A very low significance value indicates strong evidence. A very low significance value indicates strong evidence that the independent variable has a significant effect on the dependent variable in this model. In this case it can be said that the study the effect of eWOM and e-service quality on purchase intention at Artic.house restaurant has a significant and simultaneous effect according to the tests carried out Table 8. Output Simultaneous Test ANOVAa Model Sum of Squares df Mean Square F Sig. 1 Regression 248.090 32 7.753 36.699 .000b Residual 27.885 132 .211 Total 275.976 164 x Improving eWOM (Electronic Word of Mouth), because eWOM is also proven to have a significant influence on Purchase Intentions, is important to encourage and expand positive eWOM about Arctic.House Restaurants. Engaging customers to leave positive reviews and recommend this place to others via online platforms can help expand your reach and enhance a restaurant's image. Based on the test results, it was concluded that all independent variables or independent variables have a significant influence on the dependent or dependent variable. This can be seen from the calculated F value, which is equal to 36,669 > 3.05 F table. x Optimizing digital marketing strategies In order to increase Buying Interest, restaurants need to optimize digital marketing strategies. This could include engaging and informative online marketing campaigns, using social media to interact with customers and introduce special offers, as well as ensuring the availability and ease of use of digital payments to increase customer security and convenience in making payments. 4.6.2 Simultaneous test (Test F) x Based on the finding that the independent variable has a significant effect on Purchase Intention, it is suggested to improve the overall service quality. Focus on aspects such as service speed, employee friendliness, and improving the quality of products and food offered. Reference 1 N. Pratama, “Analisa Pengaruh Social Media Marketing Activities, Ewom dengan Variable Mediasi Brand Experience terhadap Consumer Based Brand Equity pada Industri Franchise Food and Beverage di Kota Batam.” Universitas Internasional Batam, (2022). 16 F. C. Sitanggang and R. Hidayat, “Pengaruh Electronic Word Of Mouth (ewom) Dan Kualitas Website Terhadap Minat Beli (studi Kasus Pada Warunk Upnormal Di Kota Bandung Tahun 2018),” eProceedings Appl. Sci., vol. 4, no. 2, (2018). 2 A. Grubor and O. Jakša, “Internet Marketing as a Business Necessity,” Interdiscip. Descr. Complex Syst. INDECS, vol. 16, no. 2, pp. 265–274, (2018). 17 T. R. Kristianti and R. Erdiansyah, “Pengaruh Electronic Word of Mouth, Tingkat Kepercayaan Konsumen dan Minat Kunjungan Wisata Kuliner Kota Bogor,” Prologia, vol. 4, no. 2, pp. 393–401, (2020). 3 K. K. Halim and S. Halim, “Business Intelligence For Designing Restaurant Marketing Strategy: A case study,” Procedia Comput. Sci., vol. 161, pp. 615–622, (2019). 18 J. Djimantoro and V. C. Gunawan, “Pengaruh e- wom, Kualitas Makanan dan Kualitas Layanan Terhadap Keputusan Pembelian di Rumah Makan Kakkk Ayam Geprek,” J. Hosp. dan Manaj. Jasa, vol. 8, no. 2, (2020). 4 M. Saravanakumar and T. SuganthaLakshmi, “Social Media Marketing,” Life Sci. J., vol. 9, no. 4, pp. 4444–4451, (2012). 5 D. Novak, “Promotion as Instrument of Marketing Mix,” Eng. Manag. Compet., vol. 510, (2011). 19 M. I. Syacroni, H. Hamdan, and R. R. Ilhamalimy, “The Effect of E-Service Quality , Website Quality , Promotion , and E-Trust on Repurchase Intentions at E-Commerce Shopee in DKI Jakarta,” pp. 193–205, (2023). 6 H. G. Thiasyura, C. W. Utami, and M. Padmalia, “Plan Determination for Promotion Mix for SKETCH!,” (2017). 7 A. Rumondang et al., “Pemasaran Digital dan Perilaku Konsumen,” (2020). 20 R. Septiani, “Pengaruh E-Service Quality, Persepsi Harga, dan Promosi Penjualan terhadap Loyalitas Pelanggan melalui Kepuasan Pelanggan (Studi pada Pelanggan Jasa Transportasi Online Gojek di Yogyakarta),” J. Ilmu Manaj., vol. 17, no. 2, pp. 98–108, (2020). 8 B. Lariviere and E. G. Smit, “People–planet– Profits for a Sustainable World: Integrating the Triple-P Idea in the Marketing Strategy, Implementation and Evaluation of Service Firms,” J. Serv. Manag., vol. 33, no. 4/5, pp. 507–519, (2022). 21 A. Juhria, N. Meinitasari, F. I. Fauzi, and A. Yusuf, “Pengaruh E-service Quality Terhadap Kepuasan Pelanggan di Aplikasi E-commerce Shopee,” J. Manaj., vol. 13, no. 1, pp. 55–62, (2021). 9 A. Magdalena, “Analisis Antara E-service Quality, E-satisfaction, dan E-loyalty Dalam Konteks E- commerce BLIBLI,” J. 5 Conclusion 7 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 study, factors such as price, product quality, and customer trust need to be considered in more depth. 15 A. C. Mahendrayasa, S. Kumadji, and Y. Abdillah, “Pengaruh Word of Mouth Terhadap Minat Beli Serta Dampaknya Pada Keputusan Pembelian (Survei pada Mahasiswa Pengguna Kartu Selular gsm ‘IM3’ Angkatan 2011/2012 dan 2012/2013 Fakultas Ilmu Administrasi Universitas Brawijaya Malang),” J. Adm. Bisnis, vol. 12, no. 1, (2014). Reference Strateg. Pemasar., vol. 5, no. 2, (2018). 10 N. Permatasari and S. P. Nugroho, “Pengaruh Ewom Dan E-Service Quality Terhadap Keputusan Pembelian Melalui Mediasi Kepercayaan (Studi Kasus Masyarakat Karesidenan Madiun yang Bertransaksi Kendaraan Bermotor di Marketplace).” Universitas Muhammadiyah Surakarta, (2022). 22 A. A. Satria, “Pengaruh Harga, Promosi, dan Kualitas Produk Terhadap Minat Beli Konsumen Pada Perusahaan A-36,” J. Manaj. Dan Start-Up Bisnis, vol. 2, no. 1, pp. 45–53, (2017). 23 R. C. Tantama, “Faktor-faktor Yang Mempengaruhi Pemilihan Smartphone (Studi Pembelian Konsumen Pada Produk Smartphone di Yogyakarta).” UAJY, (2017). 11 “References,” instagram, 2022. https://www.instagram.com/arctic.house.id/ (accessed Apr. 29, 2023). 24 A. Wibowo, S.-C. Chen, U. Wiangin, Y. Ma, and A. Ruangkanjanases, “Customer Behavior as an Outcome of Social Media Marketing: The Role of Social Media Marketing Activity and Customer Experience,” Sustainability, vol. 13, no. 1, p. 189, (2020). 12 S. Tsviliy, D. Gurova, and T. Kuklina, “Marketing Competitiveness of Hotel and Restaurant Enterprise: Theoretical Approach and Methods of Definition,” VUZF Rev., vol. 6, no. 2, p. 30, (2021). 25 R. Agustin and A. Hasyim, “Membangun Niat Beli Secara Online Melalui Kemudahan dan Kepercayaan Konsumen,” J. Econ., vol. 10, pp. 121–132, (2019). 13 A. Saputra, “Evaluasi Strategi Pemasaran Pada Rumah Makan Sederhana Perspektif Ekonomi Islam.” IAIN Bengkulu, (2020). 14 A. Z. E. Riung, “Pemanfaatan Digital Marketing sebagai Strategi Bisnis (Studi Kasus pada Café dan Resto di Manado).” (2022). 26 G. Armstrong, S. Adam, S. Denize, and P. Kotler, Principles of marketing. Pearson Australia, (2014). 8 8 E3S Web of Conferences 426, 02141 (2023) ICOBAR 2023 https://doi.org/10.1051/e3sconf/202342602141 27 D. Sugiyono, “Metode penelitian Kuantitatif Kualitatif dan R&D, Penerbit Alfabeta.” Jakarta, (2010). 28 J. A. Yani, “Sugiyono. 2017. Metode Penelitian Kuantitatif, Kualitatif, Dan R&D. Bandung: Alfabeta. Ferrari, JR, Jhonson, JL, & McCown, WG (1995). Procrastination And Task Avoidance: Theory, Research & Treatment. New York: Plenum Press. Yudistira P, Chandra. Diktat Kuliah Psikometri. Fakultas Psikologi Universitas”. 29 L. Laily Syahrini and M. Farmansjah, “Analisis Antrian Pada Teller di PT. Bank Rakyat Indonesia Kantor Cabang Kalimalang Duren Sawit Jakarta Timur (Analysis of Queue in Teller at PT. Bank Rakyat Indonesia Kalimalang Branch Office, Duren Sawit, East Jakarta,” Anal. Antrian Pada Teller Di PT Bank Rakyat Indones. Kant. Cab. Kalimalang Duren Sawit Jakarta Timur. 30. Wang, D., & Li, X. R. (2019). Factors influencing purchase intention of personalized products in online stores: An empirical study. Journal of Theoretical and Applied Electronic Commerce Research, 14(3), 55-71. 31. Lien, C. Reference H., Wen, C., & Huang, L. C. (2018). The impact of e-service quality, customer satisfaction, and loyalty on e-marketing: Moderating effect of perceived value. Social Behavior and Personality: An International Journal, 46(2), 283-297. 9
https://openalex.org/W4230219921	https://figshare.com/articles/thesis/_Shall_we_put_the_heart_in_now_A_comparative_analysis_between_creature_features_and_their_single_reel_abridgements/14647653/1/files/28127853.pdf	Walloon	null	“Shall we put the heart in now?” A comparative analysis between creature features and their single reel abridgements	null	2,021	cc-by	22,933	AUTHOR'S DECLARATION FOR ELECTRONIC SUBMISSION OF A THESIS I hereby declare that I am the sole author of this thesis. This is a true copy of the thesis, including any required final revisions, as accepted by my examiners. I authorize Ryerson University to lend this thesis to other institutions or individuals for the purpose of scholarly research I authorize Ryerson University to lend this thesis to other institutions or individuals for the purpose of scholarly research I further authorize Ryerson University to reproduce this thesis by photocopying or by other means, in total or in part, at the request of other institutions or individuals for the purpose of scholarly research. I understand that my thesis may be made electronically available to the public. ii Abstract Single reel abridgements of commercial feature films are entering moving image archives because home movie collections that contain them are slowly increasing in archival representation. The abridged commercial films occupy a liminal space in between sustained preservation efforts that focus on studio films and the current interest paid to preserving home movies. As a result, the abridged films are being neglected. The films’ liminal status stems from a dearth of information regarding their relationship to the original films and a clear definition of what they are narratively and aesthetically. After analyzing fourteen abridged horror and science fiction films found in the Ryerson Moving Image collection and comparing them to their original counterparts this project finds that the abridged films are heavily altered in terms of narrative, characters, and causality, and should be treated as individual objects instead of derivative works, thus absolving their liminal status. iii Acknowledgements First and foremost I would like to thank my parents for supporting me throughout the completion of my graduate degree. It is safe to say that from both an emotional and biological standpoint I would not be where I am today without them. I would like to thank my first reader, Bruce Elder, for guiding me through the process of writing this project. His wisdom, fairness, and his belief in my work and my ability as a writer has been a driving force in my success. I would also like to thank my second reader, Marta Braun, for helping me narrow my focus and teaching me how to complete a project at this level. I would be remiss to exclude the professors whose course material I have studied over the last two years, and whose lessons prepared me for writing this thesis project. Special thanks go to Gerda Cammaer and Izabella PruskaOldenhof for your wonderful lectures and eternal support. Finally, thank you to my classmates Stacey Turner, Alexandra Jokinen, Chelsea Keen. I could not have made it this far without you. iv iv I dedicate this project to my friends, the greatest people I have ever known. v v Table of Contents Chapter 1 Introduction Chapter 2 Literature Survey Chapter 3 Description of Materials Chapter 4 Historical Context Development of the Home Film Market Castle Films Chapter 5 Method of Analysis Chapter 6 Observations Length Frame Size Opening and Closing Credits Narrative Shot Reordering and Causality Conclusion of Analysis Chapter 7 Defining the Film Objects Chapter 8 Conclusion Appendix A Illustrations of Comparative Analysis Bibliography vi List of Appendices Comparative Analysis Illustrations Table 1.1 Table 1.1 Table 1.1 Descriptive Information of Castle Films Versions Descriptive Information of Castle Films Versions Descriptive Information of Castle Films Versions 1 Levno Plato and Aaron Meskin, Encyclopedia of Quality of Life and Well Being Researc , y p y g “Aesthetic Value,” New York: Springer Reference, 2014. 77. 1 Levno Plato and Aaron Meskin, Encyclopedia of Quality of Life and Well Being Research, 1st ed., s.v. “Aesthetic Value ” New York: Springer Reference 2014 77 Appendix A vii o and Aaron Meskin, Encyclopedia of Quality of Life and Well Being Research, 1st ed., s.v. , y p Q y g , , alue,” New York: Springer Reference, 2014. 77. Levno Plato and Aaron Meskin, Encyclopedia of Quality of Life and Well Being Research, Chapter One Introduction Chapter One Introduction Moving image archives, still chiefly concerned with studio productions, documentaries, and avantgarde cinema, are slowly beginning to increase the representation of home movies in their collections. Single reel 16mm and 8mm versions of feature length films, versions that were constructed by abridging feature films so they can be distributed on a single 400ft film reel, are entering moving image archives as part of home movie collections because of this increased, albeit comparatively minor, interest. The abridged commercial film occupies a liminal space in between sustained preservation efforts that focus on studio films and the current interest in preserving home movies. Their liminal status consequently affords them a much lower preservation priority even though their presence in home movie collections illustrates the viewing habits of families, provides historical information similar to that of the home films themselves, and, more broadly, illustrates nontheatrical distribution practices employed by Hollywood before the advent of VHS. The aesthetic value, or “the value that an object...possesses in virtue of its capacity to elicit pleasure (positive value) or displeasure (negative value)," also contributes to the liminality of abridged films. They are perceived as ugly, deformed versions of the originals that are 1 not deserving of attention in moving image archives. Abridged commercial films have yet to be studied in depth and there is no literature that closely examines these objects in comparison to the original films. The lack of academic inquiry has left the abridged films undefined. This project will seek to define abridged films so that their place and priority in archival collections can be accurately assessed by archivists and absolved of their liminality. Fourteen such films, created by Castle Films between 1957 and 1971 and found in the Ryerson Film Preservation collection, will be compared to their original films to assist defining 1 the abridged films more broadly. I contend that the abridged versions of feature length films must undergo a radical transformation to be distributed on a single 400ft film reel, and because of that transformation they can not be considered derivative works, but individual works created from previously produced images and sound. The abridged versions stand apart as separate entities with separate identities, and their fate in moving image archives should be determined on the individual merits of the object as opposed to their relation to feature length studio films. Chapter One Introduction The project is separated into eight chapters, and each chapter is dedicated to increasing the understanding of the abridged films, their method of creation, and how they should be defined. Chapter 3 begins the analysis by detailing the individual objects, including information regarding how the films are housed, the material nature of the objects, and their current condition. Chapter 4 places the film objects in a historical context as they relate to commercial films being purchased for use in the domestic space. The chapter is separated into two sections. The first examines how commercial films penetrated the home market at the end of the nineteenth century and into the twentieth century. The second section provides an overview of Castle Films that includes the founding of the company, its eventual purchase by Universal International, and their movement towards serving the niche market of commercial films for nontheatrical distribution. The comparison of the feature length films and the shortened versions required the use of digitization technologies and the use of digital editing software. Chapter 5 outlines the method for the comparison so researchers can compare other abridged films with the original material have a basis for doing so at a high technical level. The method is also designed so the comparison can be repeated by a researcher with little technical skill. Chapter 6 outlines the findings of the comparison in detail. The findings are separated into five sections that explore one specific area of comparative difference with the assistance of examples from the fourteen object films. The five sections are length, frame size, opening and closing credits, narrative, and cause and effect 2 2 relationships. Chapter 6 ends with an overview of my conclusions based on the observations made while also offering insight into the guiding principle that was followed by editors and producers during condensation process. The film objects are referred to as excerpts on their film canisters, but based on the findings of the comparative analysis that is not an appropriately descriptive term. Chapter 7 reviews different terms that might be used to describe the film objects, such as digest, excerpt, and remix, and determines which is most appropriate for defining the film objects. Chapter One Introduction Chapter 8 concludes the project by summarizing the findings of the comparative analysis as well as offering avenues of possible study for future researchers in the areas of Castle Films’ business practices, other types of abridgements including those made for silent projectors, and the sound design of abridged films. relationships. Chapter 6 ends with an overview of my conclusions based on the observations made while also offering insight into the guiding principle that was followed by editors and producers during condensation process. The film objects are referred to as excerpts on their film canisters, but based on the findings of the comparative analysis that is not an appropriately descriptive term. Chapter 7 reviews different terms that might be used to describe the film objects, such as digest, excerpt, and remix, and determines which is most appropriate for defining the film objects. Chapter 8 concludes the project by summarizing the findings of the comparative analysis as well as offering avenues of possible study for future researchers in the areas of Castle Films’ business practices, other types of abridgements including those made for silent projectors, and the sound design of abridged films. 3 Chapter 2 Literature Survey My research, outside of the comparative analysis, has focused on three central areas, namely the history of Castle Films, how feature length films entered the domestic space before the creation of VHS, and defining the film objects as individual entities. The goal of researching Castle Films and how feature films penetrated into the home was to compile available sources, synthesize the material, and present capsule histories that can be used by future researchers. My research, outside of the comparative analysis, has focused on three central areas, namely the history of Castle Films, how feature length films entered the domestic space before the creation of VHS, and defining the film objects as individual entities. The goal of researching Castle Films and how feature films penetrated into the home was to compile available sources, synthesize the material, and present capsule histories that can be used by future researchers. Surveying possible ways to define the film objects led to a contemporary movement of shortening works of literature, a movement that was spearheaded by Reader's Digest. The literature referenced for defining the abridged film centres largely on Reader’s Digest, its business practices, and their cultural impact. The literature available on Castle Films is limited. Chapter One Introduction Even though the company released thousands of films to the home market a major academic study of the company has not been published. Information about the company and its history can be drawn from two main sources: Castle Films: A Hobbyist’s Guide by Scott MacGillivray and a PhD dissertation authored by Eric Hoyt, entitled “Hollywood Vault: The Business of Film Libraries, 19151960." MacGillivray’s guide catalogues every film that Castle Films released throughout its history, and each entry contains basic information such as year of release, production number, and pieces of additional information specific to each individual title. The text also contains a short history of the company that begins in 1914 and includes sections dedicated to Universal International’s takeover of the company, distribution, promotion, and the company’s rebranding as Universal 8 in 1977. The text is essential to my project because of the complete catalogue of released films. However, the company history does not provide source information or citations, meaning that none of what MacGillivray wrote can be independently verified or traced back to its source for scrutiny. Furthermore, at the 4 beginning of the catalogue MacGillivray states that his book “is intended not as an academic textbook, but as a handy resource for the curious film enthusiast or collector." Based on this 2 statement and the lack of source information and citations it is best to rely on this text as a basic point of reference for information on the specific films that are being investigated. In terms of providing a historical narrative of Castle Films it will be treated as subordinate to texts whose sources of information are catalogued and cited. Eric Hoyt’s 2012 Ph.D dissertation, which was written for his Doctor of Philosophy in CinemaTelevision: Critical Studies at the University of Southern California, starkly contrasts MacGillivray's style of historical writing. Hoyt’s dissertation is an examination of “how old movies became valuable and how the marketplace for film libraries emerged and evolved…." Hoyt 3 explores the creation of the Universal International owned subsidiary United World Films, and the purchase of Castle Films in 1948 by United World. Hoyt uses the Castle Films example to 4 discuss the broader practice of creating and screening 16mm abridged versions of popular films and how that practice benefited film studios in the preVHS era. 2 Scott MacGillivray, Castle Films: A Hobbyist's Guide, New York: IUniverse, 2004. 23. y y 3 Eric Hoyt, Hollywood Vault: Film Libraries Before Home Video, Ann Arbor, MI: Proquest LLC, 2012, xxi. 4 Hoyt, Hollywood Vault, 2012. 245. y y 3 Eric Hoyt, Hollywood Vault: Film Libraries Before Home Video, Ann Arbor, MI: Proquest LLC, 2012, xxi. 4 H t H ll d V lt 2012 245 y y 4 Hoyt, Hollywood Vault, 2012. 245. y y 3 Eric Hoyt, Hollywood Vault: Film Libraries Before Home Video, Ann A y y 4 Hoyt, Hollywood Vault, 2012. 245. Chapter One Introduction Overall this is the most substantive academic study completed on Castle Films and it will figure largely in my project as a way of contextualizing the addition of abridged versions of Universal science fiction and horror films to the Castle Films catalogue. Beyond the work of Hoyt and MacGillivray, Castle Films has been left on the periphery of academic writing. The company is cited in discussions of home movies and collections, but is not described beyond the fact that they were distributors. For example, two articles written by Dwight Swanson and Eric Schaefer, published in The Moving Image and Cinema Journal respectively, 5 examine the collection of 16mm pornographic film in the 1950s. Castle Films is named as a 5 distributor of professionally made films to the amateur market in each article, but the company is not discussed beyond its market association of being advertized in the same home movie magazines as the pornographic films. I will also focus on the historical conditions that allowed for Castle Films to penetrate into home markets in the pre VHS era and find its niche as distributors of abridged commercial films to supplement the lack of academic writing on Castle Films as a subject. The first publication of note in the area of pre VHS distribution is Anthony Slide’s Before Video: A History of the NonTheatrical Film. Slide’s text chronologically organizes the history of industrial, educational, and travel film, as well as commercial films made for the home. He begins at the genesis of the film medium and ends with the release of VHS and Betamax systems in the mid to late 1970s. Slide refers to Castle Films over two pages and broadly speaks of the company’s creation and the kinds of films they distributed during the 1940s. However, Slide offers 6 no details beyond what MacGillivray covered in his more substantial history of the company. At The first publication of note in the area of pre VHS distribution is Anthony Slide’s Before Video: A History of the NonTheatrical Film. Slide’s text chronologically organizes the history of industrial, educational, and travel film, as well as commercial films made for the home. He begins at the genesis of the film medium and ends with the release of VHS and Betamax systems in the mid to late 1970s. g p ( ) 6 Anthony Slide, Before Video: A History of the NonTheatrical Film, Westport, CT: Greenwood Press, 1992. 99. Dwight Swanson, "Home Viewing: Pornography and Amateur Film Collections, A Case Stu g , g g p y , Moving Image 5, no. 2 (2005): 13640. & Eric Schaefer, "Gauging A Revolution: 16mm Film Of The Pornographic Feature" in Cinema Journal 41, no. 3 (2002): 326. 5 Dwight Swanson, "Home Viewing: Pornography and Amateur Film Collections, A Case Study" in The Moving Image 5, no. 2 (2005): 13640. & Eric Schaefer, "Gauging A Revolution: 16mm Film And The Rise Of The Pornographic Feature" in Cinema Journal 41, no. 3 (2002): 326. Chapter One Introduction Slide refers to Castle Films over two pages and broadly speaks of the company’s creation and the kinds of films they distributed during the 1940s. However, Slide offers 6 no details beyond what MacGillivray covered in his more substantial history of the company. At most the history that Slide provides can be used to verify elements contained in MacGillivray’s history of the company, which is important given the previously stated issues with MacGillivray’s text. Alan Kattelle’s Home Movies: A History of the American Industry, 18971979, published in 2000 by Transition Publishing, chronologically charts the history of home movies by examining the medium’s origins, the technological advances that brought motion pictures into the home, companies, such as Kodak and Bell & Howell, who aided in the development of the medium, and 6 further technological advances such as colour and the invention and dissemination of video technologies in the 1970s. Kattelle’s text will be chiefly used to address the broader history of home films and their penetration into the domestic space. Kattelle, a retired engineer who has previously been published in the Journal of Film and Video and Film History, focuses specifically on the American market while looking to Europe at appropriate moments, such as discussing the technological origins of projecting films in the home, to create a thorough history of films in the domestic space. As with Swanson and Schaefer, Castle Films is referred to only in passing. Kattelle describes the company as a “wellknown distributor of professional films for the amatuer market…” and doesn’t elaborate any further. The text focuses on amateur home movies from a 7 technological, sociological, and business perspective. Castle Films, as a company who sold professional films to the amateur market, falls outside of the text’s purview. Clare Watson’s article “Babies, Kids, Cartoons and Comedies: Children and Pathéscope’s 9.5mm Home Cinema in Britain," found in Movies on Home Ground: Explorations in Amateur Cinema, discusses the technological history of the 9.5mm film format by the Pathé Frères company for the home market. Watson’s approach is similar to Kattelle, but instead focuses on professional films which were released for the fledgling format. Castle Films is not discussed in the article because the company did not release films for the 9.5mm market. 7 Alan Kattelle, Home Movies: A History of the American Industry, 18971979. Nashua: Transition Publishing, 2000. 229. 7 Alan Kattelle, Home Movies: A History of the American Industry, 18971979. Nashua: Transition P bli hi 2000 229 y y Publishing, 2000. 229. Chapter One Introduction However, focusing on professional films is uncommon in academic writing dedicated to the home film market, and due to that scarcity I have included Watson’s article even though Castle Films is not addressed. Swanson’s aforementioned text on pornography in home film collections points out that many professional films came into archives as part of larger amateur home movie collections, stating that the “commercial films that accompanied homemovie collections were either not accessioned or were largely ignored by regional film archives, condemned to the fringes of 7 Alan Kattelle, Home Movies: A History of the American Industry, 18971979. Nashua: Transition Publishing, 2000. 229. 7 collections." Additionally, Slide wrote a brief section in his text that examines condensed 8 commercial films. Slide does not offer a concrete definition of what a condensed film is or how it functions in relation to the original film that is was adapted from, but he does acknowledge that condensed films were apart of Castle Films’ distribution strategy. Slide’s exclusion of a detailed 9 examination and definition of condensed films parallels the general trend of discussing aspects of nontheatrical home film distribution while ignoring how condensed films were constructed. What is missing from the current academic canon is a dissection of abridged films, both in terms of how they were made and how they differed from their original. Finally, in attempting to define the abridged films, I discovered that Reader’s Digest, at approximately the same time as Castle Films creation, began condensing articles from popular magazines and publishing them in a single monthly edition. I researched their methods of condensation in order to determine if “digest” more accurately describes the film objects as opposed to “excerpts," the term that Castle Films used to describe them. The relevant published work on Reader’s Digest is confined to historical overviews, a PhD dissertation, and the critical social observations of Ariel Dorfman. Of Lasting Interest The Story of the Reader’s Digest by James Playsted Wood, originally published in 1958 and then again in 1967 in a new edition, is the first published historical overview of the company. Swanson, "Home Viewing" in The Moving Image, 136. Slide, Before Video, 101. 10 James Playsted Wood, Of Lasting Interest: The Story of the Reader's Digest, 2nd ed, Garden City, N.Y.: Doubleday, 1967. g g g Slide, Before Video, 101. , , 10 James Playsted Wood, Of Lasting Interest: The Story of the Reader's Digest, 2nd ed, Garden City, N Y : Doubleday 1967 y g y N.Y.: Doubleday, 1967. Swanson "Home Viewing" in The Moving Image 136 , , 10 James Playsted Wood, Of Lasting Interest: The Story of the Reader's D g g g re Video, 101. , g g g , 9 Slide, Before Video, 101. , pp , y, 12 Ariel Dorfman, The Empire's Old Clothes: What the Lone Ranger, Babar, and Other Innocent Heroes Do to Our Minds, 2nd ed, Duke University Press, 2010. 126. Shawny Anderson, Condensed Hegemony: A Cultural/ideological Critique of Reader s Di 19801992, Ed. Edward Schiappa, Purdue University, 1994. vii. , pp , y, 12 Ariel Dorfman, The Empire's Old Clothes: What the Lone Ranger, Babar, and Other Innocent Heroes Do to Our Minds 2nd ed Duke University Press 2010 126 11 Shawny Anderson, Condensed Hegemony: A Cultural/ideological Critique of "Reader’s Digest," 1980 1992 Ed Ed d S hi P d U i it 1994 ii 1 Shawny Anderson, Condensed Hegemony: A Cultural/ideological Critique of "Reader’s Di , p g , , to Our Minds, 2nd ed, Duke University Press, 2010. 126. , pp , y, 2 Ariel Dorfman, The Empire's Old Clothes: What the Lone Ranger, Babar, and Other Innoc , g y g q g , Edward Schiappa, Purdue University, 1994. vii. p to Our Minds, 2nd ed, Duke University Press, 2010. 126. 13 Dorfman, The Empire’s Old Clothes, 2010. 131. Chapter One Introduction The text employs language akin to a 10 promotional piece for Reader’s Digest as opposed to an objective overview of the company, but it does contain within it three chapters of interest: “The First Condensed Book," “Cutting and Condensing," and “The Reader’s Digest Condensed Book Club." Each chapter examines the editorial process required to condense both articles and books and thus provides insight into how the finished product was made. I was initially hesitant to include Wood’s insights due to the biased 8 style of prose he uses and the amount of time that elapsed from the text’s initial publication to the present. However, the questions that Wood’s text raises in regards to respecting the work of the original author provides an excellent avenue for discussing the cultural place of abridged films, and was therefore included to aid in defining the film objects. Shawny Anderson, in her PhD dissertation entitled “Condensed Hegemony: A Cultural/Ideological Analysis of Reader’s Digest, 19801992," “analyzes the ideology of Reader’s Digest during the Reagan/Bush era.” Her focus on ideology as it relates to the political 11 landscape of the United States during the rise of neoconservatism does not directly relate to my own work, but she includes an in depth history of the company that balances the biased writing of Wood with a critical viewpoint. Ariel Dorfman’s book The Empire’s Old Clothes: What the Lone Ranger, Babar, and Other Innocent Heroes Do to Our Minds, first published in 1985 and then revised and reprinted in 2010, also takes a critical stance against Reader’s Digest. Chapter One Introduction While at times arrogant, Dorfman outlines his concerns with cultural domination through the infantilization of “mass market adult literature," specifically taking direct aim at Reader’s Digest in the chapter “The Infantilization of the Adult Reader." Dorfman argues that Reader’s Digest plants articles in other publications for their own use to appear worldly and well researched, structures their abridgements in such a way so that there is always a problem to be solved by an “average man," and reduces what is faraway and famous “to its most comprehensible, immediate, noto say vulgar, form." The actions of Reader’s Digest, according to Dorfman, reinforce capitalistic notions of 12 selfimprovement and upward social mobility because “North American society requires that each person believe that the world offers unlimited opportunities, and that everyone is equal to the task 9 of conquering that magical horizon." I intend to use Dorfman’s assessment of Reader’s Digest 13 Magazine to critique works of literature that were selected for condensation and the process of condensation that Wood claimed did not essentially alter the original work. The results of my critique will be applied to the fourteen abridged films to better understand their relationship to condensed works of literature, but also their role in infantilizing adult viewers. 10 Chapter 3 Description of Materials The fourteen film objects I will be examining were produced by Castle Films and purchased by the Ryerson Library in approximately 1975. The fourteen films came from Castle Films’ “Science Fiction and Horror” series, and the titles include It Came From Outer Space (1957), The Creature from the Lagoon (1957), Bride of Frankenstein (1960), The Mummy (1962), Dracula (1963), Frankenstein Meets the Wolfman (1963), Son of Frankenstein (1965), Revenge of the Creature (1965), House of Frankenstein (1967), The Mummy’s Ghost (1968), The Incredible Shrinking Man (1969), ManMade Monster (1969), The Wolf Man (1971), and Frankenstein (1971). Please note that the included dates refer to the release of the Castle Films 14 versions and not the release of the corresponding Universal Studios feature films. The objects are English language 16mm triacetate Eastman Kodak safety film release prints with synced optical sound. Each film is presented in the Academy Standard aspect ratio of 1.37:1. The edge code on each of the films is a plus sign and a circle, indicating that the film stock was manufactured in 1974. 14 It was common practice for Castle Films to give its abridged versions new names. In the case of the fourteen objects examined for this project only one, The Creature from the Lagoon, had its name altered. For the sake of clarity when I refer to the feature length version of the film I will use the full title, The Creature from the Black Lagoon, and when referring to the shortened version I will refer to it by its Castle Films moniker. Chapter One Introduction Each film is contained in a grey plastic film canister with exterior labels that contain the name of the film, its running time, the year the film was produced, and indicators that the films are sound with black and white images. The library call number associated with each object is still present on the front of each canister. On the inside of each canister is a square sticker that contains four separate series of numbers and a barcode. Ryerson University librarians were not able to identify the purpose of these numbers. Ophelia Cheung, the Audio Visual and Reserve Services Librarian for the Ryerson Library, commented that the numbers were likely applied before Ryerson became a university. At that time items would be sent to Biblocentre, a 11 consortium of colleges that centralized the cataloguing of library objects. Biblocentre ceased operations in 2009, and no trace of records for the objects has been located. The grey canisters are not vented, therefore they do not meet current archival standards. The films are wound on plastic projection reels, and nine of the fourteen films have a sticker attached to the reel indicating when they were last cleaned and inspected. The dates range from October 17th, 1991 to March 8th, 1999. Eleven of the fourteen films have identifying inscriptions either at the head, tail, or at both positions just after the leader. The inscriptions vary in detail, but most contain the title of the film, the name of the distributor, and a production number. The distributor names vary, but all appear to be linked with Universal Studios. For a detailed list of the film titles, their corresponding inscriptions and other specific information pertaining to each film object please see Table 1.1. Table 1.1 Table 1.1 Name Release Number Release Year Length (s) Colour Frame Production Note (on film) Production # (on film) It Came from Outer Space 1007 1957 500 BW 1.37:1 Picture It Came From Outerspace C6976 The Creature from the Lagoon 1008 1957 493 BW 1.37:1 N/A N/A Bride of Frankenstein 1013 1960 492 BW 1.37:1 UWF D1431 Bride of Frankenstein Long Version D1431 The Mummy 1021 1962 482 BW 1.37:1 UNIV EDUC The Mummy With Prologue Subtitles Sound Version D4924 Frankenstein Meets the Wolfman 1022 1963 493 BW 1.37:1 United World Frankenstein Meets the Wolfman GFL 90294 Dracula 1023 1963 494 BW 1.37:1 Dracula Sound Ver Picture N/A Son of Frankenstein 1033 1965 501 BW 1.37:1 UNIV EDUC Son of Frankenstein Sound Vers Recut New Vers 61771 D7460 12 12 Revenge of the Creature 1037 1965 490 BW 1.37:1 UW Revenge of the Creature Sound Vers N/A House of Frankenstein 1043 1967 488 BW 1.37:1 N/A N/A The Mummy's Ghost 1049 1968 492 BW 1.37:1 UWF Mummy's Ghost Sound Version N/A The Incredible Shrinking Man 1053 1969 490 BW 1.37:1 N/A N/A Man Made Monster 1054 1969 494 BW 1.37:1 Universal Education + Visual Arts D9647 The Wolf Man 1060 1971 482 BW 1.37:1 UNIV EDUC The Wolfman Sound Vers E2967 Frankenstein 1061 1971 508 BW 1.37:1 Universal ED. E(C?) 3987 Frankenstein E3987 Evidence of inspection and repair can be found on each of the objects. There is new green and red leader at the head and tail of the films, respectively. A majority of the films had sprocket holes repaired, with Bride of Frankenstein requiring nearly the entire first half of the film to be repaired in this manner. There are also multiple instances throughout the films of clear splicing tape being placed on the image. However, these instances did not always correspond with the repair of broken splices or torn film and, in a few cases, hair was visible underneath the tape splices. The films were inspected and are in relatively good condition considering they had previously been used as lending copies at the Ryerson Library. Each film shows light to medium base and emulsion scratching throughout, a fading of the luminosity of the image, and mild warping of the film strip. Table 1.1 The audio fluctuates during projection due to the warping, but individual aspects of the audio, such as dialogue, music, and sound effects, are discernible. The instances of severe damage are minimal, and are limited to two of the prints. Son of Frankenstein has sprocket holes punctured through the image in the middle of each frame throughout the opening credits, 13 and the sprocket side of Frankenstein has a prominent wave which could make it difficult to project. There is no evidence of vinegar syndrome in any of the films. 14 14 Chapter 4 Historical Context Chapter 4 Historical Context 15 Kattelle, Home Movies: 18971979, 2000. 52. 17 Ibid, 54. 15 Kattelle, Home Movies: 18971979, 2000. 52. 16 Ibid , 18 Ibid. , , 16 Ibid. 16 Ibid. 17 Ibid 54 16 Ibid. 17 17 Ibid, 54. , 18 Ibid. 17 Ibid, 54. 16 Ibid. , 18 Ibid. id. id, 54. Development of the Home Film Market To understand how abridged feature length films became a part of home film collections it is necessary to outline how film first penetrated the domestic space. According to Alan Kattelle, “almost as soon as motion pictures became commercially successful in the closing years of the nineteenth century, entrepreneurs began looking for ways to bring this new amusement into the home." Although safety film was not made available by Eastman Kodak and Pathe until 1911, 15 there were initial efforts in Europe to bring moving pictures into the home as early as 1897. Middle to upper class families began to install madeforthehome magic lanterns in domesticated spaces at the end of the nineteenth century, and the first moving image projectors made for the home relied on the magic lantern’s light source for projection. It is worth noting that these machines were being installed in the home even though they were designed to project highly flammable nitrate film. 16 In the United States the movement towards the exhibition of moving images in the home was also swift, but the immediate need to install professional equipment in the home was not as advanced as in Europe. The first moving image machines to be marketed for home use, the Parlor Kinetoscope and the Vitak 11mm film projector, released in 1897 and 1902 respectively, were marketed as toys. The Parlor Kinetoscope was a smaller version of the Edison Kinetoscope, the 17 only difference between the two, besides the diminutive dimensions of the Parlor version, being the Parlor Kinetoscope used paper film rolls instead of celluloid. Rector, the developer of the 18 15 Vitak 11mm projector, continued to develop projectors for the home market. In 1904 they released the Ikonograph 17.5mm projector along with a series of short subjects that could be purchased separately. A section of the the Sears, Roebuck & Co catalogue from 1905 featured an 19 advertisement for what they called their “Premier Moving Picture Outfit." The advertisement enticed customers with statements like “You can make lots of money by giving evening entertainments right in your own neighborhood.” The Premier Projector, like the Ikonograph, 20 projected 17.5mm film that was sold separately at the price of approximately ninety cents per ten feet of film. Ibid. 22 Haidee Wasson, "Electric Homes! Automatic Movies! Efficient Entertainment!: 16mm and Cinema's Domestication in the 1920s," Cinema Journal 48, no. 4 (2009): 121. 23 Haidee Wasson, Electric Homes! Automatic Movies! Efficient Entertainment!: 16mm and Cinema s Domestication in the 1920s," Cinema Journal 48, no. 4 (2009): 121. , , ( ) 23 Haidee Wasson, “Electric Homes!," 2009. 6. , 23 Haidee Wasson, “Electric Homes!," 2009. 6. Haidee Wasson, Electric Homes! Automatic Movies! Efficient Entertainment!: 16mm Domestication in the 1920s," Cinema Journal 48, no. 4 (2009): 121. Ibid. 2 Haidee Wasson, "Electric Homes! Automatic Movies! Efficient Entertainment!: 16mm and , , ( ) on, “Electric Homes!," 2009. 6. , 21 Ibid. Ibid, 56. 21 Ibid Development of the Home Film Market The Premier Projector, while still using a film format that did not allow for 21 professional projection of 35mm film, marked a difference in marketing techniques in North America; it was being marketed as an investment as opposed to a toy, indicating that public perception was beginning to shift in regards to the role of moving images in the home. The perceptual shift was partially influenced by the vagrant nature of early film projection. Haidee Wasson, in the journal article “Electric Homes! Automatic Movies! Efficient Entertainment!: 16mm and Cinema’s Domestication in the 1920s," argues that films were able to make their way into the home due to the initially unstable nature of cinema exhibition during the early years of film. Specifically, she states that it was not understood in the first two decades of 22 cinema that film needed to be watched in a movie palace or a theatre because, at that time, they did not exist. Films could be exhibited in any space such as schools, churches, basements, and 23 store fronts. There was no tradition of cinema being viewed in a theatre. When 16mm film was released as an amateur format in the 1920s, and along with it the 16mm projectors needed to view amateur works, it was not odd to view these works in the home, the place where they were 16 most often made. Portable film projectors, along with phonographs and radios, were marketed in 24 a way that placed them as an essential piece of the family circle, a circle that existed squarely in the domestic space of the home. 25 For the home market to be secured as a viable space for motion picture exhibition, a safety film that was not nitrate based was required. In 1912 French film company Pathé Frerès “announced the first complete system of safety film, camera, and projector for the amateur." The 26 28mm film stock gauge utilized a celluloseacetate base that tended to melt if overheated as opposed to ignite like nitrate based films were prone to do. The projector, the Pathé Kok, also came equipped with a light source that could be powered by an onboard generator. The generator drew power from the same cranking motion that ran the film through the gate of the projector. Ibid, 52. 29 Dwight Swanson, “Inventing Amateur Film: Marion Norris Gleason, Eastman Kodak and the Rochester Scene, 19211932.” Film History 15, no. 2, 2003. 127. Ibid, 4. 26 Kattelle, Home Movies: 18971979, 2000. 61. 27 Ibid 7 Ibid. 8 Ibid, 52. 24 Ibid, 6. 25 Development of the Home Film Market The films released for the Pathé Kok Projector were “abbreviated versions of Pathé’s commercial films, reduction printed to the smaller gauge," indicating that even in the infant stages of home film distribution there was a market for commercially made films. 27 The movement of film into the home grew steadily, but, as Kattelle noted when writing about the first home projectors in Europe, home projection and home filmmaking were expensive hobbies that only high class families could afford. The Pathé Kok system was not able to lower 28 the cost of production and projection to a significantly low enough threshold. In 1919 George Eastman began research into an amateur film gauge but “prudently demanded that only safety film be provided for inhome use…." Advocating for safety film meant that 17.5mm film, a gauge 29 which would have produced an image exactly half the size of 35mm, was not pursued due to fears that other companies might have tried to produce a type of 35mm film that could be cut in half for 17 amateurs and thus bring nitrate based film stock back into the home. Scientists at Eastman 30 Kodak were also aware of the reversal development process that had been created by Rodolfo Namias, an Italian photochemist, in 1900. The reversal process cut the cost of developing film in half because it removed the need to have a camera negative developed and a positive print produced from that negative. With the reversal process, the film that ran through the camera, once developed, would produce a positive image that could be projected. Combining the smaller gauge with the reversal process would make filmmaking more affordable for the amateur.” The 16mm 31 film gauge was presented to the public in 1923 along with Kodak’s CineKodak camera, the first 16mm film camera. Following the release of the CineKodak and the 16mm film gauge, Kodak’s 32 own camera and projector systems, along with the 16mm camera and projector systems of Bell & Howell, slowly became the American standard for amateur film gauges. Other amateur gauges, like the above mentioned 28mm gauge, began to disappear due to the popularity of 16mm film. 33 16mm filmmaking was still “expensive and therefore accessible primarily to the upper classes and the more enthusiastic hobbyists," however the lower price allowed the format to proliferate beyond what previous gauges had accomplished. 36 Mark Clark, “For Monster Fans." Ibid. 35 Wasson, “Electric Homes!," 2009. 9. 30Swanson, “Inventing Amateur Film," 2003. 127. 31 31 Ibid. 32 Ibid, 128. 34 Ibid. Swanson, Inventing Amateur Film, 2003. 127. 31 Ibid. 33 Ibid. Ibid. 35 Wasson, “Electric Homes!," 2009. 9. 36 34 Ibid. 31 Ibid. 32 Ibid, 128. 34 Ibid. 35 Wasson, “Electric Homes!," 2009. 9. 30Swanson, “Inventing Amateur Film," 2003. 127. 31 Ibid. 32 Ibid 128 Development of the Home Film Market As the 16mm market continued to 34 grow, an increasing number of slapstick comedies, animation, travel, sports and nature films were made available for people to view on their personal 16mm projectors. Castle Films was founded 35 at the beginning of the 16mm period of growth, and it was one of many companies distributing 16mm films for home viewing. The pace of growth of the 16mm market in the late 1920s was so 36 great that in 1927 Variety declared that “all Hollywood studios would either reduce their theatrical 18 35mm prints to the 16mm standard or would make titles exclusively for what was variably termed “the home” or “amateur” market. 37 Variety's prediction did not come to pass. The distribution of 16mm Hollywood films was strictly controlled by firms like Films Inc., a company that acted as a gatekeeper against distribution to home viewers. In fact, Films Inc. was trusted with the 16mm films of the major 38 Hollywood studios because it would refuse any rental request from a home viewer. However, the initial penetration of 16mm into the homes of amateur filmmakers in the 1920s allowed some companies, such as Castle Films, to survive well into the 40s, 50s, and beyond. The place of home films in the domestic space has been taken up by two other scholars, Dwight Swanson and Eric Schaefer, each of whom seek to contextualize the rise of pornography as a collectable item that existed in home film collections. Swanson specifically argues that the curator of the home film collection was “Dad," the evidence being examples of advertisements that were included in amateur filmmaking magazines such as Home Movies. Swanson points to an 39 example from an issue of Home Movies published in the 1950s that includes an advertisement for Castle Films near the front while advertisements for pornographic films such as Goldilocks Goes Glamorous and Sweethearts of Burlesque were found in the back. The popularity of 40 The place of home films in the domestic space has been taken up by two other scholars, Dwight Swanson and Eric Schaefer, each of whom seek to contextualize the rise of pornography as a collectable item that existed in home film collections. Swanson specifically argues that the curator of the home film collection was “Dad," the evidence being examples of advertisements that were included in amateur filmmaking magazines such as Home Movies. y , y , 39 Swanson, "Home Viewing" in The Moving Image, 136. 37 Wasson, “Electric Homes!," 2009. 10. , , 38 Hoyt, Hollywood Vault, 2012. 245. g g g 40 Ibid. 38 Hoyt, Hollywood Vault, 2012. 245. 41 Ibid, 138. 41 Ibid, 138. 40 Ibid. , , 43 MacGillivray, Castle Films, 1. y 44 Ibid, 2. 42 Slide, Before Video, 99. 45 Ibid. 44 Ibid, 2. Development of the Home Film Market Swanson points to an 39 example from an issue of Home Movies published in the 1950s that includes an advertisement for Castle Films near the front while advertisements for pornographic films such as Goldilocks Goes Glamorous and Sweethearts of Burlesque were found in the back. The popularity of 40 pornographic films relative to abridged commercial films in home collections is difficult to gauge. Swanson notes that the collection he is using as a case study is “split right down the middle," with sixteen reels of home films, fifteen reels of pronographic films, and one commerical film. 41 However, this anecdotal example should not be seen as representative of all home film collections due to the small sample size. However, this anecdotal example should not be seen as representative of all home film collections due to the small sample size. 19 Neither article focuses on the work of Castle Films, but a general enthusiasm for collecting abridged commercial films can be discerned from the time period that the articles focus on, namely the 1950s. The fact that both point to Castle Films as an example of a company that penetrated the home reinforces the notion that the abridged films that Castle Films were releasing, as some of the 16mm films that “Dad” would collect, were part of a much larger and more significant culture of home film distribution than previously thought. 43 MacGillivray, Castle Films, 1. 44 , 45 Ibid. 46 Ibid. 44 Ibid, 2. , 45 Ibid. 46 Ibid. 42 Slide, Before Video, 99. Castle Films Castle Films Eugene Castle, the founder of Castle Films, began his career in the San Franciscan film industry in 1914. He was employed as a “freelance cameraman photographing local events” who 42 was “trying to break into the film industry via newsreels." He was later employed by Pathé News 43 and then Fox News where he held an “editorial position … at the newsreels’ west coast bureau." 44 According to Scott MacGillivray, Eugene Castle believed that “movies had potential outside of conventional theatrical distribution," and by following this belief Castle began distributing educational and industrial films in 1918. 45 In 1924 Castle founded Castle Films with an investment of 10 000 dollars, or approximately 146 000 dollars adjusted for inflation to 2015. The company’s initial focus was creating films that could be used by business professionals for presentations. Castle himself thought of films as “business tools." Castle Films was distributing both 35mm film and 16mm 46 films, and the first large scale expansion of the company came in 1933 when Kodak released , 45 Ibid. 46 Ibid. 20 16mm film with sound, a format that Castle Films adopted immediately as a distribution format. Castle, having spent years in the newsreel industry and with his experience in “commercial and advertising fields," pushed the company towards distributing films made specifically for the growing home film market. Castle realized that more projectors were being sold to home users 47 and that, in turn, would create a higher demand from consumers to own films. Castle Films 48 eventually became a nontheatrical producer and distributor of short excerpt films that would be sold to churches, schools, libraries, and home viewers. Their niche, as described by Eric Hoyt, 49 was in “outright sales to home viewers and collectors, a method of distribution (sales, rather than rentals) to a market of nontheatrical endusers (home viewers)." 50 In 1937 Castle’s entry into the home market began with three newsreel shorts: Hindenburg Explodes!, England’s Coronation, and The Life of EdwardBritain's ExKing. 51 According to MacGillivray, Castle “edited the films himself, giving his wares a personal, professional touch." The films Castle produced for the home market, which quickly ballooned to 52 twelve by the end of 1937, were offered in 16mm and 8mm film formats in both sound and silent versions so that the films could be projected by smaller toy projectors as well their 16mm counterparts . , 51 MacGillivray, Castle Films, 2004. 3. 49 Hoyt, Hollywood Vault, 2012. 245. 48 Slide, Before Video, 1992. 99. 50 Ibid, 245. y 52 Ibid. y , 50 Ibid, 245. 47 Ibid. 58 MacGillivray, Castle Films, 2004. 5. 56 Slide, Before Video, 1992. 100. 57 Ibid. 61 Hoyt, Hollywood Vault, 2012. 245. 57 Ibid. MacGillivray, Castle Films, 2004. 5. 59 Ibid. 60 60 Ibid. 62 Ibid, 244. Castle Films In 1937 the company settled on the policy of releasing single reel films only. Local 53 movie theatres, at the time, would show “newsreels, sports films, travelogues, and cartoons” that were “generally no more than 10 minutes in length each." MacGillivray states that the one reel 54 format “became an industry standard," a standard that Castle Films deviated from only once when they experimented with “slightly longer films in the 1950s." 55 53 Ibid, 4. 54 Ibid. 55 Ibid. 21 In the 1940s Castle Films expanded once again to include “scenics, actuality shorts, and Terrytoon cartoons," these films comprising “the bulk of Castle’s offerings." The Second World 56 War, on the strength of Castle Films’ series of short subjects that focused on specific battles, also proved to be a successful period for the company. Exuberant titles and promises of releasing 57 battle footage as soon as it could be compiled and edited kept the public informed about events overseas. Advertisements for the war shorts even encouraged people to collect all of the films so 58 that customers could own a “complete record of the war." The company released thirty seven 59 films in their war series by the time the Second World War had come to an end. 60 In 1947 United World Films, a wholly owned subsidiary of Universal International, purchased a 75 percent stake in Castle Films for 2.25 million dollars, or 28.2 million dollars adjusted for inflations to 2015. United World Films had been created by Universal International in 61 1946 in order to meet the growing demand of the 16mm market. WWII, Hoyt contests, caused 62 an upswing in production of 16mm films resulting in “mainstream educators and the public [growing] more familiar and accepting of nontheatrical films…." The total revenue that Universal 63 International was accumulating through 16mm made up approximately 1.5% of their gross revenue, but the market for 16mm films had grown steadily leading up to 1946. United World 64 Films was created for Universal International to further monetize their film holdings and expand their film library while becoming a leader in the distribution of 16mm film. 65 63 Ibid, 243. 64 Ibid, 244. 65 Ibid. 22 United World Films first acquisition was the Bell & Howell FilmoSound library that featured a large collection of 16mm shorts. Ibid, 251. 71 MacGillivray, Castle Films, 2004. 186. 69 Ibid, 248. 70 Ibid, 251. , 68 Ibid, 246. 69 Ibid, 248. 67 Ibid, 245. Ibid, 251. 71 MacGillivray, Castle Films, 2004. 186. Ibid, 245. 68 Ibid, 246. 70 Ibid, 251. 71 MacGillivray Castle Films 2004 186 70 Ibid, 251. 71 MacGillivray Castle Films 2004 186 72 Mark Clark, "For Monster Fans of the 1960s and '70s, Castle Films Abridgements Were The Best Things on Earth!" Monsters from the Vault, Accessed December 1, 2014. http://www.monstersfromthevault.com/LittleGiants.html. 73 MacGillivray, Castle Films, 2004. 186. Castle Films At the same time, United World purchased Bell & Howell’s 16mm distribution network, which gave them a strong foothold in the 16mm market. 66 Castle Films was purchased soon after. The deal made United World/Universal the leaders in nontheatrical film distribution, but it also put Universal two million dollars into debt. For Castle 67 Films to be monetized effectively United World adapted the practice of condensing feature length films from Universal’s back catalogue for the home market. The sales of 16mm film through 68 United World did not perform as expected, and the market of growth that Universal originally foresaw had become stale. United World’s financial issues were exacerbated because the postwar market for 16mm films centred around educational and industrial films, and these were two areas that United World’s library did not cover. In an attempt to change the fortunes of the 69 company, president Matty Fox started examining contracts of films that were already in Universal’s library to determine if they could be sold to television stations. 70 Castle Films continued to produce abridged films for the niche market it served, and ten years after its initial purchase by Universal it began to release a selection of the science fiction and horror films found in the Universal library. In 1957 Castle Films released two of the most popular science fiction films of the 50s as abridgements: It Came From Outer Space (1953) and The Creature from the Black Lagoon (1954). Both films featured Richard Carlson in the starring role 71 and were directed by Jack Arnold. The success of these two films spurred Castle Films to release 28 more films as part of the science fiction and horror series that would see films from the 30s, 23 40s, 50s, 60s, and 70s condensed for release. Later films added to the series of excerpts 72 included the Jack Arnold directed Tarantula! (1955) and horror classics like Dracula (1931) and Frankenstein (1931). The series was so successful that after the introduction of the first two films no film from the series was taken off sale until Castle Films was rebranded as Universal 8 in 1977. 73 73 24 Chapter 5 Method of Analysis This section, which will detail how I compared the original films to the Castle Films versions, is designed as a guide for other researchers to use with other abridged films. 74 A lower file size allows Avid Media Composer, the program that I used to compare the films, to interact with the media at more efficient speeds. p , with the media at more efficient speeds. 74 A lower file size allows Avid Media Composer, the program that I used to compare the with the media at more efficient speeds. Castle Films To start, I needed to obtain digital copies of both versions of each film. I obtained the long form versions with ease because some of the more popular films Bride of Frankenstein and Dracula, for example were recently released for home viewing on DVD and BluRay. The more obscure films, such as Man Made Monster, I eventually found on compilation DVDs included with other horror films from the Universal Studio catalogue. I ripped the DVD image and audio files to a hard drive at the DVDs original resolution of 720x480 pixels using a program called Handbrake. I needed to encode the ripped files at an image size and bit rate that would maintain a high fidelity image while keeping the file size as low as possible. For that reason, I encoded the video with the x264 74 derivative of the H.264 codec at the same frame size of the original DVDs. I encoded the audio using the standard AAC audio encoder. Regardless of the technical efficiencies of x264 and AAC’s compression algorithms, the original low resolution images of the DVDs are less detailed than the 16mm films. After the digital versions of the feature films were properly stored and backed up on a secondary hard drive, I had to check each film to ensure that each file was identical to the film’s original release. Each of the DVD versions had been digitized and restored in order to be presented digitally, and it was possible that the films may have been altered during that process by a digitizer, technician, colourist, or producer. I cross checked technical aspects that included running time, which would indicate scenes being added or taken away from the film, and frame size. I verified the frame sizes because in the 1950s film studios began releasing films in a variety 74 A lower file size allows Avid Media Composer, the program that I used to compare the films, to interact with the media at more efficient speeds. 25 of widescreen formats (Cinemascope, Panavision, and Superscope, for example). It is possible that the original films that were released in the 1950s could have been altered in order to accommodate the now standard 16:9 aspect ratio used for home viewing on widescreen televisions. The Creature from the Black Lagoon, for example, was originally released in the 1.37:1 Academy Standard aspect ratio as a 3D film. 75 The Film Daily, "Dracula Review," Feb 15, 1931. www.mediahistoryproject.org (accessed May 7, 2015). 11 75 The Film Daily, "Dracula Review," Feb 15, 1931. www.mediahistoryproject.org (accessed May 7, 2015). 5 The Film Daily "Dracula Review " Feb 15 1931 www mediahistoryproject org (accessed 11. 75 The Film Daily, "Dracula Review," Feb 15, 1931. www.mediahistoryproject Castle Films The studio filmed Creature “widescreen safe", meaning that no important action took place in the upper and lower portions of the screen. The studio shot the film this way to ensure it could later be matted and presented in the 1.85:1 widescreen aspect ratio. The technicians who restored the version of The Creature from the Black Lagoon that I used for my comparison cropped out the upper and lower portions of the frame in order to present the film in 1.85:1 aspect ratio, a deviation from its original release. Castle Films released each of the abridged films that I am studying in 1.37:1 aspect ratio, and knowing that the original film was released in the same aspect ratio meant that I would not make a false observation in the analysis. Dracula’s running time is difficult to determine because multiple versions of it were made in 1931. The Film Daily review of Dracula lists a running time of 85 minutes, whereas other reviews state the film is 75 or 65 minutes in length. Two other versions of Dracula were released in 1931, 75 a Spanish language version that was filmed at night using the same sets as the English language film and a silent version of the English language film. It is possible that the variety of running times reported in print and online are attributable to the existence of three distinct versions of the film. In any case, when I compared Dracula to its Castle Films counterpart all of the shots used in the Castle Films version were matched back to their original source. No elements from other possible versions remained in the shorted version, so I am confident including the Dracula comparison in 26 this project. In all other cases I confirmed that the feature length films accurately represented their original presentations. I digitized the Castle Films versions using the 16mm release prints because no other reliable digital versions could be found. I digitized the films at a resolution of 1920 x 1080i pixels at 60 frames per second using the Sniper 16 HD film digitizer found at the Ryerson University film lab. The Sniper 16 HD software processed the footage to playback at 24 frames per second. The Sniper also records the visual information as one continuous video file. For the purposes of film preservation the Sniper 16 is not ideal, but for this project it was adequate. Castle Films Handbrake further processed the files, encoding them in the same x264 codec and MKV container as the feature length versions. I also decreased the frame size from 1920 x 1080i to 1280 x 720p. The smaller resolution decreases the size of the files, therefore freeing system resources and allowing the computer to run more efficiently when interacting with the files. At the same time, I kept the Castle Films versions at an HD resolution to provide enough visual detail to mitigate any difficulty in comparing the restored feature film versions to their faded and warped 16mm film counterparts. After I had completed digitizing all of the films I imported the feature and Castle Films versions into Avid Media Composer and converted the files to Avid’s MXF 1:1 format. This method created new files that Avid can read and operate with easily and ensures that any changes I make to the files in Avid as necessitated by my comparison will not affect the digital objects. Once imported, I paired the films together in bins (a file inside which film clips, audio, effects, etc can be stored and organized) that corresponded with each specific film. For example, I moved the feature and Castle Film version of Bride of Frankenstein to the same bin. I then placed both versions on a single timeline, the feature version on video track 1 (V1) and the Castle Films version on video track 2 (V2). I moved the Castle Films version along V2 until its first frame corresponded with a frame on V1. After finding it, I played both clips simultaneously until the images no longer 27 matched. I cut the Castle Films version at the frame where the difference occurred, and the shot was labeled numerically. Numbering the segments maintained the structural chronology of the original Castle Films version. The process was repeated until all of the segments in the Castle Films version were paired with their feature counterparts. If a shot from the Castle Films version came from a chronologically different section of the original film it was placed on video track 3 (V3) to visually separate it. This was done to help identify how cause and effect relationships, temporal construction, and the visual context for each shot were altered in the creation of the abridged version. matched. Castle Films I cut the Castle Films version at the frame where the difference occurred, and the shot was labeled numerically. Numbering the segments maintained the structural chronology of the original Castle Films version. The process was repeated until all of the segments in the Castle Films version were paired with their feature counterparts. If a shot from the Castle Films version came from a chronologically different section of the original film it was placed on video track 3 (V3) to visually separate it. This was done to help identify how cause and effect relationships, temporal construction, and the visual context for each shot were altered in the creation of the abridged version. Following the completion of the comparison, I created visual representations of the Avid timelines in Adobe Illustrator. The visual representations, which make up the entirety of Appendix A, demonstrate where the images used in the abridged versions came from in the original films and the degree to which the images were rearranged in order to best complete the narrative. This final step was not necessary for the purposes of the comparison, and the visualizations were included merely as a visual aide for how each abridged film was constructed. 28 28 Chapter 6 Observations Once I successfully cut the Castle Films versions into their component pieces and matched their corresponding shots, I compared both visual sources based on the content of the footage that was not used in the Castle Films version, how the absence of footage affected the causality of the original film, the extent that shots were reordered in the Castle Films versions, the effect that reordering shots had on the cause and effect relationships established in the original film, and if any new images were created for the abridged version. I observed five differences between the two versions after the comparison was complete: length, frame size, opening and closing credits, narrative, and the chronology of the component shots. There are analytical considerations that go beyond these boundaries. When analyzing moving images the construction of motion in frame and how that motion persists from shot to shot, the temporal arrangement of shots in relation to each other, and the facets of constructing visual imagery, such as tone, line, and shape, each offer different possible approaches for interpretation. Length Length The difference in length between the original film and the abridged version best illustrates the degree to which the original films were altered. The fourteen original films vary in length, the shortest being Man Made Monster (59 minutes) and the longest being Son of Frankenstein (99 minutes). The Castle Films versions range in length from 8 minutes and 2 seconds to 8 minutes and 20 seconds. The range indicates that the editors at Castle Films were required to cut the films down to a near uniform size. In the case of Man Made Monster, 51 minutes (86%) had to be excised from the original. In the case of Son of Frankenstein, 91 minutes (92%) were taken out. The amount that each film needed to be shortened directly affected the credits, narrative, and cause and effect relationships in the Castle Films versions. This point is necessary for understanding why the abridged versions differ to the extent that they do. Castle Films However, I have elected to limit the detailed aspect of my analysis to a narrative focus in order to demonstrate the high degree to which the original films were transformed during the condensation process. My aim in this thesis, as stated in the introduction, is to define the abridged films as separate and distinct entities, and I believe demonstrating that the abridged films are narratively separate from the originals will accomplish that goal. At the conclusion of this section I will briefly discuss the guiding principles that informed how images were chosen to be included in the abridged films, and in this particular case I do not believe narrative is the best means for doing so. Discussing the existence of guiding principles at Castle Films requires examining the abridgements in relation to each other as opposed to their original films. A narrative comparison between the fourteen abridged films revealed the existence 29 of a basic three act structure, but no indication for why certain images were used and others were not. Therefore, I believe that in this case it is more appropriate to examine visual elements of the abridged films in order to determine what guiding principles were followed in their creation. Length Frame Size I noted in Chapter 5 that each of the Castle Films releases was released in the Academy Standard 1.37:1 aspect ratio. Each of the original films, save one, was originally released in the same aspect ratio. The exception, The Incredible Shrinking Man, was originally released in 1957 during the transition period to widescreen viewing formats in North American theatres. Its original aspect ratio was 1.85:1, and it was altered to bring it in line with the rest of the abridged releases. 30 The reason for altering the frame size of The Incredible Shrinking Man, however, is curious. In the VHS era of home distribution it was common practice for films to be “pan and scanned," a process that altered films that were originally released in widescreen formats. The process involved cropping or moving the image so that all of the important narrative information could be displayed on a television with a 4:3 aspect ratio. There was little need for such an alteration in the preVHS era of home film viewership. Home projection could, with the appropriate lens on the projector, produce an image several feet high and wide, making the need to crop the image for the sake of saving space on the viewable surface irrelevant. The process that Castle Films used to alter the frame in order to give the image an aspect ratio of 1.37:1 is equally puzzling. In the VHS era the image was altered by deciding which part of the frame was the narrative focus and cropping out what remained. For example, in an over the shoulder shotreverseshot sequence where a character occupies the right side of the frame in medium close up while the other character’s head and shoulder are shown out of focus on the left hand side of the frame, the right side of the frame would become the main focus of the image on the VHS version and the left side would be cropped out; the reverse shot, which would show the other character’s face, would focus on the left hand side of the frame. However, The Incredible Shrinking Man is cropped using a crude method: instead of panning or adjusting each individual shot to best suit the narrative, approximately one fifth of the image on the right hand side of the frame is simply removed. 76 MacGillivray, 186. 76 MacGillivray, 186. Frame Size The removal of the right hand side of the frame is uniform throughout, and it leads to awkward instances where a character is cut in half by the frame or cut out of the frame entirely, even if that character speaks in the scene. The dearth of published material on the business and technical practices of Castle Films and on the ways it produced its abridged films make it difficult to understand why such a crude cropping method would be used. The Incredible Shrinking Man is the only film out of the fourteen 31 film objects I have assessed that is formatted this way, therefore there is no basis for comparison or for determining if this was a common practice. However, the use of such a thoughtless method of cropping suggests that Castle Films did not respect the original film in the first place and points to a general mode of thought that was pervasive throughout the company. The third film that was released in the Science Fiction/Horror series, a shortened version of This Island Earth (1955), that was renamed War of the Planets for release in 1958, was originally screened in the 2.00:1 aspect ratio. Comparing the visual treatment of War of the Planets, along with other films that were 76 originally released in widescreen aspect ratios, to the Castle Films version of The Incredible Shrinking Man would provide insight into the technical practices of Castle Films and how those practices may have evolved over time. 76 MacGillivray, 186. Opening and Closing Credits For Creature from the Lagoon, released in 1957, completely new title cards were created and shown at a rapid pace. The Mummy and Dracula, released in 1962 and 1963 respectively, used the same technique. House of Frankenstein, released in 1967, and the films released thereafter indicate that at some point between 1965 and 1967 Castle Films normalized the opening credit sequences and made them uniform. Overlaying the credits on shot sequences that contained no credits, first observed in Bride of Frankenstein, became the standard. Moreover, aspects such as the typographical layout and the font used for the credits also became ubiquitous. The average length of the opening credits is approximately 16 seconds, or 3% percent of the total run time of the average Castle Films release. same, except that a title card, which states “Castle Films Presents," is superimposed over the image of the fireball traveling from right to left across the screen. Bride of Frankenstein, released three years later in 1960, uses different credits than the feature film. The original Bride of Frankenstein opens with a credit sequence that fades in and out over a shot of smoke billowing in the background. The Castle Films version places credits that are approximately eighty seconds shorter over an exterior shot of Frankenstein's Castle. This technique is used again in Son of Frankenstein, which was released by Castle Films in 1965. Frankenstein Meets the Wolfman’s credits faded in and out over top of a single shot of steam emanating from a laboratory glass. The sequence was shortened by first removing the fade transitions, and then using crossfades to transition from one credit to the next. Using this method reduced the length of the credits by 69 seconds. For Creature from the Lagoon, released in 1957, completely new title cards were created and shown at a rapid pace. The Mummy and Dracula, released in 1962 and 1963 respectively, used the same technique. House of Frankenstein, released in 1967, and the films released thereafter indicate that at some point between 1965 and 1967 Castle Films normalized the opening credit sequences and made them uniform. Overlaying the credits on shot sequences that contained no credits, first observed in Bride of Frankenstein, became the standard. Moreover, aspects such as the typographical layout and the font used for the credits also became ubiquitous. Opening and Closing Credits The target running time of the Castle Films versions made it necessary to truncate, speed up, or alter the opening and closing credits for each film. The opening credit sequences, measuring as long as 90 seconds in Son of Frankenstein, would have accounted for approximately 18% of the total running time of the abridged version if left untouched. Beginning with It Came From Outer Space and ending with Revenge of the Creature, various styles of altering the credits were employed. There is no discernable visual evidence that points to why a particular method was used. The original It Came from Outer Space begins with the Universal International logo and is followed by a shot of a fireball streaking from right to left across the frame. The following shot shows the fireball racing towards the audience and eventually exploding when it “hits” the screen. The fireball hitting the screen is synchronous with the bold faced title of 32 32 the film appearing, and there are no other credits used. The Castle Films version is largely the same, except that a title card, which states “Castle Films Presents," is superimposed over the image of the fireball traveling from right to left across the screen. Bride of Frankenstein, released the film appearing, and there are no other credits used. The Castle Films version is largely the same, except that a title card, which states “Castle Films Presents," is superimposed over the image of the fireball traveling from right to left across the screen. Bride of Frankenstein, released three years later in 1960, uses different credits than the feature film. The original Bride of Frankenstein opens with a credit sequence that fades in and out over a shot of smoke billowing in the background. The Castle Films version places credits that are approximately eighty seconds shorter over an exterior shot of Frankenstein's Castle. This technique is used again in Son of Frankenstein, which was released by Castle Films in 1965. Frankenstein Meets the Wolfman’s credits faded in and out over top of a single shot of steam emanating from a laboratory glass. The sequence was shortened by first removing the fade transitions, and then using crossfades to transition from one credit to the next. Using this method reduced the length of the credits by 69 seconds. Opening and Closing Credits I found similar rearrangements and omissions in all of the Castle Films versions examined with the exception of The Mummy’s Ghost. The presentation of the ending credits also vary from film to film. It Came From Outer them entirely. Technical and artistic staff, such as the art director, cinematographer, and film editor were also commonly omitted from the credits. In the original Dracula, for example, the opening credit sequence uses four title cards whereas the Castle Films version uses three. The first card in each version displays the name of the film. The original film also displays a credit for Bram Stoker, the two writers who wrote the play that the film adapted (Hamilton Deane and John L. Balderston), and a credit for Tod Browning and Carl Laemmle Jr, the director and producer respectively. The Castle Films version, on the other hand, only displays the title of the film, the text “Excerpts from the Photoplay ‘Dracula’," and the original Universal Pictures copyright. The second card in the original shows the names of the technical and artistic staff who worked on the film. This card is omitted from the Castle Films version and the names are not reproduced elsewhere in the opening credit sequence. The third card from the original film contains only Tod Browning’s director credit. The Castle Films version displays this credit on the third (and last) card, but “directed by” is changed to “A Tod Browning Production." The final card of the original film’s opening contains the actors’ credits nine in total. The actor card in the Castle Films version contains only four: Bela Lugosi as Count Dracula, Helen Chandler as Nina, David Manners as John, and Edward Van Sloan as Van Helsing. There are also differences in the names of the characters between each version. In the original film Helen Chandler plays “Mina," not “Nina," and David Manners plays “John Harker” instead of simply “John." The changes are minor, and the reason for changing the names of the characters is unknown. I found similar rearrangements and omissions in all of the Castle Films versions examined with the exception of The Mummy’s Ghost. them entirely. Technical and artistic staff, such as the art director, cinematographer, and film editor were also commonly omitted from the credits. In the original Dracula, for example, the opening credit sequence uses four title cards whereas the Castle Films version uses three. Opening and Closing Credits The average length of the opening credits is approximately 16 seconds, or 3% percent of the total run time of the average Castle Films release. Castle Films would also change how much information and what kind of information would be present in the opening credits. The credits for Frankenstein Meets the Wolfman, for example, were left intact and no information was altered or omitted. However, credit sequences that were made from scratch for the abridged versions would rearrange the placement of actors or omit 33 them entirely. Technical and artistic staff, such as the art director, cinematographer, and film editor were also commonly omitted from the credits. In the original Dracula, for example, the opening credit sequence uses four title cards whereas the Castle Films version uses three. The first card in each version displays the name of the film. The original film also displays a credit for Bram Stoker, the two writers who wrote the play that the film adapted (Hamilton Deane and John L. Balderston), and a credit for Tod Browning and Carl Laemmle Jr, the director and producer respectively. The Castle Films version, on the other hand, only displays the title of the film, the text “Excerpts from the Photoplay ‘Dracula’," and the original Universal Pictures copyright. The second card in the original shows the names of the technical and artistic staff who worked on the film. This card is omitted from the Castle Films version and the names are not reproduced elsewhere in the opening credit sequence. The third card from the original film contains only Tod Browning’s director credit. The Castle Films version displays this credit on the third (and last) card, but “directed by” is changed to “A Tod Browning Production." The final card of the original film’s opening contains the actors’ credits nine in total. The actor card in the Castle Films version contains only four: Bela Lugosi as Count Dracula, Helen Chandler as Nina, David Manners as John, and Edward Van Sloan as Van Helsing. There are also differences in the names of the characters between each version. In the original film Helen Chandler plays “Mina," not “Nina," and David Manners plays “John Harker” instead of simply “John." The changes are minor, and the reason for changing the names of the characters is unknown. Opening and Closing Credits The first card in each version displays the name of the film. The original film also displays a credit for Bram Stoker, the two writers who wrote the play that the film adapted (Hamilton Deane and John L. The presentation of the ending credits also vary from film to film. It Came From Outer Space, which only featured the name of the film in the opening credits, shows the entire cast and crew in the closing credits. The Castle Films version, however, highlights Richard Carlson as the main actor and omits the rest of the cast. In addition, the abridged version shows four cards of 34 credits over the same shot of the fireball streaking across the sky, only this time the fireball is moving in reverse. The original film shows a clip of each actor with his or her name present at the bottom followed by a series of technical and artistic credits, all of which are shown over a sparkling starlit sky. The Castle Films version of The Mummy ends with three cards, one dedicated to the screenwriters, one dedicated to the director, and a “The End” card that features the Castle Films logo. In Frankenstein Meets the Wolfman, the end credits begin with the “produced by” and “directed by” sections from the revised opening credits, then fade to a different style of the Castle Films “The End” card. credits over the same shot of the fireball streaking across the sky, only this time the fireball is moving in reverse. The original film shows a clip of each actor with his or her name present at the bottom followed by a series of technical and artistic credits, all of which are shown over a sparkling starlit sky. The Castle Films version of The Mummy ends with three cards, one dedicated to the screenwriters, one dedicated to the director, and a “The End” card that features the Castle Films logo. In Frankenstein Meets the Wolfman, the end credits begin with the “produced by” and “directed by” sections from the revised opening credits, then fade to a different style of the Castle Films “The End” card. Dracula, Frankenstein Meets the Wolfman, Son of Frankenstein, House of Frankenstein, The Mummy’s Ghost, The Wolf Man, and Frankenstein each end with a solitary “The End” card adorned with the Castle Films logo. Opening and Closing Credits The Incredible Shrinking Man and Man Made Monster also end with the appearance of the lone “The End," however it is not displayed on a card; the words are superimposed over an image that comes from the end of the film. The Castle Films logo is not seen. Using the single “The End” credit was the prefered method of ending the abridged films, but unlike the standardization of the opening credits there was no movement towards uniformity. Narrative The changes that were made to the narrative are the most striking example of the differences created by the condensation process. Each film, regardless of how long the original film was, had to be edited to remove entire scenes, characters, and plot points to meet the time constraints of a single reel release. The result is that the Castle Films versions have different narratives altogether. The new narratives use different numbers of characters, change the 35 development and motivations of the characters who remain, and alter the ending to better suit the new narrative. House of Frankenstein is the best example of a narrative changing completely. The original House of Frankenstein opens with Professor Lampini’s Chamber of Horrors traveling by horse and buggy through a dark forest. The traveling side show passes a prison from which two of the main characters, Dr. Gustav Niemann and Daniel, a hunchback, have just escaped. Niemann and Daniel encounter Professor Lampini’s carriage and, after talking with Lampini, murder him. Niemann impersonates Lampini and the two travel to the town of Reigelberg. Upon arriving they join a larger circuslike festival that features other traveling sideshows. During a demonstration Niemann shows the bones of Dracula to an onlooking crowd and then pulls out the stake that had killed the vampire. Dracula is resurrected as soon as the stake is removed. Dracula then attempts to seduce and finally kidnap a woman at the fair, culminating in a final chase scene. The chase ends with Dracula realizing that the sun is about to rise and his attempt to crawl back into his coffin. He turns back into a skeleton and is left on the mountain side. The next scene depicts Lampini’s Chamber of Horrors entering the town of Frankenstein. The opening section of the feature film, approximately 28 minutes in length, is entirely omitted from the Castle Films version. Instead, the abridged film begins with Lampini’s Chamber of Horrors entering the outskirts of the town of Frankenstein with no indication of what happened in the opening act of the original version. All of the characters who made up the town of Reigelberg, Neimann’s and Daniel’s prison escape, and the murders they recently committed are omitted. In fact, there is no indication that Neimann is not Professor Lampini. Suppression of Neimann and Daniel’s story alters the motivations for actions that the pair take later in the film. Narrative House of Frankenstein also illustrates another commonality between the abridged versions: the omission of romantic subplots. As Neimann and Daniel enter the town of 36 Frankenstein they see a woman named IIoka performing a traditional dance. Daniel falls in love with her, and his love acts as the principal motivating factor for assisting Neimann to revive both the Frankenstein monster and the Wolfman. Once the Wolfman is revived and he returns to the 77 human form of Lawrence Talbot, Daniel and Talbot feud over the love of Ilonka. Their conflict, along with Talbot’s courtship of IIonka are not shown in the abridged film. Consequently, Daniel’s motivation for helping Neimann shifts from his love of Ilonka to his desire to be healed by the science that created the Frankenstein monster. Moreover, after Ilonka’s appearance at the beginning of the Castle Films version she is not seen again. Romantic subplots are removed from Castle Films abridgements as a matter of routine. The removal of the romantic subplot in Bride of Frankenstein directly affects the ending of the film, and consequently requires the ending of the Castle Films version to be altered. In the original version, the monster becomes enraged after the Bride rejects him. The monster reaches for a lever that, if pulled, would cause the destruction of the entire lab and everyone in it. Dr. Frankenstein’s partner, Elizabeth Lavenza, arrives at the door of the lab and begs Frankenstein to come with her. The monster, feeling sympathy towards the couple, allows Frankenstein and Elizabeth to escape before finally destroying the lab and, in the process, killing himself, his Bride, and the evil Dr. Pretorius. In the Castle Films version Elizabeth is not shown coming to the door of the lab, and the monster grants no final act of mercy. Frankenstein is not shown leaving the lab, and he is killed along with Dr. Pretorius, the monster, and the Bride. The removal of the romantic subplot in Bride of Frankenstein directly affects the ending of the film, and consequently requires the ending of the Castle Films version to be altered. In the original version, the monster becomes enraged after the Bride rejects him. The monster reaches for a lever that, if pulled, would cause the destruction of the entire lab and everyone in it. Dr. 77 Niemann promises Daniel that if he helps him to unlock the secrets of Dr. Frankenstein’s early experiments Niemann will be able to turn Daniel into an “Adonis." p p experiments Niemann will be able to turn Daniel into an “Adonis." Narrative The shots of Elizabeth’s arrival and her escape with Frankenstein would have added no more than 10 seconds to the running time of the Castle Films version, which in its current state is 8 minutes and 12 seconds long. The abridged version of Frankenstein is 8 minutes and 20 seconds long, so it would have been possible to include the additional footage of Frankenstein and 37 Elizabeth escaping. These shots were removed because within the new narrative Elizabeth’s arrival and Frankenstein’s escape had no narrative basis. Elizabeth is not seen in the Castle Films version, so she would appear as a character without discernible motivations. Frankenstein, in the original film, is forced to help Dr. Pretorius create the Bride after Pretorius convinces the monster to kidnap Elizabeth. Frankenstein is often shown protesting against his involvement, but he ultimately helps create the Bride because he does not want Elizabeth to be harmed. However, in the Castle Films version, the kidnapping of Elizabeth is never shown and Frankenstein’s motivations for helping Pretorius are implied to be scientific, much in the same way Frankenstein was driven to originally create the monster in Frankenstein. This lack of motivational clarity on Frankenstein’s part is ultimately what leads to his destruction in the Castle Films version; he has done nothing to deserve the sympathy of the monster and is shown to be a collaborator instead of a forced participant. For this reason he shares the same fate as Dr. Pretorius. Shot Reordering and Causality The original films each have their own specific chronology that dictates the specific cause and effect relationships between events and characters that drive the narrative forward. The chronology is built around the premise that a specific shot will always come after another specific shot, and so on throughout the film. However, when creating the abridged versions the original chronology of the feature films was disrupted by the excision of large parts of the narrative. The Castle Films versions, however, are not clumsily edited together highlight reels. In each of the films, there is a visible effort to create a new chronology that, consequently, creates new cause and effect relationships. In order to create a new chronology the Castle Films editors would move shots from other sections of the feature films and insert them into narratively strategic places. The 38 editors of the abridged films used shot reordering for three distinct purposes: to hide the presence of a character in a scene, creating a narrative bridge between two scenes that are separated in the original film, and to change the character motivations for a specific action. The monster is shown prominently in the original Bride of Frankenstein, but he is primarily absent from the Castle Films version. The first appearance of monster is withheld until after the Bride comes to life at the end of the film. It could be argued that the monster’s late appearance is merely coincidental due to the large amount of footage that needed to be cut away, but one scene in particular that was included in the abridged film indicates a conscious decision to withhold the monster until the end. While fleeing capture in the original film the monster hides in a crypt. As he hides, Dr. Pretorius and two hired helpers enter the crypt to search for a body to be used for the Bride. The monster watches them as they select a coffin to open. The following shot sequence begins with the two helpers prying at the edges of a coffin, cuts to a medium shot of the monster looking on, and then cuts back to a shot of the helpers opening the coffin. The first and third shots of this sequence are used in the Castle Films version, but the second shot of the monster is cut out. Shot Reordering and Causality Instead of cutting directly from the first shot to the third shot the editor inserted a close up of a dead woman’s face, however it is not the face of the Bride. This decision points to a conscious choice on the part of the producer of the Castle Films version to remove the monster from the scene while maintaining the basic sequential structure of the original sequence. It also indicates that the producers wanted to withhold the first appearance of the monster until the end for maximum impact. A more substantial example of shot reordering is the creation of a narrative bridge, connecting two scenes that were separated temporally in the original film. Narrative bridges collapse the plot of the film and allow the abridged versions to ignore large amounts of narrative without affecting the new plot. Man Made Monster creates a narrative bridge to move the 39 narrative as quickly as possible to the end chase scene, thus removing narrative elements that were not required. The original Man Made Monster centres on Dan McCormick, an ordinary man who was the sole survivor of a bus crashing into an electrical station. All of the passengers were electrocuted, signaling that McCormick has a high tolerance for electrical shock. McCormick is later visited in the hospital by Dr. John Lawrence who requests that McCormick come to his research centre after leaving the hospital. Upon arriving at the research centre, McCormick is tricked into being used as the subject of illegal experiments by Dr. Paul Rigas, one of Dr. Lawrence's colleagues. The experiments turn McCormick into a battery able to release electrical energy into anyone he touches, killing them. McCormick becomes addicted to electricity as a side effect of the experiments, and his addiction allows Dr. Rigas to enslave him. McCormick is shown at escalating stages of Dr. Rigas’ experiments in the feature film. The Castle Films version can not show the different stages of experimentation, and instead connects the first successful experiment and the last experiment using a narrative bridge. After the first experiment is complete McCormick is ordered by Rigas to kill the interfering Dr. Lawrence. With Lawrence dead, Rigas hands an object to McCormick that will bleed away his energy, essentially turning him off. Shot Reordering and Causality Rigas then convinces McCormick to take responsibility for Lawrence's murder, and as a result McCormick is arrested, found guilty, and taken to the electric chair for execution. The electric chair reenergizes McCormick who then returns to Rigas’ laboratory to confront him. When McCormick arrives he discovers that Rigas was about to murder June Lawrence, Dr. Lawrence’s daughter and McCormick’s love interest. McCormick kills Rigas and takes June out of the laboratory, instigating the final chase scene of the film. In the Castle Films version the narrative elements that centre on McCormick’s arrest, trial, and execution are removed. Instead, the abridged film moves directly from the scene where Rigas hands McCormick the object that bleeds away his energy, cuts to a shot of Dr. Lawrence’s dog 40 waiting outside of the laboratory, and then cuts to the moment when McCormick kills Rigas and kidnaps June. The two separate shots of McCormick and Rigas use identical shot scales, figure placement of the actors, and location within the lab. The inclusion of the shot of Dr. Lawrence’s dog eases the transition between these two shots. A cut from one to the other would have been jarring due to the minor visual differences between the shots; the clothes worn by both characters, for example, are different in each scene. The three shot sequence makes up the whole of the narrative bridge. The final approach to reordering shots that producers at Castle Films would employ involved swapping the placement of two scenes in order to create new motivation for an action. Moving scenes became necessary during the abridgement process in order to maintain a coherent cause and effect relationship inside the Castle Films version. The Creature from the Lagoon uses this strategy to change the motivation for escape from the Black Lagoon. The original film depicts a scientific expedition committed to finding evidence of a Devonian era creature. The expedition travels down the Amazon River and into the Black Lagoon where they instead find the creature itself. A number of conflicts between the crew and the creature leave members of the expedition dead or severely injured. Dr David Reed, the leader of the expedition, implores the crew to leave the Black Lagoon and return to safety. The crew agrees, but as their ship travels towards the exit of the lagoon they run into a series of toppled trees that block their exit. Dr. Shot Reordering and Causality Reed and Mark Williams, one of the other main characters, dive into the water to clear away the debris. While working on the debris they are attacked by the creature, and Williams is killed. The Castle Films version uses a narrative bridge to collapse the narrative content between Dr. Reed and Williams’ first dive into the lagoon and the dive where Williams is killed. The scene begins with a shot from much earlier in the film that depicts the first time Reed and Williams dive into the lagoon. The Castle Films version then cuts to the attack by the creature that 41 leads to the death of Williams. Moving from one scene to the other was seamless because Reed and Williams were underwater in both scenes and dressed identically. The narrative bridge eliminates all of the content in between the two separate dives and allows the narrative to move forward. However, the ship must still escape the Black Lagoon. The chronology of the ship crashing into the downed trees and Williams’ death is reserved, meaning that Williams’ death becomes the motivating factor for escape from the lagoon as opposed to the impassioned pleas of Reed shown in the original version. A logical cause and effect relationship is maintained, but how the events relate to each other in the cause and effect chain is altered to suit the newly established narrative. In short, the characters in the abridged film are taking the same actions but for different narrative reasons. Conclusions of Analysis The conclusion that I draw from this analysis is that the shortened versions are new films with different lengths, credits, casts of characters, narratives, and cause and effect relationships. They are drawn from the same genetic makeup as the originals, but their relationship should be thought of as two siblings born at different times and raised under different circumstances, with only their original genes as proof of their relation. Based on these findings I believe it is necessary for moving image archivists who currently have abridged commercial films in their collections to reconsider the preservation priority of the objects. The objects represent a section of film history that, to a large extent, has yet to be mined by researchers and, as such, require a higher degree of care than they currently receive. At the very least, the objects need to be thought of not as compromised or inferior films, but instead as wholly new works. 42 42 The analysis also provided insight into the process that an editor might have followed when condensing a feature length film. Based on what I have observed the abridged versions are not created to fit a mould, and they do not conform to easily identifiable patterns. The illustrations found in Appendix A show where the images from each abridged version came from, and the finale is the only consistent section that the editors draw images from. It may seem natural for all of the abridged films to contain images from the finales of the original versions, but we can not forget that entire sections were left out of the abridged films. In the case of a film like House of Frankenstein, it is conceivable that the editor at Castle Films could have taken the first section of the film, which depicts Dracula’s resurrection and eventual demise in just 28 minutes, and created an abridged version from that section alone. The film would have had to be renamed (there are several examples from Castle Films’ catalogue of films being renamed after they have been abridged) and then it could have been distributed as part of the science fiction and horror series. The process would have taken less time, and a more coherent narrative that more closely resembled the original narrative could have been told. Instead, the editor or producer chose to focus on the final 43 minutes of the film. 8 "The Shining Recut, HD," YouTube, Web, 18 June 2015, g , , , , , <https://www.youtube.com/watch?v=6s40Q6ODSI8>. ng Recut, HD," YouTube, Web, 18 June 2015, Conclusions of Analysis Without documentation from the offices of Universal International or Castle Films it is impossible to be sure if editors and producers had specific rules for condensing the feature films with which they had to abide. However, it is possible to determine why an editor would choose to keep certain shots while omitting others by focusing on the one element that is always present in the abridged versions: dynamic motion both in and between images. It is clear that the editors of the abridged versions could potentially cut the feature films into any form they pleased. To use a modern example, an amateur editor released a trailer for Stanley Kubrick’s The Shining online that arranged the footage to give the impression the film was a romantic comedy. The abridged films 78 43 made by Castle Films could have similarly focused on different aspects of the narrative. The romantic subplots that were completely removed could have easily been the narrative focus. There was, however, a concerted effort to define the narrative of the abridged versions by the creature/monster/alien, their creation, the terror they wrought, and their eventual defeat in the finale. Each abridgement’s compliance to that three part structure demonstrates that there were guiding principles in place. It could be said that the abridged films are similar to found footage films that repurpose images from a film in order to create a new work. Typically images in found footage films are repeated rhythmically, juxtaposed with other images, rephotographed using colourization techniques, etc., for an artistic purpose. However, the abridged films were not constructed by editors who were allowed to express themselves artistically through their work. I discussed previously how film elements from other sections of the feature films could be used to create smooth transitions between sections using a narrative bridge. However, the use of subtle and unobtrusive editing techniques was not a primary concern. As demonstrated by the thoughtless cropping of The Incredible Shrinking Man, there was little care taken to preserve the visual fidelity created by the unobtrusive, Classical Hollywood style of editing found in the original films. Each of the abridged versions contain awkward cuts within scenes where, for example, at one moment a character has his back to the camera, and then in the next shot he is facing the camera and speaking to another character off screen. 79 The feeling is similar to what a viewer experiences while watching a Hollywood film that has committed to the oppressive editing style popularized by the films of Michael Bay and which is now commonplace amongst highbudget Hollywood cinema Conclusions of Analysis Visual impact, as opposed to visual fidelity and artistic motivations, is the concern of the editors. Due to the constraints of reducing a feature film to an eight minute short film the editors were at the mercy of the narrative, but they typically gravitated towards images that contained a high degree of dynamic motion regardless if they matched graphically with the surrounding shots or if they created awkward editing rhythms. If a Castle Films editor used a jump cut it was not 44 because the editor was attempting to undermine classical Hollywood conventions of editing, but because using a jump cut was the most effective way to join two narratively relevant images whose combination created a visually impactful juxtaposition. The reliance on highly dynamic images underlines the guiding principle used to determine whether a shot would be selected for use in the condensation process: the image and audio had to efficiently portray the new narrative while using the most impactful images to do so. By following this principle the abridged films not only condensed the narrative, but concentrated the excitement a viewer would experience over the course of a feature length film. The effect is numbing as we are only presented with moments of decision and action, with the subtleties and nuances of the original film removed in favour of highlighting the most visually dynamic images. It is for this reason that the abridged versions 79 draw a majority of their shots from the final third of the original film, the section of the film that would contain the final chase, the destruction of the lab, the death of the monster and quite often his creator. These images are indicative of the kind of spectacle a viewer would expect from the “grand finale” of any horror or science fiction film from the Classical Hollywood period, except in the case of the abridged film they do not need to wait through the majority of the film to be stimulated. The idea of concentrated entertainment, or fast entertainment, directly relates to Ariel Dorfman’s notion of the infantilized adult reader, or in this case viewer. His cultural critique will be expanded on in the following chapter in relation to the Reader’s Digest Condensed Book Club. p g y y http://www.oed.com/view/Entry/65746?rskey=8GUs3T&result=1#eid 80 "Excerpt." Oxford English Dictionary. Web. 6 May 2015. Conclusions of Analysis 79 The feeling is similar to what a viewer experiences while watching a Hollywood film that has committed to the oppressive editing style popularized by the films of Michael Bay and which is now commonplace amongst highbudget Hollywood cinema 45 Chapter 7 Defining the Film Objects Referring to a Castle Films version as a different film or a film with a new narrative lacks the specificity necessary to accurately describe it. There are several possible defining terms that could be used to describe the film objects including excerpt, digest, and remix. This section will focus on defining each term and applying that definition to the film objects and their method of construction to determine which term best describes the film objects. Excerpts is the term that Castle Films used to describe its own products. In the opening credits of each film there is a title card that features the name of the film and a variation of the following statement: “Excerpts from the photoplay [film title]." The librarians who created the labeling for each film also included the term “excerpts” in brackets beside the title on the film canister, indicating that the statement shown above was viewed during initial inspection or on literature that accompanied the film objects. When describing the film objects to my peers they have used terms that are synonymous with excerpts, such as “highlights” and “compilations”, to describe the objects. These three terms, in the context of print media, can be defined as “a passage taken out of a printed book or manuscript; an extract, quotation, selection." But this definition does not 80 account for the fact that these film objects are more than sections pulled from a longer film and laid out in chronological order. More often than not the majority of footage used in the abridged versions came from the final act of the original film, but the way that the entire abridged version is constructed creates a new context for the final climactic ending. In other words, how the characters reach the climax of the respective version is different, and the shortened version requires the construction of a new narrative in order to reach that climax. Defining the abridged 46 versions as excerpts, highlights or compilations neglects the narrative transformation that each film object was subjected to during editing, and are thus not ideal terms to use when defining them. 81 Wood, Of Lasting Interest, 1967. 175. 82 Ibid. 83 Ibid, 176. 82 Ibid. 83 Ibid, 176. 82 Ibid. Ibid, 178. 85 Nicholas Monsarrat, Smith and Jones, London: Cassell & Company LTD, 1963, 90. Conclusions of Analysis In order to access the validity of using the term digest to describe the abridged films, I will compare and contrast how Castle Films created its abridgements to how Reader’s Digest, a publishing contemporary of Castle Films, created digest versions of magazine articles and novels. There are parallels between the creation of an abridged Castle Films short and the creation of digest versions of novels and articles made for public consumption. Beginning in 1934 Reader’s Digest began publishing digest versions of nonfiction books, the first of which was “How to Live on TwentyFour Hours” by Arnold Bennett. In 1949 the The Reader’s Digest Condensed Book 81 Club began circulation in the United States. The book club compendium was published four 82 times a year, and each would contain two longer fictional stories and shorter nonfictional works typically totalling 575 pages per volume . The condensing of longer works into digest form 83 matches the business practices of Castle Films: take a long form subject, remove content to truncate the overall narrative, and rerelease the shorter version to the public. There are, however, differences between the how literary works were condensed and how feature length films were abridged that calls into question the use of a term like digest to describe the film objects. This section will examine the similarities and the differences in creating condensed literature and abridged films, after which I will discuss the appropriateness of using digest as a term to define abridged films. James Playsted Wood, who wrote a corporate history of Reader’s Digest in 1958, James Playsted Wood, who wrote a corporate history of Reader’s Digest in 1958, James Playsted Wood, who wrote a corporate history of Reader’s Digest in 1958, described the process that was used to condense a book: described the process that was used to condense a book: described the process that was used to condense a book: “Once the decision is made to accept a complete book, the Digest buys the rights to condense it, and the book is assigned to an editor for the first cut. Three more editors go over this first condensation, making further cuts, perhaps restoring some 47 already made, making sure that the contents, the spirit, and the style of the author are retained in the shortened version. Nothing essential is changed. 86 Nicholas Monsarrat, "Smith and Jones," Reader's Digest Condensed Books, Vol. 2, Pleasantville, N.Y.: Reader's Digest Association, 1963. 361. g , 7 Alexander Klein, The Counterfeit Traitor, London: Frederick Muller, 1958. 20. 86 Nicholas Monsarrat, "Smith and Jones," Reader's Digest Condensed Books, Vol. 2, Pl Reader's Digest Association, 1963. 361. Nicholas Monsarrat, Smith and Jones, Reader s Digest Condensed Books, Vol. 2, Pleasantville, N.Y.: Reader's Digest Association, 1963. 361. , in, The Counterfeit Traitor, London: Frederick Muller, 1958. 20. 84 Ibid, 178. , , p y , , 86 Nicholas Monsarrat, "Smith and Jones," Reader's Digest Condensed Books, Vol. 2 Reader's Digest Association, 1963. 361. 8. s Monsarrat, Smith and Jones, London: Cassell & Company LTD, 1963, 90. bid, 178. Nicholas Monsarrat, Smith and Jones, Reader s Digest Condensed Books, Reader's Digest Association, 1963. 361. , , pg 9 Alexander Klein, "The Counterfeit Traitor," Reader's Digest Condensed Books, Spring ed, g Pleasantville, N.Y.: Reader's Digest Association, 1958, 281. 88 Klein, The Counterfeit Traitor, pg 20. g p g ville, N.Y.: Reader's Digest Association, 1958, 281. n 131 Conclusions of Analysis Where deletions necessitate the insertion of transitions, these are made in the manner of the author. The condensed book, still in Digest “manuscript,” then goes to the copy desk, where a final reading is made, and the copy is prepared for the printer.” 84 Wood’s claim that “nothing essential is changed” is preposterous. When a work of literature is condensed in any way it is immediately stripped of the author’s intentions and style, two facets of its creation that are undeniably essential. For example, the eighth chapter of Smith and Jones by Nicholas Monsarrat begins with the following sentence: “They settled down very quickly; it was made easy for them by a continuing popular welcome, and a benign official smile.” 85 The same sentence taken from the condensed version reads “Smith and Jones settled down very quickly.” The narrative action of the two main characters settling down remains the same, but the 86 nuance of that action and the context in which it is performed is removed. The condensed novels and articles are systematically stripped of the words, phrases, and paragraphs that define the original tone of the work so the reader only has to engage with the content on a purely narrative level. In addition to Smith and Jones I have also compared Anya Seton’s The Winthrop Woman, Alexander Klein’s The Counterfeit Traitor, and E.R. Braithwaite’s To Sir, With Love with their respective condensed versions. The editors of Reader’s Digest, beyond removing sections of a sentence, would also combine sentences in order to accelerate conversations between characters. For example, in The Counterfeit Traitor, Ambassador Steinhardt asks Eric Erickson, the main character, if he will become a spy. Erickson is astonished and Steinhardt replies “Yes, Eric. When the time comes, will you be willing to help us?” Erickson agrees, questions 87 Steinhardt what his role will be, and wonders out loud if he will be sent to the infantry. Steinhardt 48 replies “No. As you’ve probably guessed, it’s oil. Intelligence about the German oil industry.” In 88 the Reader’s Digest version Steinhardt’s answers are combined into one reply: “Yes, Eric. When the time comes, we will badly need intelligence about the German oil industry. Would you be willing to help us?” In terms of narrative nothing essential has changed. Alexander Klein, The Counterfeit Traitor, Reader s Digest Condensed Books, Spring ed, Vol. 2, Pleasantville, N.Y.: Reader's Digest Association, 1958, 281. ille, N.Y.: Reader s Digest Association, 1958, 281. n, 131. 91 Ibid. 90 Dorfman, 131. 89 Alexander Klein, "The Counterfeit Traitor," Reader's Digest Condensed Books, Spring ed, Vol. 2, Pleasantville, N.Y.: Reader's Digest Association, 1958, 281. 90 Dorfman 131 88 Klein, The Counterfeit Traitor, pg 20. 90 Dorfman, 131. 92 Ibid, 120. 88 Klein, The Counterfeit Traitor, pg 20. Conclusions of Analysis Erickson has still agreed 89 to help, and Steinhardt has still told Erickson that they will need intelligence regarding the German oil industry. The original pacing and style, however, has been altered significantly. replies “No. As you’ve probably guessed, it’s oil. Intelligence about the German oil industry.” In 88 the Reader’s Digest version Steinhardt’s answers are combined into one reply: “Yes, Eric. When the time comes, we will badly need intelligence about the German oil industry. Would you be willing to help us?” In terms of narrative nothing essential has changed. Erickson has still agreed 89 to help, and Steinhardt has still told Erickson that they will need intelligence regarding the German oil industry. The original pacing and style, however, has been altered significantly. The condensed products created by Reader’s Digest and Castle Films are systematic of what Ariel Dorfman describes as the need to satisfy rampant consumption in North American society. In a capitalistic society it is necessary to gain knowledge as quickly as possible because 90 it is knowledge that separates the successful from the failures in a supposedly equal community. The condensed novels and abridged films are products of the need to consume knowledge as quickly as possible and, to use Dorfman’s term, are best described as “fast ideas," ideas equal to that of fast food in that they stimulate the consumer but ultimately do not nourish them. They are 91 products made so consumers believe they are learning and gaining new experiences, but in actuality they are consuming the same narrative repeatedly. This is true for both the abridged films and the condensed literature. While the abridged films and the condensed novels were born in the same cultural context to satisfy rampant consumerism, the differences between how they were created questions the appropriateness of using digest to define the abridged films. Dorfman points out that Reader’s Digest would often plant articles in other magazines so they could be reproduced, in condensed form, in the Reader’s Digest magazine. 93 Ibid, 121. 93 Ibid, 121. Conclusions of Analysis Seeding articles in other publications for condensation 92 49 allowed Reader’s Digest to control the style and content of articles it was publishing for its readership, a practice that Dorfman mocks: “Each “selected” [article] cannot help but repeat the same language, procedure, technique, and ideology of all other [articles]." There was, in other 93 words, a discernible hegemony of style that was pervasive throughout all of the articles that Reader’s Digest published and that dictated which articles would be seeded. Castle Films, on the other hand, did not have the opportunity to seed films in theatres and then select them for abridgment. As a wholly owned subsidiary of Universal International it could only choose from the feature length films produced by Universal. Castle Films could not request that a feature film be made in a particular way so that it could be abridged and conformed to a hegemonic style. Despite this, the abridgements released by Castle Films still contain the repeated language, procedure, and technique that Dorfman saw in Reader’s Digest condensations. However, I argue that those repeated elements owe more to the ubiquitous Classical Hollywood style of film construction than the more dubious associations to cultural domination that Dorfman makes in regards to Reader’s Digest. The other main difference between the creation of abridged films and condensed literature is the degree to which each adheres to the structure of the original’s plot. As discussed above, it was routine for the editors at Castle Films to move a scene in order to create new cause and effect relationships and character motivations, to remove characters from the plot, and to change the ending of the original to better suit the narrative of the abridged film. While the condensations of Reader’s Digest would remove words, sentences, paragraphs, and even whole chapters while trying to streamline the narrative, it would only do so if the narrative cause and effect relationship and plot remained in tact. Additionally, Reader’s Digest did not create new endings for their condensed works that better suited the shortened narrative. The progression of the plot, in its 50 50 basic form, would not be altered. The movement of scenes that led to the alteration of the cause and effect relationships and the creation of new endings are some of the main distinguishing factors of the abridged films. http://www.oed.com/view/Entry/246356?rskey=Qi8SG6&result=1&isAdvanced=false#eid 94 "Remix," Oxford English Dictionary, Web, 6 May 2015. 95 “Remix," Oxford English Dictionary. 95 “Remix," Oxford English Dictionary. 4 "Remix," Oxford English Dictionary, Web, 6 May 2015. http://www.oed.com/view/Entry/246356?rskey=Qi8SG6&result=1&isAdvanced=false#eid 95 “Remix " Oxford English Dictionary 95 “Remix," Oxford English Dictionary. Conclusions of Analysis Despite that fact that both the abridged films and the condensed novels were stripped of their original nuance and style, the differences between both styles of condensation are rooted in the reorganized nature of the abridged films, an aspect that is tied directly to their identity as objects separate from the original films. For that reason, and for the reason that the works were selected for condensation based on different circumstances, digest is not an acceptable term to define the abridged films. The last term I would like to explore is remix. The term is generally reserved for audio recordings of songs in which a new version of a recording is made where “the separate instrument or vocal tracks are rebalanced or recombined," or, in a more modern context, a song is radically changed by “altering the rhythm and instrumentation." In extended use the term remix can also 94 be used to define “a reworked version." Both definitions can be applied to the abridged films, but 95 the definitions are too broad to specifically define them. The radical alteration of a sound recording is the most relevant aspect of remixing that can be applied to the Castle Films versions. The findings of my comparison show the abridged films have been radically altered. Similarly, the broader and nonmedium specific definition that a remix is any reworked version also applies. The issue with these definitions is that they do not specifically address the shortening and elongating of a work as an aspect of the remixing process. The exclusion of language that specifically speaks about length and the removal of content means that “remix," as defined here, is not specific enough to properly define the shortened Castle Films. However, I would argue that the radical alteration a song experiences while being remixed covers 51 the differences in length, content, and narrative. The definition itself is biased towards sound recordings, so a film definition of remix could easily include language that applied to length, content, and narrative. Remix is not a perfect way of defining the films, but it is the most appropriate option given the lack of alternative choices. 52 52 Chapter 8 Conclusion The comparative analysis that I have undertaken in this project outlines both technical and narrative differences between feature length films and their shortened counterparts. Possible Avenues for Future Research Conclusions of Analysis The shortened versions of feature films are now beginning to enter archives as part of home film collections, and it is necessary to understand how these film objects were constructed to assess their value to archival collections as objects of study. The role that abridged commercial films played in the history of home film distribution has only begun to be explored, and now that these film objects are entering moving image archive collections the opportunity to explore the history of this often overlooked section of film history is growing. However, it is important that film archivists who accession these types of shortened films understand that the films are not merely excerpts of, or highlights from, Hollywood films, but new narratives created using the same genetic material as the original. Framing the shortened films as merely excerpts undermines their status as selfcontained narratives that do not require knowledge of the original source material in order to understand or enjoy. Treating the film objects in any other way increases the possibility of neglect, both scholarly and archivally. Additionally, defining the objects as remixes allows archivists to assess the importance of the objects by removing the negativity of terms, such as excerpts, that connotatively imply the objects are mere highlight reels as opposed to new works. Declaring the objects as remixes acknowledges both the source of the material used to create the abridged films and the defining difference between the abridged films and their feature length siblings. Possible Avenues for Future Research 53 53 The sample size of this project is admittedly small. Castle Films released thousands of abridged films for the home market, and there were other companies who engaged with the same niche market. However, given that this project is, to my knowledge, the first of its kind to closely examine how abridged versions of feature length films differ from the original films, it opens several areas of future research. Castle Films’ process for creating the abridged versions is the most obvious hole in current literature about the company. Based on my observations I have suggested the guiding principles that an editor might have followed, but discovering internal documentation that dictated company policy for abridging a feature film would allow for new ways to approach the construction of the film objects. , Secret Cinema An Interview with Jay Schwartz," Framework: The Journal of Cinema and Media 49, no. 2, 2008. 116. 96 Elena Gorfinkel, “‘Shown in 16mm on a Giant Screen’: Adventures in Alternative Exhibition with The Secret Cinema An Interview with Jay Schwartz," Framework: The Journal of Cinema and Media 49, no. 2, 2008. 116. 96 Elena Gorfinkel, “‘Shown in 16mm on a Giant Screen’: Adventures in Alternative Exhibition with The Secret Cinema An Interview with Jay Schwartz " Framework: The Journal of Cinema and Media 49 no 96 Elena Gorfinkel, “‘Shown in 16mm on a Giant Screen’: Adventures in Alternative Exhib Conclusions of Analysis Internal documentation might also reveal the names of the people who worked on the abridgements, which in turn could lead to possible interviews and new sources of information regarding the creation of abridgements. Another avenue for approaching the construction of the film objects is an examination of the silent versions of each film that Castle Films released. For each 16mm sound print that Castle Films distributed, at least in the science fiction and horror series, there was a corresponding 16mm and 8mm silent copy made for consumers who had silent projectors or toy projectors. These silent versions typically ran for twelve minutes and used title cards or subtitles, much like the silent films of old, for the purpose of screen direction and dialogue. Did Castle Films insert 96 direction and speech cards into the abridged versions they created for 16mm, or did they create new versions whose image selection would be informed by the inherent limitations of silent projection? A comparison between an 8mm silent version and one of the 16mm sound versions could reveal more information about the technical practices of Castle Films. Another avenue for approaching the construction of the film objects is an examination of the silent versions of each film that Castle Films released. For each 16mm sound print that Castle Films distributed, at least in the science fiction and horror series, there was a corresponding 16mm and 8mm silent copy made for consumers who had silent projectors or toy projectors. These silent versions typically ran for twelve minutes and used title cards or subtitles, much like the silent films of old, for the purpose of screen direction and dialogue. Did Castle Films insert 96 direction and speech cards into the abridged versions they created for 16mm, or did they create new versions whose image selection would be informed by the inherent limitations of silent projection? A comparison between an 8mm silent version and one of the 16mm sound versions could reveal more information about the technical practices of Castle Films. 54 This project focuses solely on visual comparisons and did not delve into comparing how the audio was mixed and altered in order to suit the abridged films. I omitted the consideration of audio both for the sake of brevity and because it was difficult to analyze the less than ideal state of the scanned audio that the Sniper 16HD film scanner produced. 7 Ralph J Amelio, Hal in the Classroom: Science Fiction Films, Dayton, OH: Pflaum Publis melio, Hal in the Classroom: Science Fiction Films, Dayton, OH: Pflaum Publishing, 1974. 55 Conclusions of Analysis All of the films are warped, some worse than others, and this caused the audio transfer to be unusable. A comparison of the source audio to the audio used in the Castle Films versions would further expand the knowledge of Castle Films’ technical practices. For example, in order to perform the narrative bridge in Man Made Monster effectively the music in one of the scenes would need to be extended over all three of the shots that form the bridge. If this section were dissected using the appropriate tools it would reveal whether a sound editor had access to the appropriate material to remix the music, or if release deadlines permitted that level of sound editing. Examining the objects in the context of a university library lending collection could lead to a discussion of the use of abridged versions of commercial films in educational institutions. Examining the objects in the context of a university library lending collection could lead to a discussion of the use of abridged versions of commercial films in educational institutions. Inscriptions found near the head and tail on clear film leader indicate that the films were marked for sale to educational facilities. The company names present in the inscriptions include “UNIV EDUC," “Universal Education," and “Universal ED." Research into the role of science fiction films in educational settings has been done in the past, most notably the collection of articles entitled Hal in the Classroom: Science Fiction Films released in 1974 and edited by Ralph Amelio. The collection of articles argues for specific feature length films, such as 2001: A Space Odyssey and The Creature from the Black Lagoon, to be taught to high school students in ways that compliment broader discussions, such as on politics and the creation of cultural icons. The book 97 compliment broader discussions, such as on politics and the creation of cultural icons. The book 97 points to a contemporary attitude that science fiction and horror films can be used as educational tools. However, there is no mention of abridged films and their possible use in any of the articles. 55 The inscriptions on the film objects indicate that there was a belief that these abridged films could be used as educational tools. 98 MacGillivray, Castle Films, 2004. 192. 98 MacGillivray, Castle Films, 2004. 192. Conclusions of Analysis Identifying the perceived role and advantage of using abridged films in the classroom would expand both the context for the objects creation and the history of educational films, particularly the role of commercial films in the classroom. That discussion could also include Dorfman’s previously cited stance on the infantilization of the viewer to question why abridged films were purchased for the classroom in the first place. Finally, it would be interesting to learn what the original filmmakers of The Creature from the Black Lagoon, The Mummy, House of Frankenstein, and others thought about the abridged versions of their films. Determining the involvement that the filmmakers had in the condensation process would provide additional insight into how the abridged versions were created. Did the filmmakers have any input in the process, and if so to what extent? Jack Arnold, who was the director of It Came From Outer Space and The Creature from the Black Lagoon, was still under contract with Universal International at the time those two films were released by Castle Films, so it is conceivable that he could have participated in the process. For films such as Frankenstein, that was released by Castle Films forty years after its initial release, it is difficult to imagine the original filmmakers participating in the abridgement process. Further research into this area may also reveal the attitudes that filmmakers took towards their artistic work being abridged and reassembled. Discovering those attitudes would more than likely be accidental and come from musings recorded in interviews done with the filmmakers. Of particular interest would be the opinion of Alfred Hitchcock. The last two films to be abridged and released in the science fiction and horror series were two Hitchcock films, Psycho and Frenzy. 98 Hitchcock was notoriously careful when constructing his films during the shooting phase of production, and the way that Castle Films created new narratives that would have disrupted his Hitchcock was notoriously careful when constructing his films during the shooting phase of production, and the way that Castle Films created new narratives that would have disrupted his 56 carefully conceived editing patterns may not have pleased him. Conclusions of Analysis Despite the possible bemusement of an auteur these abridged films, as new films created from previously existing images and sound, deserve to be treated with respect in moving image archives so that preservation and academic study remain a possibility for scholars and the interested public. 57 57 Appendix 1.1 Appendix 1.1 What follows are visual representations of the comparisons that were made in Chapter 6. The illustrations show two key elements: the location in the full length version of a film from which the corresponding shots in the abridged version are taken and the order of those corresponding shots in the abridged version. The original version is represented by the two grey bars located in the middle of the page, and each bar corresponds to one half of the film. The visualization was split into two halves to allow for greater detail. When a corresponding shot is used in the same chronological order as the original film, it is denoted in red; shots that are not ordered chronologically are denoted in blue. In the lower section, entitled Castle Films Version Shot Order, the shot order of the abridged film is shown. The illustrations are ordered by the year of release of the Castle Films version, not by the release date of the original films. 58 58 59 59 60 60 61 61 62 63 63 64 64 65 65 66 66 67 67 67 68 68 68 69 69 69 70 70 71 71 71 72 72 Bibliography Amelio, Ralph J. “Introduction." Hal in the Classroom: Science Fiction Films. Dayton, OH: Pflaum Publishing, 1974. 15. Amelio, Ralph J. “Introduction." Hal in the Classroom: Science Fiction Films. Dayton, OH: Pflaum Publishing, 1974. 15. Anderson, Shawny. Condensed Hegemony: A Cultural/Ideological Critique of "Reader's Digest," 19801992. Ed. Edward Schiappa. Purdue University, 1994. 1992. Ed. Edward Schiappa. Purdue University, 1994. Braithwaite, E. R. To Sir, With Love. Englewood Cliffs: PrenticeHall, 1960. . R. To Sir, With Love. Englewood Cliffs: PrenticeHall, 1960. Braithwaite, E. R. "To Sir, With Love." Reader's Digest Condensed Books. Vol. 2 Braithwaite, E. R. "To Sir, With Love." Reader's Digest Condensed Books. Vol. 2. Pleasantville, N.Y.: Reader's Digest Association, 1963. 401468. N.Y.: Reader's Digest Association, 1963. 401468. Bray, Catherine. “Playing at Home." Sight and Sound 23, no. 12 (2013): 9. Clark, Mark. "For Monster Fans of the 1960s and '70s, Castle Films Abridgement Clark, Mark. "For Monster Fans of the 1960s and '70s, Castle Films Abridgements Were The Best Things on Earth!" Monsters from the Vault. Accessed December 1, 2014. http://www.monstersfromthevault.com/LittleGiants.html. Dorfman, Ariel. The Empire's Old Clothes: What the Lone Ranger, Babar, and Other Innocent Heroes Do to Our Minds. 2nd ed. Duke University Press, 2010. Gordon, Marsha. “Lenticular Spectacles: Kodacolor’s Fit in the Amateur Arsenal.” Film History: An International Journal 25, no. 4 (2013): 3661. 73 Gorfinkel, Elena. “Shown in 16mm on a Giant Screen: Adventures in Alternative Exhibition with The Secret Cinema An Interview with Jay Schwartz.” Framework: The Journal of Cinema and Media 49, no. 2 (2008): 114137. en with Castle Films." Harvard Film Archive. Accessed November 13, 2014. http://blogs.law.harvard.edu/hfacollections2/. Johnson, Ted. "Audiobooks Abridged vs Unabridged: The Tale of the Tapes." Los Angeles Times 30 Apr. 1995, Calendar; Calendar Desk sec.: 72. Print. Kattelle, Alan. “Amature Equipment Prior to 1923." Home Movies: A History of the American Industry, 18971979. Nashua: Transition Publishing, 2000. 5269 Katz, Ephraim. The Film Encyclopedia. 5th ed. New York: HarperCollins Publishers, 2005. Klein, Alexander. The Counterfeit Traitor. London: Frederick Muller, 1958. er. "The Counterfeit Traitor." Reader's Digest Condensed Books. Spring ed. Vol. 2. antville, N.Y.: Reader's Digest Association, 1958. 279352. Pleasantville, N.Y.: Reader's Digest Association, 1958. 279352. Klise, Kate. "On Condensed Books: Don't Snicker Too Much." St. Louis Post 7 May 1992, Everyday sec.: 2.C. Print. "Kodak Edge Codes." Film Forever. Accessed November 26, 2014. "Kodak Edge Codes." Film Forever. Accessed November 26, 2014. "Kodak Edge Codes." Film Forever. Accessed November 26, 2014. http://www.filmforever.org/Edgecodes.pdf. Limbacher, James L. Feature Films on 8mm, 16mm, and Videotape: A Directory of Feature Films Available for Rental, Sale, and Lease in the United States and Canada, with a Serial Section, an Index of Directors, and an Index of Foreignlanguage Films. 7th ed. New York: Bowker, 1982. Limbacher, James L. Feature Films on 8mm, 16mm, and Videotape: A Directory of Feature Films Available for Rental, Sale, and Lease in the United States and Canada, with a Serial Section, an Index of Directors, and an Index of Foreignlanguage Films. 7th ed. New York: Bowker, 1982. MacGillivray, Scott. Castle Films: A Hobbyist's Guide. New York: IUniverse, 2004. 74 Marchand, Philip. "Remembrance of Libraries Past; On the Reader's Digest Condensed Library." National Post 15 Mar. 2015, Books sec.: WP.5. Print. Marchand, Philip. "Remembrance of Libraries Past; On the Reader's Digest Conden National Post 15 Mar. 2015, Books sec.: WP.5. Print. Masson, Eef. “Celluloid Teaching Tools: Classroom Films in the Netherlands (1941 1953).” Film History: An International Journal 19, no. http://www.monstersfromthevault.com/LittleGiants.html. 4 (2007): 392400. McGuire, Philip. "Doing the Right Thing: Reader's Digest's Lasting Appeal." The Sun 25 Aug. 1993, Features sec.: 1C. Print. McGuire, Philip. "Doing the Right Thing: Reader's Digest's Lasting Appeal." The Sun 25 Aug. 1993, Features sec.: 1C. Print. Monsarrat, Nicholas. Smith and Jones. London: Cassell & Company LTD, 1963. Monsarrat, Nicholas. Smith and Jones. London: Cassell & Company LTD, 1963. Monsarrat, Nicholas. "Smith and Jones." Reader's Digest Condensed Books. Vol. 2. Pleasantville, N.Y.: Reader's Digest Association, 1963. 317400. Plato, Levno and Aaron Meskin. Encyclopedia of Quality of Life and Well Being Research, 1st ed., s.v. “Aesthetic Value.” New York: Springer Reference, 2014. s.v. “Aesthetic Value.” New York: Springer Reference, 2014. Reemes, Dana M. Directed by Jack Arnold. Jefferson, N.C.: McFarland, 1988. "Remix." Oxford English Dictionary. Web. 6 May 2015. Schaefer, Eric. "Gauging A Revolution: 16mm Film And The Rise Of The Pornographic Feature." Cinema Journal 41, no. 3 (2002): 326. Schneider, Alexandra. “Time Travel with Pathe Baby: The Smallgauge Film Collection as Historical Archive.” Film History: An International Journal 19, no. 4 (2007): 353360. Historical Archive.” Film History: An International Journal 19, no. 4 (2007): 353360. Historical Archive.” Film History: An International Journal 19, no. 4 (2007): 353360. Schreiner, Samuel Agnew. “The Condensed World of the Reader's Digest.” New York: Stein and Day, 1977. Seton, Anya. The Winthrop Woman. New York: Houghton Mifflin Harcourt, 1958. Seton, Anya. "The Winthrop Woman." Reader's Digest Condensed Books. Spring ed. Vol. 2. Pleasantville, N.Y.: Reader's Digest Association, 1958. 41278. Singer, Penny. "New Editor at Digest's Condensed Books." New York Times 2 Aug. 1987, Late Edition (East Coast) ed., sec. A: A.14. Print. 75 Slide, Anthony. Before Video: A History of the NonTheatrical Film. Westport, CT: Greenwood Press, 1992. "The Shining Recut, HD." YouTube. YouTube. Web. 18 June 2015. <https://www.youtube.com/watch?v=6s40Q6ODSI8>. Streible, Dan, Martina Roepke, Anke Mebold. “Introduction: Non Theatrical Film.” Film History: An International Journal 19, no. 4 (2007): 339343. Streible, Dan. "Moving Image History and the FWord; Or, "Digital Film" Is an Oxymoron." Film History: An International Journal 25, no. 12 (2013): 22735. Swanson, Dwight. "Home Viewing: Pornography and Amateur Film Collections, A Case Study." The Moving Image 5, no. 2 (2005): 13640. Swanson, Dwight. “Inventing Amateur Film: Marion Norris Gleason, Eastman Kodak and the Rochester Scene, 19211932.” Film History 15, no. 2 (2003): 126136. Haidee. "Electric Homes! Automatic Movies! http://www.monstersfromthevault.com/LittleGiants.html. Efficient Entertainment!: 16mm and Cinema's Domestication in the 1920s." Cinema Journal 48, no. 4 (2009): 121. Weaver, Tom. Universal Filmscripts Series, Classic Science Fiction: Creature from the Black Lagoon. Vol. 2. New Jersey: MagicImage, 1992. Wood, James Playsted. Of Lasting Interest: The Story of the Reader's Digest. 2nd ed. Garden City, N.Y.: Doubleday, 1967. 76
https://openalex.org/W2784247690	https://europepmc.org/articles/pmc5768609?pdf=render	English	null	MicroRNA 27a-3p Regulates Antimicrobial Responses of Murine Macrophages Infected by Mycobacterium avium subspecies paratuberculosis by Targeting Interleukin-10 and TGF-β-Activated Protein Kinase 1 Binding Protein 2	Frontiers in immunology	2,018	cc-by	13,645	Keywords: Mycobacterium avium subspecies paratuberculosis, microRNAs, miR-27a, mitogen-activated protein kinase p38, TAK1 binding protein 2, interleukin-10 Original Research published: 10 January 2018 doi: 10.3389/fimmu.2017.01915 MicroRNA 27a-3p Regulates Antimicrobial Responses of Murine Macrophages Infected by Mycobacterium avium subspecies paratuberculosis by Targeting Interleukin-10 and TGF-β-Activated Protein Kinase 1 Binding Protein 2 Tariq Hussain†, Deming Zhao†, Syed Zahid Ali Shah, Jie Wang, Ruichao Yue, Yi Liao, Naveed Sabir, Lifeng Yang and Xiangmei Zhou* Tariq Hussain†, Deming Zhao†, Syed Zahid Ali Shah, Jie Wang, Ruichao Yue, Yi Liao, Naveed Sabir, Lifeng Yang and Xiangmei Zhou* Edited by: Uday Kishore, Brunel University London, United Kingdom State Key Laboratories for Agrobiotechnology, Key Laboratory of Animal Epidemiology and Zoonosis, Ministry of Agriculture, National Animal Transmissible Spongiform Encephalopathy Laboratory, College of Veterinary Medicine, China Agricultural University, Beijing, China Reviewed by: Kushagra Bansal, Harvard Medical School, United States Anthony George Tsolaki, Brunel University London, United Kingdom Reviewed by: Kushagra Bansal, Harvard Medical School, United States Anthony George Tsolaki, Brunel University London, United Kingdom Mycobacterium avium subspecies paratuberculosis (MAP) persistently survive and replicate in mononuclear phagocytic cells by adopting various strategies to subvert host immune response. Interleukin-10 (IL-10) upregulation via inhibition of macrophage bactericidal activity is a critical step for MAP survival and pathogenesis within the host cell. Mitogen-activated protein kinase p38 signaling cascade plays a crucial role in the elevation of IL-10 and progression of MAP pathogenesis. The contribution of microRNAs (miRNAs) and their influence on the activation of macrophages during MAP pathogenesis are still unclear. In the current study, we found that miRNA-27a-3p (miR-27a) expression is downregulated during MAP infection both in vivo and in vitro. Moreover, miR-27a is also downregulated in toll-like receptor 2 (TLR2)-stimulated murine macrophages (RAW264.7 and bone marrow-derived macrophage). ELISA and real-time qRT-PCR results confirm that overexpression of miR-27a inhibited MAP-induced IL-10 production in macrophages and upregulated pro-inflammatory cytokines, while miR-27a inhibitor counteracted these effects. Luciferase reporter assay results revealed that IL-10 and TGF-β-activated protein kinase 1 binding protein 2 (TAB 2) are potential targets of miR- 27a. In addition, we demonstrated that miR-27a negatively regulates TAB 2 expression and diminishes TAB 2-dependent p38/JNK phosphorylation, ultimately downregulating IL-10 expression in MAP-infected macrophages. Furthermore, overexpression of miR- 27a significantly inhibited the intracellular survival of MAP in infected macrophages. Our data show that miR-27a augments antimicrobial activities of macrophages and inhibits the expression of IL-10, demonstrating that miR-27a regulates protective innate immune responses during MAP infection and can be exploited as a novel therapeutic target in the control of intracellular pathogens, including paratuberculosis. Correspondence: Xiangmei Zhou zhouxm@cau.edu.cn †These authors have contributed equally to this work. †These authors have contributed equally to this work. Specialty section: This article was submitted to Molecular Innate Immunity, a section of the journal Frontiers in Immunology Received: 03 July 2017 Accepted: 14 December 2017 Published: 10 January 2018 INTRODUCTION Paratuberculosis or Johne’s disease is characterized by chronic granulomatous enteritis predominantly observed in ruminants. It is caused by an obligate non-tuberculous Mycobacterium avium subspecies paratuberculosis (MAP), a member of M. avium com- plex. Paratuberculosis has a global distribution and is listed under the World Organization for Animal Health (OIE) Terrestrial Animal Health Code. Studies demonstrated that more than 50% of dairy cattle herds were positive for MAP antibodies in Europe and USA (1, 2). In addition, the USDA reports suggested an increasing trend of MAP infections in US dairy herds, from 21.6% in 1996 to 91.1% in 2007 (3). A recent study from northeastern part of China reported that the prevalence of paratuberculosis is 4.8% at the animal level and 50.0% at the herd level (4). Paratuberculosis inflicts heavy economic losses to the dairy industry in terms of weight loss, reduced milk production, infertility, culling of young animals, and the treatment of secondary diseases (2). Animals of all ages are susceptible to paratuberculosis, and the infection is acquired by fecal–oral route or ingestion of contaminated milk in young animals (5, 6). MAP infection poses a serious threat to human population apart from affecting a range of animal species. Contaminated food or water is the major source of MAP infection in humans (7). The relation between MAP and Morbus Crohn’s disease (CD) of humans was reported for the first time by Dalziel in 1913 (8). Various studies have documented the involvement of MAP in CD, but the causation of CD by MAP has not been confirmed (7, 9). In light of the current knowledge about MAP and its relationship with human disease, the majority of research- ers support the theory that MAP causes CD in some genetically susceptible human hosts, but additional proof is needed for the confirmation of MAP as a causative agent of CD (10). Mitogen-activated protein kinase p38 signaling pathway is implicated as an emerging signaling pathway involved in the pathogenesis of MAP in mononuclear phagocytic cells (24, 25). MAPK-p38 signaling can be pro-apoptotic or anti-apoptotic, depending on the cell type and stimulus strength (26). MAP infection leads to the activation of MAPK-p38 pathway by altering antibacterial activity of macrophages and promoting IL-10 pro- duction required for the survival of MAP in infected macrophages (15, 27). Furthermore, a set of six recombinant proteins of MAP promote MAPK-p38 phosphorylation and strongly induce IL-10 transcription in bovine monocyte-derived macrophages (16). INTRODUCTION p y p g ( ) MicroRNAs (miRNAs) are small non-coding endogenous RNA fragments, 21–25 nucleotides in length, regulating gene expression by binding at the 3′-UTR of the targeted mRNA genes (28, 29). Recent studies have determined the role of miRNAs as a key mediator of the host immune response to infection, mostly by regulating proteins involved in innate and adaptive immune pathways (30). In addition, miRNAs are involved in macrophage polarization and T cell and B cell differentiation (31, 32). The miR-24 and miR-27a modulate immune response by inhibiting Th2 regulation through targeting IL-4 and GATA binding protein 3 (GATA3) of mouse CD4 T cells (33). Early studies reported that miR-27a and miR-27b have antiviral activities against murine cytomegalovirus infection in different mouse cell lines (34). Ji and colleagues determined that both miR-27a and miR-27b target the 3′-UTR of the retinoid X receptor α and play a similar role in regulating fat metabolism and cell proliferation during rat hepatic stellate cell activation (35). In addition, Tak et al. reported that miR-27a/b similarly promotes mitochondrial membrane poten- tial and mitochondrial ATP level by targeting the 3′-UTR of the mitochondrial fission factor (36). Recent studies reported the pro- file of miRNA’s in MAP-infected animals that can be considered as an emerging strategy for the control of paratuberculosis. These miRNAs maintain their integrity during long-term storage and are considered an effective diagnostic tool (37, 38). However, the regulatory mechanism of innate immunity in paratuberculosis by miRNAs is unclear. IL-10 is an important mediator to suppress the early innate immune responses against paratuberculosis (20). Recent studies have demonstrated that miR-27a participates in i Macrophages are important cells in the regulation of pro- tective immune responses for the elimination of intracellular pathogens. On the other hand, they are the key cells involved in the pathogenesis of paratuberculosis by providing a suitable intracellular environment for the persistent survival and growth of MAP (11). Pattern recognition receptors (PRRs) are crucial in the phagocytosis of mycobacterium by mononuclear phagocytic cells such as macrophages (12). Toll-like receptors (TLRs) are a group of important PRRs involved in the activation of various signaling pathways in macrophages during mycobacterium infec- tion (13). It has been reported that the signaling cascades initiated by TLR2 are important for the pathogenesis of mycobacterium via inhibition of antimicrobial responses in macrophages (14). Citation: Hussain T, Zhao D, Shah SZA, Wang J, Yue R, Liao Y, Sabir N, Yang L and Zhou X (2018) MicroRNA 27a-3p Regulates Antimicrobial Responses of Murine Macrophages Infected by Mycobacterium avium subspecies paratuberculosis by Targeting Interleukin-10 and TGF-β-Activated Protein Kinase 1 Binding Protein 2. Front. Immunol. 8:1915. doi: 10.3389/fimmu.2017.01915 January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 1 Role of miR-27a in Paratuberculosis Hussain et al. the production of other cytokines, either by inverse-regulatory function or sharing a common signaling pathway in similar types of cells (18). IL-10 is one of the potent inhibitory cytokines that inversely regulate the expression of various pro-inflammatory mediators induced by TLR signaling cascades in immune- modulating cells (19). IL-10 also plays an important role in repairing tissue injury, but on the other hand it is involved in the persistent survival of mycobacterium in macrophages (20). IL-10 promotes the ability of Mycobacterium tuberculosis (M.tb) to evade immune responses and mediate long-term pathogenesis in the lungs (21). Furthermore, anti-IL-10 antibodies treatment increased immunological balance and controlled M.tb burden and pathogenesis in monkeys (22). In addition, IL-10 produc- tion is highly specific to MAP infection in comparison to other Mycobacterium species, suggesting that IL-10 detection can be used as a diagnostic tool in subclinical paratuberculosis (20). Overall, IL-10 promotes anti-inflammatory genes and is associ- ated with a dampening of the protective immune responses (23). Frontiers in Immunology \| www.frontiersin.org January 2018 \| Volume 8 \| Article 1915 INTRODUCTION Mitogen-activated protein kinase p38 (MAPK-p38) signaling pathway in bovine mononuclear phagocytic cells is activated by MAP via TLR2, resulting in overexpression of interleukin (IL)-10 and downregulation of IL-12 (15, 16). In addition, the binding of antigens to TLR2/4 and CD14 receptors of phagocytic cells induces an increased production of IL-10, while inhibiting the development of protective immune responses, ultimately pro- moting the survival and growth of intracellular pathogens (17). Cytokines are a diverse group of secretory proteins playing an important role in intercellular signaling mechanisms for the generation of immune responses in various diseases. As a result of the use of immunomodulatory drugs, some cytokines inhibit January 2018 \| Volume 8 \| Article 1915 2 Hussain et al. Role of miR-27a in Paratuberculosis JNK phosphorylation, subsequently downregulated IL-10 in MAP-infected macrophages. Our results provide a novel role for miR-27a as a potential biomarker that can be exploited for control strategies of MAP infection. the inflammatory responses of macrophages stimulated by LPS (39). In addition, miR-27a is one of the differentially regulating miRNAs in U937 macrophages activated with M.tb heat shock protein Hsp16.3 (40).hf The main objective of our study was to evaluate the effect of miR-27a on the production of IL-10, the IL-10 regulatory pathway—especially MAPK-p38—and macrophages antimicro- bial response for the control of MAP pathogenesis. We observed a gradual decrease in the level of miR-27a in MAP-infected murine macrophages and in the spleen and intestinal tissues of MAP-challenged mice. Overexpression of miR-27a decreased the production of IL-10 in MAP-infected macrophages. Also, miR- 27a significantly inhibited the expression of TGF-β-activated protein kinase 1 (TAK1) binding protein 2 (TAB 2)/3, an impor- tant component of the MAPK signaling pathway (Figure 1). Upregulation of miR-27a diminished TAB 2-dependent p38/ Ethics Statement Animal experiments were performed according to the protocols for the care of laboratory animals, Ministry of Science and Technology People’s Republic of China and approved animal care and use committee (IACUC) protocols at China Agricultural University of Beijing (20110611-01). All other experiments were carried out in accordance with the University Institutional Biosafety Committee approved protocol number 20110611-0. Figure 1 \| Schematic diagram depicting the role of miR-27a in Mycobacterium avium subspecies paratuberculosis (MAP)-mediated signaling pathways in macrophage. Mitogen-activated protein kinase (MAPK) pathway is triggered by toll-like receptors (TLRs), and TLR2/4 is more critical in signaling through various adopter proteins. A complex of protein integrations are involved in the downsignaling cascades consisting of TGF-β-activated protein kinase 1 (TAK1), TAK binding protein 1 (TAB 1), and TAB 2 or TAB 3. This interaction induces phosphorylation of TAK1, which results in the activation of the MAPK-p38 (mitogen-associated protein kinases). This leads to the activation and nuclear translocation of transcription factors, which play a critical role in the expression of pro-inflammatory and anti-inflammatory cytokines such as interleukin (IL)-10. MAP predominantly induces an immunosuppressive cytokine, IL-10, which subverts protective immune responses and promotes MAP growth and survival. MAP downregulates the expression of miR-27a, while miR27a inhibits the activation of the MAPK-p38 signaling pathway by targeting TAB 2. Furthermore, miR-27a negatively regulates the expression of IL-10 by directly targeting IL-10 mRNA at its 3′-UTR. Figure 1 \| Schematic diagram depicting the role of miR-27a in Mycobacterium avium subspecies paratuberculosis (MAP)-mediated signaling pathways in macrophage. Mitogen-activated protein kinase (MAPK) pathway is triggered by toll-like receptors (TLRs), and TLR2/4 is more critical in signaling through various adopter proteins. A complex of protein integrations are involved in the downsignaling cascades consisting of TGF-β-activated protein kinase 1 (TAK1), TAK binding protein 1 (TAB 1), and TAB 2 or TAB 3. This interaction induces phosphorylation of TAK1, which results in the activation of the MAPK-p38 (mitogen-associated protein kinases). This leads to the activation and nuclear translocation of transcription factors, which play a critical role in the expression of pro-inflammatory and anti-inflammatory cytokines such as interleukin (IL)-10. MAP predominantly induces an immunosuppressive cytokine, IL-10, which subverts protective immune responses and promotes MAP growth and survival. MAP downregulates the expression of miR-27a, while miR27a inhibits the activation of the MAPK-p38 signaling pathway by targeting TAB 2. Furthermore, miR-27a negatively regulates the expression of IL-10 by directly targeting IL-10 mRNA at its 3′-UTR. Quantitative Real-time PCR Total RNA (including miRNA) was extracted from RAW264.7 and BMDM cells with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s guidelines. To evaluate the expression levels of miRNAs and mRNAs, reverse transcription was done by using miRcute Plus miRNA First strand cDNA synthesis kits (obtained from Thermo Scientific and TIANGEN BIOTECH Beijing, China). To construct cDNA we used Thermo Scientific RevertAid First Strand cDNA synthesis kit according to the manufacturer’s instructions. Quantitative real-time PCR was performed by means of the miRcute miRNA qPCR detection kit (SYBR Green) for miRNA’s [miR-27a and internal references, small nucleolar S5 (mouse)]. Sybr Green Master Mix Kit (Vazyme Biotech Co., Ltd.) was used for the amplification of mRNA’s genes, including cytokines, by using the 700 Fast Real-Time PCR Systems (ViiA7 Real-time PCR, ABI). GAPDH was selected as housekeeping gene for the normalization of various cytokine Ethics Statement January 2018 \| Volume 8 \| Article 1915 Role of miR-27a in Paratuberculosis Hussain et al. RAW264.7 cells and sheep blood monocyte-derived macrophages show similar expression of TLR2 receptors (42). Furthermore, RAW264.7 cells are more differentiated than THP-1 cells, thus hav- ing higher phagocytic ability (43). Mouse bone marrow-derived macrophages were prepared and cultured as described previously (44). In brief, BMDMs were prepared from cells obtained from femurs, tibia, and humerus of 6- to 8-week-old C57BL/6J mice and cultured in RPMI supplemented with 10% heat-inactivated FBS (Hyclone) in the presence of 10 ng/ml of M-CSF (Peprotech) or GM-CSF (Peprotech) and 1% Penicillin-Streptomycin. At day four, non-adherent cells were collected and cultured for three more days in the presence of fresh RPMI containing M-CSF (10 ng/ml). These primary macrophages were cultured for 7–10 days in cell culture flasks. After that, the adherent BMDM cells (6–8 × 106 cells per dish) were collected and plated in 12-well cell culture plates for further experiments (45). RAW264.7 macrophages were taken from cold storage (−80°C) and cultured in cell culture flask in DMEM supplemented with 10% FBS and 1% penicillin– streptomycin for 3 days. The cells were transferred to 12-well cell culture plates 12–18 h before transfection. Reagents g MicroRNA miR-27a negative control, mimic, inhibitor control, and inhibitor were purchased from Shanghai GenePharma. Pam3Cys-Ser-(Lys)4 (ab142085), a TLR1/TLR2 agonist, was purchased from Abcam Biotechnology; lipofectamine 2000 from Invitrogen Thermo Fisher Scientific; and BBL Middlebrook OADC (211886) enrichment and mycobactin J from Becton, Dickinson and Sparks Company (USA), respectively. Macrophage colony-stimulating factor (M-CSF) was purchased from Peprotech Technology and total RNA extraction kit (ca. RN2802) from Aidlab Biotechnologies. The miR-27a (CD202-0033) and internal control S5 (CD202-0012) primers were from TIANGEN BIOTECH (Beijing). Primers for IL-1β, IL-6, IL-10, IL-12, interferon (IFN)-β, TNF-α, and GAPDH were purchased from Genewiz Technology. The calorimetric mouse IL-10, TNF-α, and IL-6 ELISA kit (KE10008, KE10007, and KE10002) were from Proteintech. Mouse IL-12 ELISA kit (ca. CSB-E07360m) was from Cusabio Co. Ltd. Rabbit polyclonal anti-p38 MAPK antibody (9211), rabbit polyclonal anti-phospho-p38 antibody (9211), rabbit monoclonal anti-phospho-JNK antibody (4668T), and rabbit monoclonal anti-phospho-ERK antibody (4370) were purchased from Cell Signaling Technology. Rabbit polyclonal anti-TAB 2 (MAP3K7IP2) polyclonal antibody (14410-1-Ap), rabbit polyclonal anti-GAPDH antibody (10494-1-AP), and goat anti-rabbit IgG (heavy and light chains) peroxidase-conjugated secondary antibody (SA000012) were purchased from Proteintech Biotechnology. Rabbit polyclonal anti-MAP3K7IP3 (TAB 3) antibody (bs5424R) was from Beijing Biosynthesis Biotechnology. Plasmid of mRNA for IL-10 and TAB 2 wild and mutant were constructed by Synbio Technology. Dual luciferase Reporter Assay kit (RFE: E1910, 205415) from Promega technology, USA. CellTiter 96Aqueous one solution cell proliferation assay kit (ca. G3580, 0000180214) from Promega technology, USA. Preparation of Bacterial Culture Preparation of Bacterial Culture We used two strains of MAP (MAP-0908 and MAP k-10). MAP-0908 was isolated from a naturally infected cattle farm in the Shandong province of China. Molecular characterization of the MAP-0908 strain was done at the Transmissible Spongiform Encephalopathy Diagnostic Laboratory (China Agricultural University Beijing). Organisms were confirmed as MAP by cultural characteristics, while further species-specific DNA sequences by a standard PCR were performed for molecular characterization (41). K-10 strain of MAP 35716 was a kind gift from Prof. Paul Barrow, from the University of Nottingham at Leicestershire, UK. A stock culture of both strains of MAP was maintained in Middlebrook 7H9 medium supplemented with 1% glycerol, 10% oleic-acid-dextrose-catalase (OADC), and mycobactin J (2.0 mg/l) (Allied Monitor, Fayette, MO, USA) at 37°C. The organisms were grown to a concentration of 108/ml for 2–3 weeks before being used for cell infection. Cells Transfection and Infection Cells were allowed to attach overnight in 12-well plates (2 × 105 cells in each well). The following day, miR-27a mimic, inhibi- tor, and control were individually transfected into RAW264.7 and BMDM cells using lipofectamine 2000 reagent, according to manufacturer’s instructions. After 6 h, the original medium was replaced with fresh medium supplemented with 10% FBS (46). After 48 h post-transfection, macrophages were infected at a multiplicity of infection (MOI) 20:1 (bacteria/cell) with live MAP in cell culture medium without antibiotic (47). Following incubation for 3 h at 37°C in 5% CO2, the supernatant was discarded and each well was washed three times with sterile phosphate-buffered saline (PBS) to remove non-adherent MAP bacilli. After washing, fresh DMEM medium supplemented with 10% serum was added for the specified time period. The cells were harvested 6 and 18 h postinfection for analysis of total protein, total RNA, and cell supernatant. All samples were stored at −80°C until further use. Isolation of Bone Marrow-Derived Macrophages (BMDMs) We used murine macrophages (RAW264.7 and BMDM) for our experiments. It has been demonstrated that MAP-infected January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 4 Role of miR-27a in Paratuberculosis Hussain et al. genes because it is considered one of the most stable endogenous controls (48). using twofold dilutions of the standard for each independent experiment. To obtain the optical density (OD), plate was read at 450 nm with a wavelength correction of 630 nm by using an ELISA plate reader. The ELISA results were obtained from two independent experiments. Samples were added into triplicate wells in each independent experiment. To normalize the expression of miRNAs, Ct values were obtained to calculate fold change for miRNA. A universal miR qPCR primer (possessing the binding site with universal adaptor primer, included in the kit), miRNA-27a primer, and 5S primer as internal reference (sequence complementary to the miRNA) were used to complete the real-time PCR reaction. miRNA qRT-PCR cycle parameters were as follows: 94°C for 2 min, followed by 45 cycles at 94°C for 20 s and 60°C for 34 s. The expression of mRNAs for various cytokines was determined by conducting qRT-PCR as described before (49). All the primers used in the present study for cytokines detection are mentioned in Table 1. To normalize the expression level of miRNAs and mRNAs, internal controls of S5 and GAPDH were used. The relative expression levels of miRNAs and mRNAs were determined as fold change, ΔCt values were obtained as follows: ΔCt = Ct of miRNAs or mRNAs − Ct of small nucleolar RNA S5or GAPDH. ΔΔCt values were obtained as follows: ΔΔCt = ΔCt of treated groups − ΔCt of untreated control groups. Fold change was calculated by the 2−ΔΔCt method (50, 51). Western Blot Analysis For total protein extraction, cells were washed twice with ice-cold PBS and homogenized with RIPA buffer containing a cocktail of protease and phosphatase inhibitor (Sigma Aldrich, St. Louis, MO, USA) for 20 min on ice. Afterward, samples were sonicated for 20 s and then centrifuged at 12,000 × g for 20 min at 4°C. The resulting supernatants were collected and boiled for 10 min after the addition of loading buffer (250 nM Tris HCl 6.8 pH, 10% sodium dodecyl sulfate, 0.5% bromophenol blue, 50% glycerol, and 0.5 M dithiothreitol) (52). Equal amounts of protein were separated by 12% or 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). After blocking with 5% skim milk prepared in Tris buffered saline Tween-20, membranes were incubated overnight at 4°C with primary anti- bodies. After incubation, membranes were rinsed three times in TBST and then incubated for 60 min at 37°C with HRP-labeled secondary antibodies. Signals were detected using an enhanced chemiluminescence detection kit (Bio-Rad, USA). Dual-Luciferase Reporter Assayh Dual-Luciferase Reporter Assay The binding elements for miR-27a at the 3′-UTR of IL-10 and TAB 2 mRNAs were obtained by PCR amplification using mouse genomic DNA as template and cloned into pMir-Reporter Luciferase vector (Synbio Tech). A mutant form of mouse IL-10 and TAB-2 3′-UTR was cloned into a pmirGLO vector by site-directed mutagenesis using a WT clone. HEK293 and BMDM cells were cultured in 24-well plates (2.5 × 104 cells/ well) overnight and transfected with 50 nM control mimics or 50 nM/ml miR-27a mimics using a Lipofectamine 2000. Six hours post-transfection, cells were again transfected with 100 ng pMir-Reporter constructs of IL-10 and TAB 2 wild- and mutant type of plasmid. Twenty-four hours after transfection, cells were lysed in lysis buffer, and a luciferase assay was performed using the Dual Luciferase reporter system (Promega), according to the manufacturer’s instructions (53). The results were obtained from three independent experiments, and all samples were collected in three replicates in each experiment. Table 1 \| Primers used in this study. Primer name and sequence (5′–3′) Interleukin (IL)-1β (forward) 5-AAGGAGAACCAAGCAACGACAAAATA-3 IL-1β (reverse) 5-TTTCCATCTTCTTCTTTGGGTATTGC-3 IL-6 (forward) 5-CCCAATTTCCAATGCTCTCCTA-3 IL-6 (reverse) 5-AGGAATGTCCACAAACTGATATGCT-3 IL-10 (forward) 5-AGCATTTGAATTCCCTGGGTGA-3 IL-10 (reverse) 5-CCTGCTCCACTGCCTTGCTCTT-3 IL-12 (forward) 5-CCAAATTACTCCGGACGGTTCAC-3 IL-12 (reverse) 5-CAGACAGAGACGCCATTCCACAT-3 TNF-α (forward) 5-AGAGCTACAAGAGGATCACCAGCAG-3 TNF-α (reverse) 5-TCAGATTTACGGGTCAACTTCACAT-3 Interferon (IFN)-β (forward) 5-AAGAGTTACACTGCCTTTGCCATC-3 IFN-β (reverse) 5-CACTGTCTGCTGGTGGAGTTCATC-3 GAPDH (forward) 5-CGACTTCAACAGCAACTCCCACTCTTCC-3 GAPDH (reverse) 5-TGGGTGGTCCAGGGTTTCTTACTCCTT-3 5-AAGGAGAACCAAGCAACGACAAAATA-3 5-TTTCCATCTTCTTCTTTGGGTATTGC-3 5-CCCAATTTCCAATGCTCTCCTA-3 5-AGGAATGTCCACAAACTGATATGCT-3 5-AGCATTTGAATTCCCTGGGTGA-3 5-CCTGCTCCACTGCCTTGCTCTT-3 5-CCAAATTACTCCGGACGGTTCAC-3 5-CAGACAGAGACGCCATTCCACAT-3 5-AGAGCTACAAGAGGATCACCAGCAG-3 5-TCAGATTTACGGGTCAACTTCACAT-3 5-AAGAGTTACACTGCCTTTGCCATC-3 5-CACTGTCTGCTGGTGGAGTTCATC-3 5-CGACTTCAACAGCAACTCCCACTCTTCC-3 5-TGGGTGGTCCAGGGTTTCTTACTCCTT-3 ELISA for Cytokinesh The concentration of cytokines in the cells supernatant was measured by using an ELISA kit. Cells supernatant was collected at various time intervals (6 and 18 h) post-MAP (k-10 and 0908 strains) infection of different miR-27a transfected groups. All reagents, standard, and samples were prepared according to the manufacturer’s protocols. Standards and samples (100 µl each) were added into appropriate wells of 96-well ELISA plates and incubated for 2 h at 37°C in a humid environment. After incubation, the liquid was discarded and the plates washed 4–5 times with washing buffer. Then, 100 µl of detection antibody solution was added into each well, incubated in humid condi- tions at 37°C for 1 h, and washed as previously. Then we added 100 µl of HRP-conjugated antibodies to each well for 40 min and incubated at 37°C in a humid environment and washed again. After that, we added 100 µl of TMB substrate for 10–30 min at room temperature in the dark. To stop the reaction, we added 100 µl stop solution to each well. A standard curve was obtained CFU Assay To assess bacterial viability, BMDM and RAW264.7 mac- rophages (2 × 105 cells in each well) were cultured in 12-well plates overnight and infected with MAP (0908 and k-10, MOI is 20:1) strains for 3 h and then washed thrice with warm PBS to remove extracellular bacteria. Thereafter, the infected cells were incubated for the indicated time periods and lysed at the specified time points with 0.1% Triton X 100. Appropriate dilutions were prepared for all transfected groups and plated on Middlebrook 7H10 agar plates supplemented with mycobactin-J and OADC in January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 5 Role of miR-27a in Paratuberculosis Hussain et al. miR-27a Attenuates the Regulation of IL-10 in Macrophages Infected with MAP It has been reported that TLR2-activated macrophages elevate the expression of IL-10 in response to MAP infection (20). IL-10 is one of the potent anti-inflammatory mediators that prevent tis- sue injury from excessive inflammatory reactions and also con- tribute to the pathogenesis of MAP (15, 54). More recently, it has been reported that miR-27a controls various targets to regulate the immune system (34, 39). We asked whether miR-27a partici- pates in inflammatory responses associated with MAP (0908 and k-10) infection. We found that the level of miR-27a in BMDM cells increased by transfecting the cells with miR-27a mimic. We also found that upregulation of miR-27a reduced MAP-induced IL-10 expression at both 6 and 18 h postinfection (Figure 3A) and simultaneously reduced L-10 protein levels in BMDM cells after infection with MAP (Figure 3B). A similar IL-10 expression was recorded in RAW264.7 macrophages treated with miR-27a at both protein and mRNA levels (Figure 3D). In contrast, an upregulation of IL-10 was observed at protein and mRNA levels in both BMDM and RAW264.7 macrophages transfected with miR-27a inhibitor (Figures 3E,F,H). To check the transfection efficiency, BMDM cells were transfected with miR-27a con- trol, mimic, and inhibitor for 48 h and then the expression of miR-27a observed with or without MAP infection (0908). The expression level of miR-27a was significantly higher in MAP- infected groups than in the MAP non-infected mimic-treated group (Figure 3C). In addition, miR-27a was downregulated in inhibitor treatment of both MAP-infected and MAP non-infected groups (Figure 3G). These data suggest that miR-27a negatively regulates the expression of IL-10 at both mRNA and protein levels in MAP-infected macrophages. RESULTS MAP Infection Decreases miR-27a Expression in Macrophages and Mice To determine whether miR-27a plays a role in the immune responses against MAP infection, we investigated the expression of miR-27a in vivo and in vitro. We infected murine BMDMs and RAW264.7 cells with two strains of MAP (MAP k-10 and MAP 0908) at MOI 20:1 (bacteria/cell) and with TLR2 agonist. To ensure maximum infection of macrophages with MAP, we conducted phagocytosis assays at different levels of MOI (5:1, 10:1, and 20:1) before establishing the dose of infectivity (Figures S1A,B in Supplementary Material). The expression of miR-27a was measured by using qRT-PCR analysis. MAP infection decreased the expression of miR-27a in a time-dependent manner in BMDM (Figures 2A,B) and RAW264.7 cells (Figures 2D,E). Additionally, TLR2 agonist also showed a decrease in the expres- sion of miR-27a in both types of macrophages (Figures 2C,F). The minimum level of miR-27a was detected at 12–24 h time intervals in all groups. Moreover, for in vivo determination of miR27a we challenged mice with MAP (0908) via intraperito- neal route. After 4 weeks postinfection, the expression level of miR-27a was significantly downregulated in different tissues of infected mice (Figures 2G–I). Various studies have reported that TLR2 receptors play an important role in the initiation of signal- ing cascades in macrophages to favor the persistent survival and growth of MAP (17). Our data suggest that downregulation of miR-27a in response to MAP strains and TLR2 treatment shares a common response in both types of macrophages and indicate that the expression of miR-27a in MAP- and TLR2-treated cells may share a common signaling mechanism. miR-27a Attenuates the Regulation of IL-10 in Macrophages Infected with MAP triplicate. Inoculated plates were incubated at 37°C, and colonies were counted 5 weeks after plating. Statistical Analysis All experiments were performed on three separate occasions. Data are expressed as mean ± SD. Student’s t-test was used for comparison between two groups. One-way ANOVA followed by Bonferroni’s multiple comparison tests was performed for mul- tiple-group comparisons. Statistical analysis was carried out by using GraphPad Prism 5 software (GraphPad, La Jolla, CA, USA). Densitometric analysis of digitalized images was done by using ImageJ software (National Institute of Health, Bethesda, MD, USA). p Values < 0.05 were considered statistically significant. Cell Viability Assay Cell viability was assessed by the MTS tetrazolium assay. BMDM and RAW264.7 cells were cultured in 96-well plates overnight before transfection. Cells were transfected with mimic control, mimic, inhibitor control, and inhibitor of miR-27a for 48 h. MTS reagent (20 µl) was added in each well, and then cells were incubated at 37°C for 3 h in a humidified, 5% CO2 atmosphere. OD was quantified by measuring the absorbance at 490 nm wavelength with an ELISA plate reader. CFU Assay We further analyzed the effects of miR-27a transfection on macrophage viability. BMDM and RAW264.7 cells were transfected with miR-27a (control, mimic, control-inhibitor, and inhibitor) for 48 h, and then cell viability was assessed by using the MTS assay (see Materials and Methods). Cell viability among all transfected groups showed no significant difference after miR-27a transfection (Figures S4A,B in Supplementary Material). miR-27a Upregulation Improves Macrophage Activation in Response to MAP Infection We have already shown that miR-27a is downregulated in MAP- infected macrophages and also observed the inhibitory effects of miR-27a on IL-10 expression. Previous studies reported that MAP inhibit the activation of macrophages by upregulating the expres- sion of IL-10 while downregulating various pro-inflammatory mediators (15). The downregulation of miR-27a in MAP infec- tion suggested that miR-27a could participate in the activation of macrophages. To test this hypothesis, we investigated whether miR27a participates in the expression of inflammatory mediators during MAP infection. First, we increased the level of miR-27a in BMDM cells by transfecting with miR-27a mimics for 48 h, and then cells were infected with MAP 0908 strain. After 6 h postin- fection, we found a high level of pro-inflammatory cytokines, January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 6 Role of miR-27a in Paratuberculosis Hussain et al. Figure 2 \| Mycobacterium avium subspecies paratuberculosis (MAP) infection decreases miR-27a expression in macrophages and mice. (A,B) Bone marrow- derived macrophage (BMDM) and (D,E) RAW264.7 cells were infected with MAP (k-10 and 0908) for the indicated time period, and miR-27a expression was subsequently evaluated by using qRT-PCR. (C) RAW264.7 and (F) BMDMs were stimulated with 1 µg/ml Pam3Cys-Ser-(Lys)4 [PAM toll-like receptor (TLR)1/2 agonist] for the indicated time period, and miR-27a expression was examined using qRT-PCR. (G–I) The expression levels of miR-27a were measured in the intestine (ilium) (G), spleen (H), and liver (I) of negative control or MAP (0908)-infected C57BL/6 mice by qRT-PCR analysis. All data above represent mean ± SD for three independent experiments. p < 0.05 and *p < 0.001. Figure 2 \| Mycobacterium avium subspecies paratuberculosis (MAP) infection decreases miR-27a expression in macrophages and mice. (A,B) Bone marrow- derived macrophage (BMDM) and (D,E) RAW264.7 cells were infected with MAP (k-10 and 0908) for the indicated time period, and miR-27a expression was subsequently evaluated by using qRT-PCR. (C) RAW264.7 and (F) BMDMs were stimulated with 1 µg/ml Pam3Cys-Ser-(Lys)4 [PAM toll-like receptor (TLR)1/2 agonist] for the indicated time period, and miR-27a expression was examined using qRT-PCR. (G–I) The expression levels of miR-27a were measured in the intestine (ilium) (G), spleen (H), and liver (I) of negative control or MAP (0908)-infected C57BL/6 mice by qRT-PCR analysis. All data above represent mean ± SD for three independent experiments. p < 0.05 and *p < 0.001. miR-27a Upregulation Improves Macrophage Activation in Response to MAP Infection including IL-1β, IL-6, IL-12, IFN-β, and TNF-α in BMDM at both translational and transcriptional levels (Figures 4A,B). Similar results were observed for k-10 strain of MAP in infected mac- rophages (Figures 4C,D). Consistent with the findings in BMDM cells, upregulation of miR-27a also enhanced MAP-induced expression of IL-1β, IL-6, IL-12, IFN-β, and TNF-α in RAW264.7 cells (Figures 4E–H). In addition, we found that miR-27a also upregulates the expression of pro-inflammatory cytokines after 18 h of MAP infection, compared to miR-27a control (Figure S2 in Supplementary Material). These data suggest that miR-27a is a positive regulator for the induction of inflammatory cytokines in macrophages infected with MAP. and then asked whether downregulation of miR-27a will affect macrophage response against MAP. The level of miR-27a decreased in BMDM cells by transfecting the cells with miR-27a inhibitor for 48 h and then infected them with MAP 0908. After 6 h of infection, we found significantly decreased levels of pro- inflammatory cytokines, including IL-1β, IL-6, IL-12, IFN-β, and TNF-α induced by MAP 0908 in BMDM at both translational and transcriptional levels (Figures 5A,B). Similar findings were observed for MAP k-10-infected macrophages (Figures 5C,D). In addition, the downregulation of miR-27a also inhibited various cytokines expression in RAW264.7 macrophages after infection with MAP 0908 and k-10 strains (Figures 5E–H). Furthermore, we found that downregulation of miR-27a also inhibited the expression of pro-inflammatory cytokines after 18 h of MAP infection (Figure S3 in Supplementary Material). These findings suggest that miR-27a regulates macrophage activation by abrogating the anti-inflammatory environment created by MAP in infected macrophages for prolonged survival and growth. miR-27a Modulates MAPK Signaling Cascade by Targeting TAB 2 and TAB 3 in MAP-Infected Macrophagest 2 mRNA (Figures 7F,G). Additionally, qRT-PCR data demon- strated that BMDM transfected with miR-27a mimic significantly increased the expression levels of miR-27a. On the other hand, BMDMs transfected with miR-27a inhibitor significantly reduced the expression level of miR-27a (Figure 7C). These findings sug- gest that both IL-10 and TAB 2 are direct targets of miR-27a. p g Complex signaling cascades are triggered by MAP after interac- tion with TLR2/4 for the activation of MAPK-p38 pathway (15). TAB 1, TAB 2, and TAB 3 are involved in the activation of TAK1 (55). Various kinases are activated by downstream signaling of TAK1 including MAPKs. We investigated whether miR-27a par- ticipates in the regulation of signaling cascades of MAPK pathway in MAP-infected macrophages. BMDM cells were transfected with miR-27a control, mimic, and inhibitor. Forty-eight hours after transfection, cells were infected with MAP 0908 strain, and the expression of various proteins of the MAPK pathway was determined by immunoblot assays. As shown in Figure 6A, upregulation of miR-27a diminished MAP (0908) induced TAB 2 (Figure 6B) and TAB 3 (Figure 6C) expression, while down- regulation of miR-27a counteracted these effects. Furthermore, we examined whether miR-27a affects the components of MAPK pathway that regulate the expression of IL-10 during MAP infec- tion. We found that upregulation of miR-27a reduced the expres- sion of phosphorylated p38 (Figure 6D) and p-JNK (Figure 6E), while downregulation of miR-27a significantly promoted the expression of p-p38 and p-JNK at all time intervals. However, no significant difference was observed in p-ERK (Figure 6F) in BMDM cells at various time intervals among all treated groups after MAP (0908) infection. These findings suggest that miR- 27a modulates MAPK signaling pathways activated by MAP in infected macrophages. miR-27a Promotes Antimicrobial Properties of Macrophages and Inhibits Intracellular Survival of MAP Mycobacterium avium subspecies paratuberculosis adopts vari- ous strategies for intracellular survival in mononuclear phago- cytic cells. The activation of MAPK-p38 signaling pathway and upregulation of IL-10 is one of the important strategies of MAP survival and pathogenesis (58, 59). We already found that miR- 27a positively regulates the expression of pro-inflammatory cytokines in MAP-infected macrophages and that overexpres- sion of miR-27a promotes macrophage activation by inhibition of IL-10. We next investigated whether miR-27a affects the intracellular survival of MAP. To test this hypothesis, BMDM cells were transfected with miR-27a control, mimic, and inhibi- tor. Upregulation of miR-27a reduced the survival of MAP (0908) in BMDM cells, while miR-27a inhibitor counteracted that effect (Figure 8A). Additionally, transfection of RAW264.7 cells with miR-27a mimic showed a clearer difference in the intracellular survival of MAP than cells transfected with miR- 27a inhibitor (0908) (Figure 8C). We observed similar results between miR-27a mimic and inhibitor for MAP (k-10)-infected macrophages (Figures 8B,D). Although the difference for total viable bacterial count between the control and miR-27a inhibitor was less evident, there was a clear significant difference between miR-27a mimic and inhibitor group (Figures 8B,C). These results indicate that miR-27a inhibits the intracellular survival of MAP by promoting bactericidal activities of macrophages. Downregulation of miR-27a Inhibits Inflammatory Responses of MAP-Infected Macrophages We observed that mimic treatment of miR-27a promoted inflammatory responses of macrophages infected with MAP January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 7 Role of miR-27a in Paratuberculosis Hussain et al. Figure 3 \| Continued Hussain et al. Role of miR-27a in Paratuberculo igure 3 \| Continued Figure 3 \| Continued Figure 3 \| Continued Frontiers in Immunology \| www.frontiersin.org January 2018 \| Volume 8 \| Article 1915 8 Role of miR-27a in Paratuberculosis Hussain et al. Figure 3 \| miR-27a attenuates regulation of interleukin (IL)-10 in macrophages infected by Mycobacterium avium subspecies paratuberculosis (MAP). Bone marrow-derived macrophage (BMDM) (A,B) and RAW264.7 (D) cells were transfected with 50 nM control mimics or miR-27a mimics. After 48 h, cells were infected with MAP (0908 or K-10) strain for 6 and 18 h. The mRNA and protein levels of the anti-inflammatory cytokine IL-10 were calculated by using qRT-PCR and ELISA. BMDM (E,F) and RAW264.7 (H) cells were transfected with 50 nM control inhibitors or miR-27a inhibitors. After 48 h, cells were infected with MAP (0908 or K-10) strain for 6 and 18 h. The mRNA and protein levels of the anti-inflammatory cytokine IL-10 were determined by qRT-PCR and ELISA. (C,G) BMDMs were transfected with miR-27a control, mimic (C), or inhibitor (G) and then infected with MAP. The expression levels of miR-27a were determined by qRT-PCR analysis. Statistical analysis was carried out by using one-way ANOVA followed by Bonferroni’s multiple comparison tests. p < 0.05, p < 0.001, and p < 0.001. IL-10 and TAB 2 Are Direct Targets of miR-27a To clarify whether miR-27a targets IL-10 alone or IL-10 and TAB 2 together at their 3′-UTR region, we used TargetScan and miRanda algorithm to predict miRNA targets as previously described (56, 57). We identified a putative binding site of miR- 27a/b at the 3′-UTR of IL-10 and TAB 2 transcript in mice and also in other species (Figure 7A). Luciferase reporter vectors were constructed for wild and mutant 3′-UTR for IL-10 and TAB 2 target site of miR-27a (Figure 7B). We found that upregula- tion of miR-27a reduced the activity of the luciferase reporter in BMDM and HEK293 cells containing IL-10 wild-type reporter, while miR-27a failed to inhibit luciferase activity in cells having IL-10 mutant-type reporter (Figures 7D,E). We also found that miR-27a decreased the activity of the luciferase reporter that contained the wild-type 3′-UTR of TAB 2 mRNA in both BMDM and HEK293 cells. However, miR-27a had no effect on the activity of a luciferase reporter that contained the mutant 3′-UTR of TAB DISCUSSION Emerging evidence suggested that miRNAs are key regulators of genes involved in immune cell differentiation, effectors func- tions, and modulation of the host defense mechanism against pathogenic infection (60, 61). However, the biological role of miRNAs in the specific context of macrophage activation during MAP infection remains poorly understood. Macrophages are not only involved in the elimination of intracellular pathogens but also provide a niche for the survival of MAP organisms. Classically activated macrophages (M1) are pro-inflammatory and are involved in tissue damage in high population numbers (62), while alternatively activated macrophages (M2) are poorly January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 9 Role of miR-27a in Paratuberculosis Hussain et al. re 4 \| miR-27a upregulation improves macrophage activation in response to Mycobacterium avium subspecies paratuberculosis (MAP) infection. (A–D) Bone w-derived macrophage (BMDM) cells were transfected with 50 nM miR-27a control or miR-27a mimic. After 48 h, cells were infected by MAP (0908 or k-10) strain for he mRNA and protein levels of pro-inflammatory cytokines interleukin (IL)-1β, IL-6, IL-12, interferon (IFN)-β, and TNF-α were determined by qRT-PCR (A,C), and IL-6, and TNF-α were determined by ELISA (B,D). (E–H) RAW264.7 cells were transfected with 50 nM miR-27a control or miR-27a mimic. After 48 h, cells were infected AP (0908 or k-10) strain for 6 h. The mRNA and protein levels of pro-inflammatory cytokines IL-1β, IL-6, IL-12, IFN-β, and TNF-α were determined by qRT-PCR (E,G), L-6, IL-12, and TNF-α were determined by ELISA (F,H). Data represent mean ± SD from three independent experiments. p < 0.05, p < 0.001, and p < 0.001. n et al. Role of miR 27a in Paratuberculosis Figure 4 \| miR-27a upregulation improves macrophage activation in response to Mycobacterium avium subspecies paratuberculosis (MAP) infection. (A–D) Bone marrow-derived macrophage (BMDM) cells were transfected with 50 nM miR-27a control or miR-27a mimic. After 48 h, cells were infected by MAP (0908 or k-10) strain for 6 h. The mRNA and protein levels of pro-inflammatory cytokines interleukin (IL)-1β, IL-6, IL-12, interferon (IFN)-β, and TNF-α were determined by qRT-PCR (A,C), and IL-6, IL-12, and TNF-α were determined by ELISA (B,D). (E–H) RAW264.7 cells were transfected with 50 nM miR-27a control or miR-27a mimic. After 48 h, cells were infected by MAP (0908 or k-10) strain for 6 h. DISCUSSION Figure 6 \| miR-27a modulates mitogen-activated protein kinase (MAPK) signaling cascades by targeting TGF-β-activated protein kinase 1 binding protein 2 (TAB 2) and TAB 3 in Mycobacterium avium subspecies paratuberculosis (MAP)-infected macrophages. (A) Bone marrow-derived macrophage (BMDM) cells were transfected with 50 nM miR-27a control, mimic, and inhibitor for 48 h and then infected with MAP (0908) at multiplicity of infection (MOI) of 20:1 (bacteria/cell) for 30 min, 6 h, and 18 h. The expression levels of TAB 2, TAB 3, p-P38, p-JNK, and p-ERK were detected by Western blot. (B–F) Western blot analysis of protein levels of TAB 2 (B), TAB 3 (C), p-P38 (D), p-JNK (E) s, and p-ERK (F) normalized to GAPDH after transfection with miR-27a control, mimic, and inhibitor for the indicated times. Data represent mean ± SD from three independent experiments. p < 0.05, p < 0.001, and p < 0.001. Figure 6 \| miR-27a modulates mitogen-activated protein kinase (MAPK) signaling cascades by targeting TGF-β-activated protein kinase 1 binding protein 2 (TAB 2) and TAB 3 in Mycobacterium avium subspecies paratuberculosis (MAP)-infected macrophages. (A) Bone marrow-derived macrophage (BMDM) cells were transfected with 50 nM miR-27a control, mimic, and inhibitor for 48 h and then infected with MAP (0908) at multiplicity of infection (MOI) of 20:1 (bacteria/cell) for 30 min, 6 h, and 18 h. The expression levels of TAB 2, TAB 3, p-P38, p-JNK, and p-ERK were detected by Western blot. (B–F) Western blot analysis of protein levels of TAB 2 (B), TAB 3 (C), p-P38 (D), p-JNK (E) s, and p-ERK (F) normalized to GAPDH after transfection with miR-27a control, mimic, and inhibitor for the indicated times. Data represent mean ± SD from three independent experiments. p < 0.05, p < 0.001, and p < 0.001. important PRRs of phagocytic cells activating MAPK pathway after ligation with the cell wall component Man-LAM and the 19 kDa lipoprotein of MAP (15). Our results revealed that miR- 27a is downregulated in BMDM and RAW264.7 cells activated with MAP and TLR2 agonist. Similarly, the study of Xie et al. reported that miR-27a was downregulated in mouse BMDM and human macrophages stimulated by TLR4 and TLR2 agonists (51). Therefore, the regulation of miR-27a appears to be fundamental in keeping macrophage modulation and inflammatory responses. microbicidal (63, 64). In infectious diseases, the control of tissue inflammation might lead to uncontrolled pathogen multiplica- tion (65). DISCUSSION The mRNA and protein levels of pro-inflammatory cytokines IL-1β, IL-6, IL-12, IFN-β, and TNF-α were determined by qRT-PCR (E,G), and IL-6, IL-12, and TNF-α were determined by ELISA (F,H). Data represent mean ± SD from three independent experiments. p < 0.05, p < 0.001, and p < 0.001. Frontiers in Immunology \| www.frontiersin.org January 2018 \| Volume 8 \| Article 1915 10 Role of miR-27a in Paratuberculosis Hussain et al. wnregulation of miR-27a inhibits inflammatory responses of Mycobacterium avium subspecies paratuberculosis (MAP)-infected macrophages. (A–D) derived macrophage (BMDM) cells were transfected with 50 nM control inhibitors or miR-27a inhibitors. After 48 h, cells were infected by MAP (0908 or 6 and 18 h. The mRNA and protein levels of pro-inflammatory cytokines interleukin (IL)-1β, IL-6, IL-12, interferon (IFN)-β, and TNF-α were qRT-PCR (A,C), and IL-6, IL-12, and TNF-α were determined by ELISA (B,D). (E–H) RAW264.7 cells were transfected with 50 nM control inhibitors bitors. After 48 h, cells were infected by MAP (0908 or k-10) strain for 6 and 18 h. The mRNA and protein levels of pro-inflammatory cytokines IL-1β, -β, and TNF-α were determined by (E,G) qRT-PCR, and IL-6, IL-12, and TNF-α were determined by (F,H) ELISA. Similar results were observed in ent experiments. p < 0.05, p < 0.01, and p < 0.001. Figure 5 \| Downregulation of miR-27a inhibits inflammatory responses of Mycobacterium avium subspecies paratuberculosis (MAP)-infected macrophages. (A–D) Bone marrow-derived macrophage (BMDM) cells were transfected with 50 nM control inhibitors or miR-27a inhibitors. After 48 h, cells were infected by MAP (0908 or k-10) strain for 6 and 18 h. The mRNA and protein levels of pro-inflammatory cytokines interleukin (IL)-1β, IL-6, IL-12, interferon (IFN)-β, and TNF-α were determined by qRT-PCR (A,C), and IL-6, IL-12, and TNF-α were determined by ELISA (B,D). (E–H) RAW264.7 cells were transfected with 50 nM control inhibitors or miR-27a inhibitors. After 48 h, cells were infected by MAP (0908 or k-10) strain for 6 and 18 h. The mRNA and protein levels of pro-inflammatory cytokines IL-1β, IL-6, IL-12, IFN-β, and TNF-α were determined by (E,G) qRT-PCR, and IL-6, IL-12, and TNF-α were determined by (F,H) ELISA. Similar results were observed in three independent experiments. p < 0.05, p < 0.01, and p < 0.001. January 2018 \| Volume 8 \| Article 1915 11 Frontiers in Immunology \| www.frontiersin.org Role of miR-27a in Paratuberculosis Hussain et al. Frontiers in Immunology \| www.frontiersin.org DISCUSSION Macrophage activation in M1 or M2 must be tightly regulated and can be used as a therapeutic target. Recent studies mentioned that miR-27a and miR-27b play an important role in macrophage polarization by targeting the 3′-UTR of IRF4 (23). In addition, miR-27a is a member of the miR-23a/27a/24-2 gene cluster, promoting the expression of M1 cytokines while reducing the expression of M2 cytokines, suggesting that it plays an impor- tant role in the activation and polarization of macrophages (39). Similarly, our results show that miR-27a (miR-27a-3p; a mature form of miR-27a) promotes pro-inflammatory cytokines while inhibiting anti-inflammatory ones and modulating macrophage activation after MAP infection. l Farrell and colleagues demonstrated the existence of novel miRNAs in MAP-infected bovine serum and explored the potential of miRNAs as biomarkers for infectious diseases in cattle (37). Our data provide confirmation of the important role of miR-27a in the regulation of IL-10 expression in MAP-infected macrophages. We found that enforced expression of miR-27a suppressed IL-10 protein expression, whereas transfection of cells with an inhibitor of miR-27a resulted in increased IL-10 expression in MAP-infected macrophages. Similarly, it has been shown that miR-27a inhibits the production of IL-10 at both mRNA and protein levels and enhanced the signaling of NF-kB for cytokines production in hypoxia and reperfusion injury (67). Liu and colleagues showed that miR-98 inhibits TLR4-dependent IL-10 production in LPS-stimulated macrophages (68). For the progression of pathogenesis, MAP overexpresses the production of IL-10 that gradually shifts protective Th1-type cell-mediated t We found that the expression of miR-27a was downregulated by MAP (0908 and k-10 strain of MAP) both in vitro and in vivo. Our results are in accordance with Ma and colleagues findings, where they demonstrated that miR-23a/27a/24-2 was down- regulated in mouse peritoneal macrophages stimulated with LPS (39). Similarly, murine cytomegalovirus downregulated the expression of miR-27a in various mouse cell lines and primary macrophages (34). Several miRNAs are induced by TLR signaling in macrophages and target the 3′-UTR of mRNAs encoding key molecules involved in macrophage polarization (66). TLR2 and TLR4 are January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 12 Role of miR-27a in Paratuberculosis Hussain et al. Figure 7 \| Interleukin (IL)-10 and TGF-β-activated protein kinase 1 binding protein 2 (TAB 2) are direct targets of miR-27a. DISCUSSION (A) The mature sequences of miR-27a/b and the conserved sequences of the 3′-UTR of IL-10 and TAB 2 from various species are illustrated. The seed sites of miR-27a/b and its binding sites at the 3′-UTR of IL-10 and TAB 2 are shown in blue and red color, respectively. (B) Schematic illustration of the miR-27a targeting site at 3′-UTR of the mouse IL-10 and TAB 2 gene. The created mutation is represented in green color. (C) Bone marrow-derived macrophage (BMDM) cells were transfected with 50 nM miR-27a control, mimic, and inhibitor. Forty-eight hours after transfection, cells were lysed and the expression levels of miR-27a were determined by real-time PCR assay. Data represent mean ± SD from two independent experiments. (D) BMDM cells were transfected with 50 nM miR-27a control, miR-27a mimic and a wild-type (IL-10 WT) or mutant IL-10 3′-UTR (IL-10 Mut), and (F) wild-type (TAB 2 WT) or mutant TAB 2 3′-UTR (TAB 2 Mut) luciferase reporter plasmid. Luciferase activities of the BMDM cells were assessed 24 h after transfection. (E) HEK-293 cells were transfected with 50 nM miR-27a control, miR-27a mimic and a wild-type (IL-10 WT) or mutant IL-10 3′-UTR (IL-10 Mut) and (G) wild-type (TAB 2 WT) or mutant TAB 2 3′-UTR (TAB 2 Mut) luciferase reporter plasmid. Luciferase activities of the HEK-293 cells were assessed 24 h after transfection. Similar results were observed in three independent experiments. p < 0.05, p < 0.01, and p < 0.001. Hussain et al. Role of miR-27a in Paratuberculosis Figure 7 \| Interleukin (IL)-10 and TGF-β-activated protein kinase 1 binding protein 2 (TAB 2) are direct targets of miR-27a. (A) The mature sequences of miR-27a/b and the conserved sequences of the 3′-UTR of IL-10 and TAB 2 from various species are illustrated. The seed sites of miR-27a/b and its binding sites at the 3′-UTR of IL-10 and TAB 2 are shown in blue and red color, respectively. (B) Schematic illustration of the miR-27a targeting site at 3′-UTR of the mouse IL-10 and TAB 2 gene. The created mutation is represented in green color. (C) Bone marrow-derived macrophage (BMDM) cells were transfected with 50 nM miR-27a control, mimic, and inhibitor. Forty-eight hours after transfection, cells were lysed and the expression levels of miR-27a were determined by real-time PCR assay. Data represent mean ± SD from two independent experiments. DISCUSSION January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 13 Role of miR-27a in Paratuberculosis Hussain et al. Figure 8 \| miR-27a promotes antimicrobial properties of macrophages and inhibits intracellular survival of Mycobacterium avium subspecies paratuberculosis (MAP). Bone marrow-derived macrophages (BMDM) (A,B) and RAW 264.7 (C,D) cells were transfected with 50 nM miR-27a control, mimic, and inhibitor. Forty-eight hours after transfection cells were infected by MAP (0908 and k-10) strains at multiplicity of infection (MOI) of 20:1 (bacteria/cells) for 18 h. After that, cells were lysed, and intracellular survival of MAP was determined by CFU. Mean ± SD values were obtained from three independent experiments. p < 0.05 and *p < 0.01. Figure 8 \| miR-27a promotes antimicrobial properties of macrophages and inhibits intracellular survival of Mycobacterium avium subspecies paratuberculosis (MAP). Bone marrow-derived macrophages (BMDM) (A,B) and RAW 264.7 (C,D) cells were transfected with 50 nM miR-27a control, mimic, and inhibitor. Forty-eight hours after transfection cells were infected by MAP (0908 and k-10) strains at multiplicity of infection (MOI) of 20:1 (bacteria/cells) for 18 h. After that, cells were lysed, and intracellular survival of MAP was determined by CFU. Mean ± SD values were obtained from three independent experiments. p < 0.05 and *p < 0.01. anti-inflammatory mediator (72) that blocks the release of vari- ous pro-inflammatory cytokines and promotes the survival of MAP in mononuclear phagocytic cells (73). Our data also suggest that mimic treatment of miR-27a inhibits the expression of IL-10, while promoting the regulation of various pro-inflammatory cytokines in murine macrophages results in line with the study reported by Xie and colleagues (51). response to a humoral Th2 response (69). Cho and colleagues reported that miR-24 and miR-27a collectively inhibit the dif- ferentiation of CD4 T cells into Th2 cells by targeting IL-4 and GATA3 (33). Similarly, we found that overexpression of miR-27a modulated the immune responses during MAP infection by targeting IL-10 at the 3′-UTR or by inhibiting the regulatory pathway for the expression of IL-10. Previously, it was shown that the regulation of IL-10- and IL-10-dependent MAPK-p38 is required for the control of persistent survival and growth of MAP in infected macrophages (16, 20).hf Mitogen-activated protein kinase signaling pathways are playing an important role in the regulation of cell death and cell survival depending on the type of stimulus and type of cells (26). DISCUSSION (D) BMDM cells were transfected with 50 nM miR-27a control, miR-27a mimic and a wild-type (IL-10 WT) or mutant IL-10 3′-UTR (IL-10 Mut), and (F) wild-type (TAB 2 WT) or mutant TAB 2 3′-UTR (TAB 2 Mut) luciferase reporter plasmid. Luciferase activities of the BMDM cells were assessed 24 h after transfection. (E) HEK-293 cells were transfected with 50 nM miR-27a control, miR-27a mimic and a wild-type (IL-10 WT) or mutant IL-10 3′-UTR (IL-10 Mut) and (G) wild-type (TAB 2 WT) or mutant TAB 2 3′-UTR (TAB 2 Mut) luciferase reporter plasmid. Luciferase activities of the HEK-293 cells were assessed 24 h after transfection. Similar results were observed in three independent experiments. p < 0.05, p < 0.01, and p < 0.001. Figure 7 \| Interleukin (IL)-10 and TGF-β-activated protein kinase 1 binding protein 2 (TAB 2) are direct targets of miR-27a. (A) The mature sequences of miR-27a/b and the conserved sequences of the 3′-UTR of IL-10 and TAB 2 from various species are illustrated. The seed sites of miR-27a/b and its binding sites at the 3′-UTR of IL-10 and TAB 2 are shown in blue and red color, respectively. (B) Schematic illustration of the miR-27a targeting site at 3′-UTR of the mouse IL-10 and TAB 2 gene. The created mutation is represented in green color. (C) Bone marrow-derived macrophage (BMDM) cells were transfected with 50 nM miR-27a control, mimic, and inhibitor. Forty-eight hours after transfection, cells were lysed and the expression levels of miR-27a were determined by real-time PCR assay. Data represent mean ± SD from two independent experiments. (D) BMDM cells were transfected with 50 nM miR-27a control, miR-27a mimic and a wild-type (IL-10 WT) or mutant IL-10 3′-UTR (IL-10 Mut), and (F) wild-type (TAB 2 WT) or mutant TAB 2 3′-UTR (TAB 2 Mut) luciferase reporter plasmid. Luciferase activities of the BMDM cells were assessed 24 h after transfection. (E) HEK-293 cells were transfected with 50 nM miR-27a control, miR-27a mimic and a wild-type (IL-10 WT) or mutant IL-10 3′-UTR (IL-10 Mut) and (G) wild-type (TAB 2 WT) or mutant TAB 2 3′-UTR (TAB 2 Mut) luciferase reporter plasmid. Luciferase activities of the HEK-293 cells were assessed 24 h after transfection. Similar results were observed in three independent experiments. p < 0.05, p < 0.01, and *p < 0.001. Frontiers in Immunology \| www.frontiersin.org DISCUSSION Zheng and colleagues reported that the silencing of miR-195 reduced lung and liver injuries and improved the survival of septic mice (88). In addition, the downregulation of miR-328-3p by intratracheal administration of antagomirs of miR-328-3p enhanced bacterial killing when mice were infected with non-typeable Haemophilus influenza (89). In the present study, we investigated the role of miR-27a in macrophages infected with MAP in vitro, but further studies are required to explore the role of miR-27a in an in vivo model of MAP infection. the expression of TAB 2 and TAB 3 and has a putative binding site at the 3′-UTR of TAB 2. These observations prompted us to examine whether miR-27a influences the activation of the MAPK pathway during MAP infection. We found that miR-27a mimic-transfected BMDM cells after infection with MAP (0908) significantly reduced the phosphorylation of p38 and JNK at different times, while no significant difference was recorded for phosphorylated ERK. Previously, it was reported that the inhibition of MAPK-p38 activity promotes the microbicidal ability of bovine monocyte-derived macrophages to eliminate intracellular bacteria (78). In addition, MAP recombinant proteins are involved in the activation of MAPK-p38 in bovine macrophages and resulted in enhanced production of IL-10 (16). We have shown here that miR-27a inhibits IL-10 expression by targeting mRNA at the 3′-UTR of IL-10 as well as inhibition of p38 phosphorylation by targeting TAB 2 mRNA at its 3′-UTR in MAP-infected macrophages. Our results are in line with the finding of Pan and colleagues’ that miR-27a inhibits JNK/p38 expression by targeting MAP2K4 mRNA at its 3′-UTR in human osteosarcoma cells (46). Macrophages are not only killing mycobacteria by promot- ing inflammatory mediators but also provide a favorable hiding site for intracellular multiplication and survival of pathogenic mycobacteria (79). Therefore, macrophage activation should be tightly regulated for the development of protective immune responses against mycobacterial infection. An early study reported that miR-27a upregulated the expression of TNF-α in BMDM cells treated with LPS (51), while TNF-α played a crucial role in macrophage activation in killing intracellular mycobacteria (80), indicating that miR-27a might be required in the macrophage-mediated immune defense against infection. We demonstrated that miR-27a enhances pro-inflammatory cytokines, while decreasing IL-10 production and the intra- cellular survival of MAP, indicating that miR-27a promoted macrophage-mediated bacterial elimination. DISCUSSION The function of the MAPK-p38 branch in the regulation of pro- and anti-inflammatory genes expression is distinct in various cell types (74). In MAP infection, the activation of the MAPK-p38 pathway promotes the production of IL-10 and inhibits anti- microbial activity of macrophages (59). Furthermore, other studies have reported that p38, JNK, and ERK play a dual role in regulating the expression of cytokines, especially IL-12 and IL-10 (75). TAB 2 and its closely related protein, TAB 3, are essential for the activation of TAK1, a major downstream molecule for IkB-kinases and MAPKs including p38 and JNK (76). Previously, it has been shown that TAB 1 and TAB 2 are essential for the survival of LPS-activated macrophages by inhibiting apoptotic cell death via preventing reactive oxygen species production (77). Our analysis has revealed that miR-27a significantly inhibited The observed effects on IL-10 expression prompted us to examine whether miR-27a influences the activation of mac- rophages induced by MAP infection. We demonstrated that the expressions of pro-inflammatory cytokines were upregu- lated by overexpression of miR-27a, whereas these cytokines were reduced by inhibition of miR-27a by transfection with miR-27a inhibitor. Our work is in line with a previous report showing the effects of miR-23a/27a/24-2 on the expression of pro-inflammatory cytokines in murine macrophages (39). In addition, the expressions of various pro-inflammatory cyto kines were suppressed by IL-10 (70). Similarly, the blocking of IL-10 resulted in overexpression of IL-12 and TNF-α in bovine monocyte-derived macrophages infected with MAP or purified protein derivative (or Johnin) (71). Overall, IL-10 is an important January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 14 Role of miR-27a in Paratuberculosis Hussain et al. Similarly, MAP inhibits the expression of multiple miRNAs, which may contribute to MAP evasion strategies from the host defense mechanisms (37, 86, 87). Our findings show that MAP infection inhibits the regulation of miR-27a and that miR-27a is a key mediator of macrophage polarization into M1 phenotype. Therefore, identification of miRNAs related to the dynamic changes of macrophage polarization and understanding their immune regulating functions are important for discussing the molecular basis of disease progression and the development of novel miRNA-targeted therapeutic strategies. Several research groups have explored the role of miRNAs in vitro models but few research groups have achieved promising results with miRNAs in vivo experiments. ETHICS STATEMENT All animal experiments were conducted in accordance with the guidelines for the care of laboratory animals, Ministry of Science and Technology People’s Republic of China and approved animal care and use committee (IACUC) protocols at China Agricultural University of Beijing (20110611-01). All other experiments were carried out in accordance with the University Institutional Biosafety Committee (IBC) approved protocol number 20110611-0. DISCUSSION Similarly, over- expression of miR-27a substantially decreased viral infectivity by repressing the expression of many lipid metabolism-related genes, which are essential for the production of HCV parti- cles. Furthermore, miR-27a enhanced in vitro IFN signaling, and patients who expressed high levels of miR-27a in the liver showed a more favorable response to antiviral therapy (81). Forced expression of miR-155 reduced the intracellular growth of mycobacteria by promoting autophagic response in macrophages (82). These studies imply a potential role of miRNAs in the regulation of immune response against invading pathogens. Our results indicate that miR-27a negatively regu- lates the expression of IL-10 and hence promotes the ability of macrophages to eliminate intracellular MAP. In conclusion, the interaction between macrophages and MAP remains a challenging subject. However, there is evidence suggest- ing that macrophages are critical immune-mediating cells playing an important role in the control of MAP infection. Our findings explore the essential role for miR-27a in macrophage activation and MAP elimination and provide important information for the development of therapeutic interventions in paratuberculosis. Although major milestones have been achieved regarding the study of macrophages as an integral part of the innate immune responses against MAP, further molecular research is required to determine whether therapeutic interventions with miR-27a in the modulation of macrophage activation could be a realistic approach for the control of MAP infection in humans and large animals. Frontiers in Immunology \| www.frontiersin.org REFERENCES Mycobacterium paratu berculosis as a cause of Crohn’s disease. Expert Rev Gastroenterol Hepatol (2015) 9(12):1523–34. doi:10.1586/17474124.2015.1093931 23. Ahluwalia PK, Pandey RK, Sehajpal PK, Prajapati VK. Perturbed microrna expression by Mycobacterium tuberculosis promotes macrophage polarization leading to pro-survival foam cell. Front Immunol (2017) 8:107. doi:10.3389/ fimmu.2017.00107 8. Dalziel TK. Chronic interstitial enteritis. Br Med J (1913) 2:1068–70. 9. Liverani E, Eleonora S, Carla C, Paola DM, Andrea B. Mycobacterium avium subspecies paratuberculosis in the etiology of Crohn’s disease, cause or epi- phenomenon. World J Gastroenterol (2014) 20(36):13060–70. doi:10.3748/ wjg.v20.i36.13060 i 24. Keshet Y, Seger R. The MAP kinase signaling cascades: a system of hundreds of components regulates a diverse array of physiological functions. Methods Mol Biol (2010) 661:3–38. doi:10.1007/978-1-60761-795-2_1 h 10. Kuenstner JT, Naser S, Chamberlin W, Borody T, Graham DY, McNees A. The consensus from the Mycobacterium avium ssp. paratuberculosis (MAP) con- ference 2017. Front Public Health (2017) 5:208. doi:10.3389/fpubh.2017.00208 h 25. Hussain T, Shah SZA, Zhao Z, Sreevatsan S, Zhou X. The role of IL-10 in Mycobacterium avium subsp. paratuberculosis infection. J Cell Commun Signal (2016) 14(29):1–14. doi:10.1186/s12964-016-0152-z 11. Thirunavukkarasu S, de Silva K, Begg DJ, Whittington RJ, Plain KM. Macrophage polarization in cattle experimentally exposed to Mycobacterium avium subsp. paratuberculosis. Pathog Dis (2015) 73(9):1–9. doi:10.1093/ femspd/ftv085 26. Koul HK, Pal M, Koul S. Role of p38 MAP kinase signal trans- duction in solid tumors. Genes Cancer (2013) 4(9–10):342–59. doi:10.1177/1947601913507951 27. Souza CD. Blocking the mitogen activated protein kinase-p38 pathway is associated with increase expression of nitric oxide synthase and higher pro- duction of nitric oxide by bovine macrophages infected with Mycobacterium avium subsp paratuberculosis. Vet Immunol Immunopathol (2015) 164:1–9. doi:10.1016/j.vetimm.2015.01.007 t 12. Astarie DC, N’Diaye EN, Le Cabec V, Rittig MG, Prandi J, Maridonneau-Parini I. The mannose receptor mediates uptake of pathogenic and nonpathogenic mycobacteria and bypasses bactericidal responses in human macrophages. Infect Immun (1999) 67:469–77. f 13. Fratti RA, Chua J, Deretic V. Induction of p38 mitogen-activated protein kinase reduce early endosome autoantigen 1 (EEA1) recruitment to phagosomal membranes. J Biol Chem (2003) 278:46961–7. doi:10.1074/jbc.M305225200 28. Hoentjen F, Sartor RB, Ozaki M, Jobin C. STAT3 regulates NF-kappaB recruit- ment to the IL-12p40 promoter in dendritic cells. Blood (2005) 105:689–96. doi:10.1182/blood-2004-04-1309 14. Banaiee N, Kincaid EZ, Buchwald U, Jacobs WR Jr, Ernst JD. Potent inhi- bition of macrophage responses to IFN-γ by live virulent Mycobacterium tuberculosis is independent of mature mycobacterial lipoproteins but depen- dent on TLR2. AUTHOR CONTRIBUTIONS Emerging evidence suggests the importance of the link between inflammatory mediators and miRNAs during pathogenic infec- tions (83). miR-23a is a member of miR-23a/27a/24-2 cluster, which has been shown to promote macrophage polarization by targeting JAK1 and STAT-6 directly in murine macrophages (39). Previous studies also showed that several pathogens, including mycobacteria, inhibit the classical activation of macrophages and diminished bactericidal activity of macrophages, through the induction of some miRNAs such as miR-125a (84, 85). TH performed the experiments and wrote the manuscript, and SS and NS helped in critical review and English grammar check. JW and RY helped in experiments related to cell culture and animals infection. YL helped in buying reagents and chemicals for WB experiments. DZ, LY, and XZ gave the conceptual idea for experimental design and critically reviewed the article before final submission. January 2018 \| Volume 8 \| Article 1915 15 Role of miR-27a in Paratuberculosis Hussain et al. FUNDING The Supplementary Material for this article can be found online at http://www.frontiersin.org/articles/10.3389/fimmu.2017.01915/ full#supplementary-material. This work was supported by National Natural Science Foundation of China (Project No.31572487), the MOST-RCUK international REFERENCES modulate antimycobacterial functions of bovine macrophages. PLoS One (2015) 10(6):e0128966. doi:10.1371/journal.pone.0128966 modulate antimycobacterial functions of bovine macrophages. PLoS One (2015) 10(6):e0128966. doi:10.1371/journal.pone.0128966 1. Nielsen SS, Toft N. A review of prevalences of paratuberculosis in farmed animals in Europe. Prev Vet Med (2009) 88(1):1–14. doi:10.1016/j.prevetmed. 2008.07.003 1. Nielsen SS, Toft N. A review of prevalences of paratuberculosis in farmed animals in Europe. Prev Vet Med (2009) 88(1):1–14. doi:10.1016/j.prevetmed. 2008.07.003 17. Cyktor JC, Turner J. Interleukin-10 and immunity against prokaryotic and eukaryotic intracellular pathogens. Infect Immun (2011) 79(8):2964–73. doi:10.1128/IAI.00047-11 2. Lombard JE. Epidemiology and economics of paratuberculosis. Vet Clin North Am Food Anim Pract (2011) 27:525–35. doi:10.1016/j.cvfa.2011.07.012 2. Lombard JE. Epidemiology and economics of paratuberculosis. Vet Clin North Am Food Anim Pract (2011) 27:525–35. doi:10.1016/j.cvfa.2011.07.012 18. Wallner FK, Hopkins MH, Lindvall T, Olofsson P, Tilevik A. Cytokine correlation analysis based on drug perturbation. Cytokine (2016) 90:73–9. doi:10.1016/j.cyto.2016.10.015 3. Lombard JE, Gardner IA, Jafarzadeh SR, Fossler CP, Harris B, Capsel BR, et al. Herd-level prevalence of Mycobacterium avium subsp. paratuberculosis infec- tion in United States dairy herds in 2007. Prev Vet Med (2013) 108(2–3):234–8. doi:10.1016/j.prevetmed.2012.08.006 h 19. McCoy CE, Sheedy FJ, Qualls JE, Doyle SL, Quinn SR, Murray PJ, et al. IL-10 inhibits miR-155 induction by toll-like receptors. J Biol Chem (2010) 285(27):20492–8. doi:10.1074/jbc.M110.102111 4. Liu X, Li J, Yang X, Wang D, Wang J, Wu J. The seroprevalence of Mycobacterium avium subspecies paratuberculosis in dairy cattle in Xinjiang, Northwest China. Ir Vet J (2017) 70:1–5. doi:10.1186/s13620-016-0079-0 20. Nagata R, Satoko K, Yuu M, Xuebo W, Tadashi Y, Yasuyuki M. A specific induction of interleukin-10 by the Map41 recombinant PPE antigen of Myco bacterium avium subsp. paratuberculosis. Vet Immunol Immunopathol (2010) 135:71–8. doi:10.1016/j.vetimm.2009.11.002 ʹ 5. Sorge SS, Kurnick S, Streevatsan S. Detection of Mycobacterium avium sub- species paratuberculosis in the saliva of dairy cows: a pilot study. Vet Microbiol (2013) 164:383–6. doi:10.1016/j.vetmic.2013.02.021 21. Redford PS, Oʹ Garra A, Murray PJ. The role of IL-10 in immune regulation during M. tuberculosis infection. Mucosal Immunol (2017) 4(3):261–70. doi:10.1038/ mi.2011.7 j 6. Stevenson K. Genetic diversity of Mycobacterium avium subspecies paratuber culosis and the influence of strain type on infection and pathogenesis: a review. Veterinary research (2015) 46:64. doi:10.1186/s13567-015-0203-2 22. Wong EA, Kraus C, Reimann KA, Flynn JAL. The role of IL-10 during early M. tuberculosis infection in a non-human primate model. J Immunol (2017) 198(1 Supplement):123.5. 7. McNees AL, Markesich D, Zayyani NR, Graham DY. ACKNOWLEDGMENTS cooperation project (Project No. 2013DFG32500), 2016 CAU Foreign Experts Major Projects (Project No: 2012z018). High-end Foreign Experts Recruitment Program (Project No: GDT20161100071). cooperation project (Project No. 2013DFG32500), 2016 CAU Foreign Experts Major Projects (Project No: 2012z018). High-end Foreign Experts Recruitment Program (Project No: GDT20161100071). We thank Prof. Sarinand Sreevatsan and Prof. Paul Barrow for their useful comments before final submission of the manuscript. REFERENCES Toll-like receptors: from the discovery of NF kappa-B to new insights into transcriptional regulations in innate immunity. Biochem Pharmacol (2006) 72(9):1102–13. doi:10.1016/j.bcp.2006.07.010 h 37. Farrell D, Shaughnessy RG, Britton L, David E, Mac H, Bryan M, et al. The identification of circulating miRNA in bovine serum and their potential as novel biomarkers of early Mycobacterium avium subsp paratubercu losis infection. PLoS One (2015) 10(7):e0134310. doi:10.1371/journal. pone.0134310 56. Miranda K, Huynh T, Tay Y, Ang Y, Tam W, Thomson A, et al. A pattern- based method for the identification of MicroRNA binding sites and their corresponding heteroduplexes. Cell (2006) 126:1203–17. doi:10.1016/j.cell. 2006.07.031 38. Shaughnessy RG, Farrell D, Riepema K, Bakker D, Stephen V, Gordon SV. Analysis of biobanked serum from a Mycobacterium avium subsp paratuber culosis bovine infection model confirms the remarkable stability of circulating miRNA profiles and defines a bovine serum miRNA repertoire. PLoS One (2015) 10(12):e0145089. doi:10.1371/journal.pone.0145089 57. Dweep H, Sticht C, Pandey P, Gretz N. miRWalk—database: prediction of possible miRNA binding sites by “walking” the genes of three genomes. J Biomed Inform (2011) 44:839–47. doi:10.1016/j.jbi.2011.05.002 58. Weiss DJ, Evanson OA, Souza CD. Increased expression of interleukin-10 and suppressor of cytokine signaling-3 associated with susceptibility to Johne’s disease. Am J Vet Res (2005) 66:1114–20. doi:10.2460/ajvr.2005.66.1114 39. Ma S, Liu M, Xu Z, Li Y, Guo H, Ge Y, et al. A double feedback loop mediated by microRNA-23a/27a/24-2 regulates M1 versus M2 macrophage polarization and thus regulates cancer progression. Oncotarget (2016) 7(12):13502–19. doi:10.18632/oncotarget.6284 i 59. Souza CD, Evanson OA, Weiss D. Mitogen activated protein kinasep38 pathway is an important component of the anti-inflammatory response in Mycobacterium avium subsp. paratuberculosis infected bovine monocytes. Microb Pathog (2006) 41:59–66. doi:10.1016/j.micpath.2006.04.002 40. Meng QL, Liu F, Yang XY, Liu XM, Zhang X, Zhang C, et al. Identification of latent tuberculosis infection-related microRNAs in human U937 mac- rophages expressing Mycobacterium tuberculosis Hsp16.3. BMC Microbiol (2014) 14(37):2–9. doi:10.1186/1471-2180-14-37 60. Liston A, Linterman M, Lu LF. MicroRNA in the adaptive immune system, in sickness and in health. J Clin Immunol (2010) 30:339–46. doi:10.1007/ s10875-010-9378-5 41. Yue R, Liu C, Barrow P, Liu F, Cui Y, Yang L, et al. The isolation and molecular characterization of Mycobacterium avium subsp. paratuberculosis in Shandong province, China. Gut Pathog (2016) 8(9):2–9. doi:10.1186/s13099-016-0092-6 h 61. Lee HM, Nguyen DT, Lu LF. Progress and challenge of microRNA research in immunity. Front Genet (2014) 5:178. doi:10.3389/fgene.2014.00178 62. REFERENCES Redente EF, Higgins DM, Dwyer-Nield LD, Orme IM, Gonzalez-Juarrero M, Malkinson AM. Differential polarization of alveolar macrophages and bone marrow derived monocytes following chemically and pathogen-induced chronic lung inflammation. J Leukoc Biol (2010) 88(1):159–68. doi:10.1189/ jlb.0609378 42. Thirunavukkarasu S, de silva K, Whittington RJ, Plain KM. In vivo and in vitro expression pattern of toll-like receptors in Mycobacterium avium subspecies paratuberculosis infection. Vet Immunol Immunopathol (2013) 15:20–31. doi:10.1016/j.vetimm.2013.08.008 i 43. El-Etr SH, Subbian S, Cirillo SLG, Cirillo JD. Identification of two Mycobacterium marinum loci that affect interactions with macrophages. Infection and immunity (2004) 72(12):6902–13. doi:10.1128/IAI.72.12.6902- 6913.2004 63. Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activa- tion. Nat Rev Immunol (2008) 8:958–69. doi:10.1038/nri2448 f 64. Diefenbach A, Colonna M, Koyasu S. Development, differentiation, and diversity of innate lymphoid cells. Immunity (2014) 41:354–65. doi:10.1016/j. immuni.2014.09.005 44. Weaver BK, Bohn E, Judd BA, Gil MP, Schreiber RD. ABIN-3: a molecular basis for species divergence in interleukin-10-induced anti-inflammatory actions. Mol Cell Biol (2007) 27:4603–16. doi:10.1128/MCB.00223-07 65. Pineros AR, Campos LW, Fonseca DM, Bertolini TB, Gembre AF, Prado RQ, et al. M2 macrophages or IL-33 treatment attenuate ongoing Mycobacterium tuberculosis infection. Sci Rep (2017) 7:1–12. doi:10.1038/srep41240 45. Srinivasan L, Serdar AG, Benjamin EH, Jessica LM, Petros CK, Volker B. Identification of transcription factor that regulates host cell exit and viru- lence of Mycobacterium tuberculosis. PLoS Pathog (2016) 12(5):e1005652. doi:10.1371/journal.ppat.1005652 66. Wu X-Q, Dai Y, Yang Y, Huang C, Meng X-M, Wu B-M, et al. Emerging role of microRNAs in regulating macrophage activation and polarization in immune response and inflammation. J Immunol (2016) 148:237–48. doi:10.1111/ imm.12608 46. Pan W, Wang H, Jianwei R, Ye Z. MicroRNA-27a promotes proliferation, migration and invasion by targeting MAP2K4 in human osteosarcoma cells. Cell Physiol Biochem (2014) 33:402–12. doi:10.1159/000356679 67. Yeh CH, Chen TP, Wang YC, Lin YM, Fang SH. MicroRNA-27a regulate car- diomyocytic apoptosis during cardioplegia-induced cardiac arrest by target- ing interleukin 10yrelated pathways. Shock (2012) 38(6):607–14. doi:10.1097/ SHK.0b013e318271f944 47. Pooley HB, de Silva K, Purdie AC, Begg DJ, Whittington RJ, Plain KM. A rapid method for quantifying viable Mycobacterium avium subsp. paratuber culosis in cellular infection assays. Am Soc Microbiol (2016) 112(3–4):203–12. doi:10.1128/AEM.01668-16 68. Liu Y, Chen Q, Song Y, Lai L, Wang JL, Yu H, et al. MicroRNA-98 nega- tively regulates IL-10 production and endotoxin tolerance in macrophages after LPS stimulation. FEBS Lett (2011) 585(12):1963–8. doi:10.1016/j. febslet.2011.05.029 48. REFERENCES J Immunol (2006) 176:3019–27. doi:10.4049/jimmunol.176. 5.3019 29. Guo H, Ingolia NT, Weissman JS, Bartel DP. Mammalian microRNAs pre- dominantly act to decrease target mRNA levels. Nature (2010) 466:835–40. doi:10.1038/nature09267 30. Drury RE, Connor DO, Pollard AJ. The clinical application of MicroRNAs in infectious disease. Front Immunol (2017) 8:1182. doi:10.3389/fimmu.2017.01182 15. Weiss DJ, Souza CD, Evanson OA, Sanders M, Rutherford M. Bovine monocyte TLR2 receptors differentially regulate the intracellular fate of Mycobacterium avium subsp. paratuberculosis and Mycobacterium avium subsp. avium. J Leukoc Biol (2008) 83:48–55. doi:10.1189/jlb.0707490 i 31. O’Connell RM, Rao DS, Chaudhuri AA, Baltimore D. Physiological and pathological roles for microRNAs in the immune system. Nat Rev Immunol (2010) 10:111–22. doi:10.1038/nri2708 32. Turner ML, Schnorfeil FM, Brocker T. MicroRNAs regulate dendritic cell differentiation and function. J Immunol (2011) 187(8):3911–7. doi:10.4049/ jimmunol.1101137 16. Bannantine JP, Judith RS, Elizabeth L, Maria Clara DC, Souza CD. Mycobacterium avium subspecies paratuberculosis recombinant proteins January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 16 Hussain et al. Role of miR-27a in Paratuberculosis 33. Cho S, Wu CJ, Yasuda T, Cruz LO, Khan AA, Lin LL, et al. miR-23~27~24 clusters control effector T cell differentiation and function. J Exp Med (2016) 213(2):235–49. doi:10.1084/jem.20150990 52. Shah SZY, Zhao D, Taglialatela G, Khan SH, Hussain T, Dong H, et al. Early minocycline and late FK506 treatment improves survival and alleviates neuroinflammation, neurodegeneration, and behavioral deficits in prion- infected hamsters. Neurotherapeutics (2017) 14(2):463–83. doi:10.1007/ s13311-016-0500-0 34. Buck AJ, Perot J, Chisholm MA, Kumar DA, Tuddenham L, Cognat V, et al. Post-transcriptional regulation of miR-27 in murine cytomegalovirus infec- tion. RNA (2010) 16:307–15. doi:10.1261/rna.1819210 53. Huang L, Liu Y, Wang L, Chen R, Ge W, Lin Z, et al. Downregulation of miR- 301a suppresses pro-inflammatory cytokines in TLR-triggered macrophages. Immunology (2013) 140(3):314–22. doi:10.1111/imm.12139 35. Ji J, Zhang J, Huang G, Qian J, Wang X, Mei S. Over-expressed microRNA-27a and 27b influence fat accumulation and cell proliferation during rat hepatic stellate cell activation. FEBS Lett (2009) 583(4):759–66. doi:10.1016/j. febslet.2009.01.034 54. Brooks DG, Trifilo MJ, Edelmann KH, Teyton L, McGavern DB, Oldstone MB. Interleukin-10 determines viral clearance or persistence in vivo. Nat Med (2006) 12:1301–9. doi:10.1038/nm1492 36. Tak H, Kim J, Jayabalan AK, Lee H, Kang H, Cho GH, et al. miR-27 regulates mitochondrial networks by directly targeting the mitochondrial fission factor. Exp Mol Med (2014) 46:1–8. doi:10.1038/emm.2014.73 h 55. Doyle SL, O’Neill LA. REFERENCES Vandesompele J, Preter KD, Pattyn F, Poppe B, Roy NV, Paepe AD, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol (2002) 3(7):1–12. doi:10.1186/gb-2002-3-7-research0034 69. Leite FL, Livia BE, Bruce P, John PB, Timothy AR, Judith RS. ZAP-70, CTLA-4 and proximal T cell receptor signaling in cows infected with Mycobacterium avium subsp. paratuberculosis. Vet Immunol Immunopathol (2015) 167:15–21. doi:10.1016/j.vetimm.2015.06.017 49. Gao Y, Lin X, Shi K, Yan Z, Wang Z. Bovine mammary gene expression pro- filing during the onset of lactation. PLoS One (2013) 8(8):1–9. doi:10.1371/ journal.pone.0070393 70. Del Prete G, De Carli M, Almerigogna F, Giudizi MG, Biagiotti R, Romagnani S. Human IL-10 is produced by both type 1 helper (Th1) and type 2 helper (Th2) T cell clones and inhibits their antigen-specific proliferation and cytokine production. J Immunol (1993) 150(2):353–60. 50. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods (2001) 25(4):402–8. doi:10.1006/meth.2001.1262 51. Xie N, Cui H, Banerjee S, Tan Z, Salomao R, Fu M, et al. miR-27a regulates inflammatory response of macrophages by targeting IL-10. J Immunol (2014) 193:327–34. doi:10.4049/jimmunol.1400203 71. Buza JJ, Hikono H, Mori Y, Nagata R, Hirayama S, Aodon-geril. Neutralization of interleukin-10 significantly enhances gamma interferon expression in peripheral blood by stimulation with Johnin purified protein derivative and January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 17 Hussain et al. Role of miR-27a in Paratuberculosis by infection with Mycobacterium avium subsp. paratuberculosis in experimen- tally infected cattle with paratuberculosis. Infect Immun (2004) 72:2425–8. doi:10.1128/IAI.72.4.2425-2428.2004 82. Wang J, Yang K, Zhou L, Wu M, Wu Y, Zhu M, et al. MicroRNA-155 promotes autophagy to eliminate intracellular mycobacteria by targeting Rheb. PLoS Pathog (2013) 9(10):e1003697. doi:10.1371/journal.ppat.1003697 72. Pitt JM, Stavropoulos E, Redford PS, Beebe AM, Bancroft GJ, Young DB, et al. Blockade of IL-10 signaling during bacillus Calmette-Guerin vaccination enhances and sustains Th1, Th17, and innate lymphoid IFN-g and IL-17 responses and increases protection to Mycobacterium tuberculosis infection. J Immunol (2012) 189:4079–87. doi:10.4049/jimmunol.1201061 83. Xu G, Zhang Z, Xing Y, Wei J, Ge Z, Liu X, et al. MicroRNA-149 negatively regulates TLR-triggered inflammatory response in macrophages by targeting MyD88. J Cell Biochem (2014) 115:919–27. doi:10.1002/jcb.24734 84. Banerjee S, Cui H, Xie N, Tan Z, Yang S, Icyuz M, et al. miR-125a-5p regulates differential activation of macrophages and inflammation. REFERENCES J Biol Chem (2013) 288(49):35428–36. doi:10.1074/jbc.M112.426866 73. Weiss DJ, Evanson OA, Souza CD. A Critical role of interleukin-10 in the response of bovine macrophages to infection by Mycobacterium avium sub paratuberculosis. Am J Vet Res (2005) 66:721–6. doi:10.2460/ ajvr.2005.66.721 85. Kim JK, Yuk JM, Kim SY, Kim TS, Jin HS, Yang CS, et al. MicroRNA-125a inhibits autophagy activation and antimicrobial responses during mycobac- terial infection. J Immunol (2015) 194:1–11. doi:10.4049/jimmunol.1402557 74. Kim C, Sano Y, Todorova K, Carlson BA, Arpa L, Celada A, et al. The kinase p38 alpha serves cell type-specific inflammatory functions in skin injury and coordinates pro- and anti-inflammatory gene expression. Nat Immunol (2008) 9:1019–27. doi:10.1038/ni.1640 86. Malvisi M, Palazzo F, Morandi N, Lazzari B, Williams JL, Pagnacco G, et al. Responses of bovine innate immunity to Mycobacterium avium subsp. paratuber culosis infection revealed by changes in gene expression and levels of MicroRNA. PLoS One (2016) 11(10):e0164461. doi:10.1371/journal.pone.0164461 oS One (2016) 11(10):e0164461. doi:10.1371/journal.pone.0164461 75. Courties G, Seiffart V, Presumey J, Escriou V, Scherman D, Zwerina J, et al. In vivo RNAi-mediated silencing of TAK1 decreases inflammatory Th1 and Th17 cells through targeting of myeloid cells. Blood (2010) 116:3505–16. doi:10.1182/blood-2010-02-269605 87. Sharma N, Kumawat KL, Rastogi M, Basu A, Singh SK. Japanese encephalitis virus exploit the microRNA-432 to regulate the expression of suppressor of cytokine signaling (SOCS) 5. Sci Rep (2016) 6:1–12. doi:10.9734/JSRR/ 2016/17929 88. Zheng D, Yu Y, Li M, Wang G, Chen R, Fan GC, et al. Inhibition of microRNA 195 prevents apoptosis and multiple-organ injury in mouse models of sepsis. J Infect Dis (2016) 213:1661–70. doi:10.1093/infdis/jiv760 i 76. Xia ZP, Sun L, Chen X, Pineda G, Jian X, Adhikari A, et al. Direct activa- tion of protein kinases by unanchored polyubiquitin chain. Nature (2009) 461:144–99. doi:10.1038/nature08247 89. Hock LT, Kaiko GE, Plank M, Li J, Maltby S, Essilfie AT, et al. Antagonism of miR-328 increases the antimicrobial function of macrophages and neutro- phils and rapid clearance of non-typeable Haemophilus influenzae (NTHi) from infected lung. PLoS Pathog (2015) 11(4):1–21. doi:10.1371/journal. ppat.1004549 77. Mihaly SR, Morioka S, Ninomiya-Tsuji J, Takaesu G. Activated macro- phage survival is coordinated by TAK1 binding proteins. PLoS One (2014) 9(4):e94982. doi:10.1371/journal.pone.0094982 78. Souza CD, Eckstein WC, Sreevatsan TM, Weiss DJ. Mannosylated lipoarab- inomannans from Mycobacterium avium subsp. paratuberculosis alters the inflammatory response by bovine macrophages and suppresses killing of Mycobacterium avium subsp. avium organisms. PLoS One (2013) 8(9):e75924. January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org REFERENCES doi:10.1371/journal.pone.0075924 Conflict of Interest Statement: All authors read and approved the final manu- script. The authors declare that they have no competing conflict of interests. 79. Liu PT, Modlin RL. Human macrophage host defense against Mycobacterium tuberculosis. Curr Opin Immunol (2008) 20:371–6. doi:10.1016/j.coi.2008.05.014 The reviewer AT and handling editor declared their shared affiliation. The reviewer AT and handling editor declared their shared affiliation. 80. Bermudez LE, Young LS. Tumor necrosis factor, alone or in combination with IL-2, but not IFN-gamma, is associated with macrophage killing of Mycobacterium avium complex. J Immunol (1988) 140:3006–13. Copyright © 2018 Hussain, Zhao, Shah, Wang, Yue, Liao, Sabir, Yang and Zhou. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Copyright © 2018 Hussain, Zhao, Shah, Wang, Yue, Liao, Sabir, Yang and Zhou. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 81. Shirasaki T, Honda M, Shimakami T, Horii R, Yamashita T, Sakai Y, et al. MicroRNA-27a regulates lipid metabolism and inhibits hepatitis C virus rep- lication in human hepatoma cells. J Virol (2013) 87(9):5270–86. doi:10.1128/ JVI.03022-12 January 2018 \| Volume 8 \| Article 1915 Frontiers in Immunology \| www.frontiersin.org 18
https://openalex.org/W4233436786	https://biblio.ugent.be/publication/8748186/file/8748187	English	null	Self-management Support in Primary Care Practice: the Development of a Conceptual Model Using a Qualitative Approach	Research Square (Research Square)	2,021	cc-by	8,128	Self-management Support in Primary Care Practice: the Development of a Conceptual Model Using a Qualitative Approach Lotte Timmermans (  lotte.timmermans@kuleuven. Academic Centre of General Practice, KU Leuven Dagje Boeykens Ghent University Mustafa Muhammed Sirimsi University of Antwerp Peter Decat Ghent University Veerle Foulon KU Leuven Ann Van Hecke Ghent University Mieke Vermandere Academic Centre of General Practice, KU Leuven Birgitte Schoenmakers Academic Centre of General Practice, KU Leuven Results Interviews revealed in-depth insights into patients’ care experiences. A conceptual model was developed for primary care practice, including five fundamental tasks for healthcare professionals - Supporting, Involving, Listening, Coordinating and Questioning (SILCQ) – contributing to the support of self-management of chronic patients. Background Coping with a chronic disease can be really challenging. Self-management represents a promising strategy to improve daily life experiences. The role of primary healthcare professionals cannot be underestimated in supporting self-management. Due to a shortage of theory, implementation of self-management support is hindered in primary care practice. The aim of this study is to create a conceptual model for self-management support by analysing patients’ care experiences towards self-management support. Conclusion This qualitative paper emphasises the use of the SILCQ-model to develop optimal roadmaps and hands-on toolkits for healthcare professionals to support self-management. The model needs to be further explored by all stakeholders to support the development of self-management interventions in primary care practice. Methods An explorative-descriptive qualitative study was conducted in Flanders, Belgium. Semi-structured interviews were performed with 16 patients and their informal caregiver (dyads) using a purposive sampling strategy and processed by an inductive content analysis. Research Article Keywords: Self-management, patients, primary health care, health personnel, qualitative research Posted Date: October 29th, 2021 DOI: https://doi.org/10.21203/rs.3.rs-1003156/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full Lice words: Self-management, patients, primary health care, health personnel, qualitative research Page 1/15 Page 1/15 Study design This qualitative study explored experiences of chronic patients and their informal caregivers, so called dyads, with primary healthcare in Flanders. The aim of the study was to explore specific experiences regarding self-management support by healthcare professionals. For this purpose, a qualitative content approach was used to analyse the transcripts (Graneheim et al., 2017)(35). Study participants Study participants The dyads were purposively sampled. The inclusion criteria were defined based on the definition of patients with complex care needs operationalized by Iglesias(36) and adapted to the specific research question. Background For example, the WISE-model (Thompson et al., 2018) aims to support patients by focussing on active counselling by trained healthcare practitioners(26). The model seeks to encourage practices to incorporate patient-centred care. The five A’s Model (Glasgow et al., 2003) also focusses on patient counselling(20). In this model the creation of a personal action plan informed by 5 A’s elements (i.e., assess, advise, agree, assist and arrange) is central, which is considered as essential to facilitate patient self- management(20). Various other approaches to self-management support are presented in literature. In 2017, the strategic international chronic condition self-management support framework (Mills et al.) formulated guiding principles and strategic directions prior to the establishment of self-management support initiatives(27). A comparative overview for different support frameworks exists and aims to create a platform for researchers to operationalise frameworks (O’connell et al., 2018)(28). Although they are designed to be applicable in primary care practice, self-management support interventions are not yet optimal in use and their effectiveness is questioned(29–32). The models and frameworks described in literature tend to focus on the implementation of self-management support rather than on underlying mechanisms, such as the interaction between healthcare professionals and their patients. However, understanding these mechanisms is of great importance in effective self-management support(33). In addition, existing models are drawn up by researchers based on cooperation with healthcare professionals, instead of with patients. Since patients act as equal partners in a collaboration on shared responsibility(34), the importance of involving them in research cannot be underestimated. More insights are required to explore patients’ experiences towards this partnership to self- manage a chronic condition(25). Therefore, there is a need to investigate not only the underlying mechanisms that help people to self-manage a chronic condition, but also to listen to the voice of the patients themselves. These understandings should supplement insights from existing models and orientate the further development of new self-management support interventions. This study offers an exploratory view on patient’s self-management and on the support by primary healthcare professionals. The aim is to create a conceptual model for self-management support by analysing patients’ experiences. The research question addressed in this study is, “What do we learn from patients’ care experiences about the interaction with care professionals related to self-management support?”. This study is part of a larger research project of the Primary Care Academy (PCA) that aims to explore patients’ experiences towards primary care in Flanders (Belgium). Background There is a growing number of people worldwide living with multiple chronic conditions(1, 2). Approximately one on three adults has to face the challenges of living with multimorbidity(3). In Europe, this number of people is estimated at more than 50 million(4). Chronic conditions are defined by the World Health Organization (WHO) as long standing and slowly deteriorating diseases(5). Due to this chronic character, the impact on a patient’s daily life cannot be underestimated. The consequences of chronic diseases are generally reflected in limited capacity, reduced functionality and productivity, reduced quality of life and increased healthcare costs(6, 7). As a result, chronic diseases require intensive management. Multiple interventions have been developed in order to support people with chronic diseases(7–9). Key roles are mainly reserved for the healthcare professionals and the social environment of the patient(10–14) but even more important in the care process is the role of the patient himself(15). By taking charge of their own chronic diseases, patients fully commit in their own care(16). This responsibility helps them to function in daily life activities, to experience interdependency throughout the care process, and offers the possibility to optimize patient’s own care(17). In addition, taking charge of their own health contributes to the development of self-management(18). Self-management is defined as “the individual’s ability to manage the symptoms, treatment, physical and psychosocial consequences and lifestyle changes inherent in living with a chronic condition” (Barlow et al. 2002)(19). Patients can be assisted in this process of taking ownership by healthcare professionals. Health care systems are gradually changing as to support these professionals in taking up this role. This change is defined as “the process of making and refining multi-level changes in healthcare systems (and the community) to facilitate patients self-management” (Glasgow et al. 2003)(20). Self-management support implies intensive cooperation among patients, healthcare professionals and the healthcare system(21). Primary healthcare professionals are well-positioned to help patients developing the ability of self-management, since primary care often serves as the first point of contact in the care system(22). They have the opportunity to encourage patients to take part in their own care process. As a result, patients feel empowered and can actively be involved in their health care(23–25). Page 2/15 Page 2/15 Several models have been developed to guide self-management support interventions for patients with chronic conditions. Inclusion criteria The interviewed patients were purposively recruited and had to meet all the following criteria: (1) aged 18 years or above, (2) suffering from a single severe chronic condition or two or more stable chronic conditions (defined as multimorbidity), and (3) receiving support from at least three primary healthcare disciplines, in addition to the support of an informal caregiver. To achieve a heterogeneous maximal variation sample, the patients had to meet one of the following additional criteria: (1) taking four or more different medications related to their chronic condition(s), (2) demanding a higher need of care, (3) living in a low socio-economic situation, (4) estimated to have limited or low health literacy, or (5) tending to need more care according to at least one member of their primary care team. Page 3/15 Page 3/15 Exclusion criteria Exclusion criteria were defined for either ethical or practical reasons and assessed by the entire research team: (1) patients legally incapacitated to participate, (2) patients incapable to reason about care for various reasons (e.g., severe mental illness, cognitive impairment), (3) patients incapable of being interviewed during the predetermined time frame, (4) patients unable to give permission by the informed consent form, and (5) patients with terminal illness. Dyads were approached by health and welfare organisations or by their General Practitioner (GP), who provided general information on the study. Oral permission of the dyads to their GP was required for the research team to contact the participants. The researcher explained the informed consent form and provided additional information on the project. The dyads were assured participation was completely voluntary and their informed consent was obtained before the beginning of the interviews. Data collection Data analysis To address the research question in this paper, a qualitative inductive content analysis was undertaken by the main researcher LT (Graneheim et al., 2017)(35). First, transcripts were read multiple times to gain an overall naive understanding. Afterwards, the data were reduced into meaningful units addressing patients’ self-management support experiences. Subsequently, units were condensed and labelled with codes. Thereafter, the codes were compared and allocated into subcategories by classifying the codes according to similarities and differences. Similar subcategories were merged into each other and categorized into main categories. Finally, a reverse approach was used, and the overall main categories were tailored to the initial data. Data analysis was conducted using an Excel spreadsheet. Meaningful units, condensations, subcategories, and categories were checked by the principal investigator (BS) and confirmed to increase the credibility of the results. We considered data saturation as the end point of data collection and analysis, meaning no new data were generated from the interviews. R h Ethical approval for the original study was obtained from the Ethical Committee of University of Antwerp (B300201942302). All methods were carried out in accordance with relevant guidelines. The entire study was in accordance with the Helsinki Declaration. Data collection Semi-structured interviews were organised with dyads in their home setting or due to the COVID-19 pandemic by using video conferencing platforms. The interviews were conducted between January 2020 and August 2020 and were supported by an open- ended interview guide. Questions focused on patients’ primary care experiences and their interaction with healthcare professionals. More specifically, open-ended questions in the interview related to 1) daily experiences of living with a chronic condition; 2) structure and functioning of the care network; 3) empowerment, involvement, and participation in care processes; 4) needs, goals and wishes towards primary care and 5) guidance and support strategies in primary care. Every interview was scheduled to last no more than one hour a half. The interviews were addressed to the patient, but the informal caregiver was offered the possibility to add input or support the conversation to improve patient understanding. The collection of data was pilot tested, and the first interviews were conducted by two researchers of the team (DB, MS, and LT) together. The subsequent interviews were conducted independently by one researcher of the same team, individually recorded and transcribed verbatim. D t l i Research team The interviews were collected by DB, MS, and LT; the analysis of the experiences towards self-management support was performed by the main researcher LT in close collaboration with the entire research team. Before the project initiation, the team received a training on the principles and methods in qualitative research to assure a certain level of standardization. Ethics Characteristics of study participants A total of 32 persons were interviewed, including 16 patients and 16 informal caregivers. Most patients were living together with their informal caregiver (11 out of 16 patients). The characteristics of the participants are summarized in Table 1. Table 1 Characteristics of the participants Patients Informal caregivers Gender Male Female 5 11 7 9 Mean age 67.5 66.8 Additional inclusion criteria Patients taking four or more different medications Patients demanding a higher need of care Patients of low socio-economic situation Patients estimated to have limited or low health literacy Patients tending to need more care according to at least one member of their primary care team 11 6 3 3 2 - - - - - Employment Employed Unemployed Unemployed due to disability Retired 0 0 3 13 1 1 3 11 Structural analysis Table 1 Characteristics of the participants Results In total, 16 interviews with a patient-informal caregiver dyad were performed. The interviews lasted from 58 to 90 min. Although the interviews were conducted with patients in the presence of their informal caregivers, the results presented in this paper entirely Page 4/15 Page 4/15 focus on the patient’s experiences towards self-management support. Involving the informal caregivers helped patients in formulating their story and to feel at ease. Structural analysis The inductive content analysis of the interview data resulted in five main categories regarding the role of healthcare professionals in reinforcing patients’ self-management: supporting, involving, listening, coordinating, and questioning. The main categories are the result of breaking down the interview transcripts into meaning units, further condensed into multiple subcategories, and finally encapsulated into main categories (Table 2). In some cases, a subcategory applies to multiple main categories since the main categories are not strictly delineated and overlap slightly. Table 3 provides an overview of the main categories and subcategories. The categories exclusively focus on interactions between patients and their healthcare professionals. The healthcare professional network consisted in many cases of the same key actors including a GP, a pharmacist, a home nurse, and a medical specialist related to a specific disease. Typically, the primary care professionals acted as the first and central point of contact for patients. Depending on patients’ case, other primary healthcare professionals were also involved like social workers, physiotherapists, speech therapists, members of the pain clinic, dietitian, etc. A patient with Parkinson’s disease, for example, would be additionally supported by a physiotherapist and a speech therapist. In a limited number of interviews, there was also guidance by a psychologist or psychiatrist. In the following section, the different categories are presented with examples of verbatim quotes. Page 5/15 Page 5/15 Page 5/15 Table 2 Example of the inductive content analysis Meaning unit Condensation Subcategory Main Category “She [GP] helped me… I can always go to visit her or call if I do not feel okay. Not for the prescription of my medications, but to meet for a talk.” (Patient – P4) The doctor is both physically and by telephone accessible for not only the prescription of medication, but also for a listening ear. Accessibility Understanding Coordinating Listening “I have balance problems. If the physiotherapist comes at home, our first exercise is on it.” ... “We walk around. She stands behind me when she sees me shaking so she can quickly grab me.” (Patient – P2) Treatment of the physiotherapist with supportive exercise (balance problems) Practical support Formal support Supporting “When I was first diagnosed, I had many questions. And I gave them to him [doctor].” … “That doctor just answered the questions he wanted to respond. Structural analysis And when the doctor’s appointment was over, I never received the answers on the questions, he did not likely formulate a response on. So yes, that was a mess ... and I did not feel good about it.” (Patient – P6) Patient used to write the questions down before an appointment. The doctor just answered the questions he wanted to answer and then the appointment was over. Patient didn't feel good about it. Dialogue Information Questioning Table 2 Example of the inductive content analysis Condensation Table 2 “She [GP] helped me… I can always go to visit her or call if I do not feel okay. Not for the prescription of my medications, but to meet for a talk.” (Patient – P4) “I have balance problems. If the physiotherapist comes at home, our first exercise is on it.” ... “We walk around. She stands behind me when she sees me shaking so she can quickly grab me.” (Patient – P2) “When I was first diagnosed, I had many questions. And I gave them to him [doctor].” … “That doctor just answered the questions he wanted to respond. And when the doctor’s appointment was over, I never received the answers on the questions, he did not likely formulate a response on. So yes, that was a mess ... and I did not feel good about it.” (Patient – P6) (Patient – P6) (Patient – P6) Page 6/15 Table 3 Overview of subcategories and main categories resulting in an overall core category Subcategory Main category Core category Practical support ADL support Physical support Household support Medical support Information exchange Clinical expertise Follow-up Supporting The role of healthcare professionals in self-management support Communication tools Shared decision-making Participation Cooperation Care continuity Freedom of choice Involving Taking time Empathy Understanding Listening ear Dealing with help requests Emotional support Listening to questions Listening to expectations Listening to wishes and goals Listening to care barriers and facilitators Listening Accessibility Care continuity Deliberation Stability Collaboration Coordinating Supporting Within our study context, we defined supporting as all elements supplied by a healthcare professional related to treatment, follow- up, and guidance. The participating patients indicated to be accompanied by a team of healthcare professionals. According to the interview data, the main actors involved in the active support were nurses, GPs, and physiotherapists. Their support was experienced as essential to fit the chronic condition into daily life. In essence, essential to self-manage the disease. “For my care… I don’t think other persons are involved beside my physiotherapist, Table 3 Subcategory Main catego Practical support ADL support Physical support Household support Medical support Information exchange Clinical expertise Follow-up Supporting Communication tools Shared decision-making Participation Cooperation Care continuity Freedom of choice Involving Taking time Empathy Understanding Listening ear Dealing with help requests Emotional support Listening to questions Listening to expectations Listening to wishes and goals Listening to care barriers and facilitators Listening Accessibility Care continuity Deliberation Stability Collaboration Time management Support network Teamwork Point of contact Follow-up Coordinating Page 7/15 Subcategory Main category Core category Dialogue Information Questioning expectations Questioning experiences Questioning wishes and goals Questioning care barriers and facilitators Questioning the home nurse and my GP.” (patient) Besides delivering physical care (e.g., diagnosis, treatment, symptom control), various other support elements came up in the interviews. More specifically, supporting was about actively guiding through the delivery of practical tools (e.g., a wheelchair), health-related information, medical assistance (e.g., medicines) and the set-up of home care. Home health care represented most of the support. The frequent visits by home care nurses ensured a better disease management. And then in the evening, the physiotherapist comes to visit at home. And he provides exercises for my back and my neck and he applies a bandage with a tape. … “It really helps me. He supports me well.” (patient) … “It really helps me. He supports me well.” (patient) … “It really helps me. He supports me well.” (patien “The purpose of the nursery is to make the physical condition as good as possible. ( ) “The purpose of the nursery is to make the physical condition as good as possible. That goal has been realized.” (patient) That goal has been realized.” (patient) Involving was defined as working in (co-)partnership with patients. Participants described explicitly the wish to be involved in their care. Patients experienced feelings of respect and equality when being actively engaged in health care. It gave them a chance to participate in their own care and to self-manage their chronic condition. The involvement seemed to be mostly related to medical decision making. For example, patients longed for a freedom of choice related to medical treatment options. In addition, the wish was expressed to be involved in decisions concerning the support of daily life activities. Well-informed decisions did arise when there was room for open conversation and discussion. The extent to which patients wanted to be involved depended on the patient. e an appointment with D. [the doctor], he asks me if I want to try the proposed medicine or “If I have an appointment with D. [the doctor], he asks me if I want to try the proposed medicine or if I prefer something else. So yes, I'm taking part in the decisions he makes.” (patient) if I prefer something else. So yes, I'm taking part in the decisions he makes.” (patient) Moreover, patients strived for involvement throughout the entire care process. To overcome the pitfall of a provider-centred management, patients indicated empathy and understanding as essential components of good collaboration. The participants expressed the wish to share their concerns to feel comfortable in the care they receive. Page 8/15 Page 8/15 “No, I make decisions together. Uh, for example… They have asked me a couple of times what I think about a DBS [Deep Brain Stimulation]. Uh, I am quite reluctant towards it Not for the prescription of my medications, but to meet for a talk.” ( ti t) Listening to patients was defined as taking time and being into an accessible mindset. Furthermore, listening means patients felt heard. Participants clearly mentioned that plenty of opportunities can arise from open-minded interactions between care professionals and care receivers. “You have the feeling that those doctors are making time for you. It's not that you walk in their practice and they immediately start looking at the clock… If the doctor comes here for a home visit, and he always comes here, he'll just sit down for half an hour, at least, and tell you all kind of things. He is also a nice person… and he listens to you and gives advice.” (patient) The listening aspect was also mentioned in the interviews when talking about emergency situations. Patients were in need to contact their care professionals if they experienced problems. Dealing with help requests included active listening to all actors involved. The GP mostly acted as the central contact person in these cases. Coordinating Coordinating was defined as guiding responsibilities in the entire care process. According to the participants, coordinated care contributed to coping with their chronic condition, and consequently, to self-manage the chronic condition. Patients expected healthcare professionals to assume responsibility to keep the care network running. An effective follow-up was fundamental by means of coordination between the professionals. Furthermore, patients expressed the importance of being able to make an appeal to their formal network. Again, the central connection was the patients’ GP. “They work well together. P. [GP1] is the one who coordinates everything a bit. J. [GP2] has been away for a year because she had a baby. P. and J. used to come around. During the holidays we received a letter that someone is stepping in. They are all good doctors “She [GP] helped me… I can always go to visit her or call if I do not feel okay. “She [GP] helped me… I can always go to visit her or call if I do not feel okay. “She [GP] helped me… I can always go to visit her or call if I do not feel okay. Listening Listening seemed one of the most talked-about components, defined as the act of giving an audience and paying attention to someone. Patients expected from their healthcare professionals to listen to what they need, what they want and to what they strive for. Only in this way patients felt supported to manage the chronic disease. In addition, the interviews emphasised the importance of encountering a professional with a listening ear to be able to express concerns. The interview participants pointed towards the home care nurse, a psychologist, or the GP as the ideally positioned sounding board. Conceptual Model Using in-depth interviews, insights into patients’ primary care experiences related to self-management support were gained. We learned that patients could manage their chronic disease more effectively if they feel supported, involved in the care process, listened to, if their care is coordinated and if they are questioned. Based on these five main categories, we were able to identify and formulate concepts to describe the key characteristics in the support of self-management. The conceptual model for primary care practice includes five fundamental tasks that need to be performed by healthcare professionals - Supporting, Involving, Listening, Coordinating and Questioning (SILCQ) – that contribute to efficiently supporting chronic patients’ self-management. These tasks were incorporated into the model as the acronym ‘SILCQ’ (Figure 1). Patients are in most cases connected to a social network. We defined this network as ‘the social environment’. According to the participants, the social environment consists of the closest surroundings of a patient. The composition of the social environment varied and was different depending on the patients. Possible members were relatives, friends, and partners. This network was reinforced by peers in some cases. In this close environment, someone had taken upon the role of informal caregiver. “I am in contact with the health insurance fund. “I am in contact with the health insurance fund. “I am in contact with the health insurance fund. Also, with my parents, of course, and my friends and family... Also, with my parents, of course, and my friends and family... And the nurses for appointments and to accompany me to scans etc. “I think that fellow sufferers are important. Like the pain society... That's very good because there you get to know people who also understand you and who are also going through the same thing to some extent. I find the support of those people enormous.” (patient) Being able to manage a chronic condition is closely related to support by the entire care network, including the social environment. The proposed model considers these interactions in a patient-centred care network. “And if she suddenly has a disease flare, I call the medical practice to ask for help. “And if she suddenly has a disease flare, I call the medical practice to ask for help. Afterwards, I can get a medicine from the pharmacist, because in those situations she is out of medication. Everything is written inside that has happened. And if someone [one of the two central GPs] cannot come to us, there are always others doctors available. Everything is written inside that has happened. And if someone [one of the two central GPs] cannot come to us, there are always others doctors available. Everything is written inside that has happened. And if someone [one of the two central GPs] cannot come to us, there are always They coordinate with each other very well.” (patient) According to the participants, the act of coordinating involves communication and discussion between different actors through the entire care process. Participants experienced benefits of information exchange since it guaranteed care continuity. In addition, coordinating included collaborating among the healthcare professionals with a focus on the chronically ill patient. This could be facilitated using digital tools (e.g., digital patient file, electronic health platforms). The interview conversations highlighted the significance of a well-functioning structured care system, in which effective and clear agreements are made. Finally, coordinating also meant that there was good overview of all care actors involved. Page 9/15 Page 9/15 And they have their computers with them. They contain the entire patient file. Questioning For the current analysis, we defined questioning as a type of communication giving raise to conversation by using interrogation. Patients expected the care professionals to ask them questions. Posing questions created a genuine momentum between care professionals and patients. It resulted in an interactive conversation about patient’s wishes, goals, and expectations. Furthermore, formulating questions to the patient initiated valuable conversations about the care process: what goes well and what could be improved? Do patients feel comfortable in their care, do they understand the medical treatment or are there any ambiguities? Additionally, a question could boost the mechanisms of information transfer between patients and their care professionals. A great healthcare professional takes time, poses question and talks. C t l M d l A great healthcare professional takes time, poses question and talks. Discussion The findings highlight healthcare professionals’ role in supporting self-management of chronic conditions. These roles are reflected in the proposed SILCQ-model. Unfortunately, important components such as arranging follow-up are often ignored in practice(37). Nevertheless, our study emphasises the importance of active support from healthcare professionals during both treatment and follow-up. Providing medical information and tools is part of the support. Our study demonstrates that supporting means that the necessary guidance is provided or built up around a patient. In this environment, healthcare professionals and patients should act as equal partners. Fu et al.(38) also revealed the importance of patients’ partnerships with healthcare professionals on the ability to self-manage. Our interview participants expressed the desire to be involved in the entire care process. The challenge seems to shape this collaboration and define the contribution of the patient himself(15,39). Vahdat et al. (40) stresses the importance of this collaborative relationship between the healthcare professional and the patient by focussing on patient's involvement. According to our interview data, involvement contributes to self-management, but the extent depends very much on the individual. Consequently, determining patients’ wishes and goals is an important component of self- management support. Several studies confirm this association with goal-oriented care(41–43). Because of the strong association, healthcare professionals are challenged to actively question patients to determine goals. Conversation is shaped using interrogation. This questioning aspect was strongly emphasised in the interviews. As a result, healthcare professionals should commit to active listening. Unfortunately, healthcare professionals are used to telling, rather than actively questioning, what patients want or what feels good for them(44). Miles et al.(45) confirms these findings by identifying communication as crucial fo effective self-management. The study analyses patients’ experiences and reveals that patients expect to be listened to. To optimally deliver self-management support, a multidisciplinary approach is required. Tocchi et al.(46) described a positive effect of multidisciplinary teams on self-management of symptoms. Our research emphasises the role of a central healthcare professional to coordinate care. According to the interview data, the main actor in supporting self-management is context and patient dependent. Nevertheless, we can argue that in most cases the GP is mentioned as the pivotal figure. Unfortunately, not all GPs are able to fulfil this pivotal role(47,48). This must be considered when organising care around chronic patients. GPs must be taught the necessary competencies and skills in leadership(49). Strengths and limitations Some limitations should be mentioned. First, patients were interviewed in the presence of their informal caregiver for additional support. This may have caused the participants withholding personal stories resulting in some bias. However, involving the care network of patients resulted in valuable insights. As a result, the SILCQ-model can be applied to the entire care network. Secondly, since participation in this research was completely voluntary, volunteer bias can occur. However, our research sample can be considered clinically representative in the international primary care context, when focusing on patients with complex care needs (i.e., mostly an aging population). Thirdly, the SILCQ-model is the result of discussions only with patients. Not including healthcare professionals allowed patients’ voice to be the focus of the model. Involving healthcare professionals should be the next step to validate the results. Lastly, the SILCQ-model does not provide any recommendations for future self-management support interventions. We aimed to provide insights, rather than formulate specific guidelines because effectiveness of interventions is highly context and patient dependent. In addition, we believe that the SILCQ-components are the fundaments of every self- management support intervention in primary care. Conceptual Model When I call, the doctor says: 'Yes, go ahead.' So, I call the pharmacist and I say to them: 'I have a problem, but I don't have the prescription for the medication yet. Can I already pick up the medication?' I’ll bring it later, and that is not a problem.” (informal caregiver) Page 10/15 Discussion Although our findings are in agreements with other studies regarding self-management support, most studies focus on one single aspect of the support. Few comprehensive models have been designed that include multiple aspects (20,26,27). In addition, most models don’t target multimorbidity and only focus on self-management support from a specific healthcare professional(28). Our SILCQ-model addresses this issue by providing insights into the fundamentals of support strategies, independent of the type of disease and healthcare professional. Self-management support interventions are not applicable to every patient and subject to change(50). Therefore, it is important to consult the SILCQ-model which represents the foundation on which to build interventions. Implications for research and practice The SILCQ-model provides a holistic approach to self-management support, while focussing on the central interaction in the care network between healthcare professionals and patients. Programme developers are encouraged to keep the formulated elements Page 11/15 Page 11/15 in mind when setting up new self-management support interventions in primary care. Further research is required to understand how these SILCQ-elements are implemented in care practice and contribute to self-management outcomes. in mind when setting up new self-management support interventions in primary care. Further research is required to understand how these SILCQ-elements are implemented in care practice and contribute to self-management outcomes. Funding The PCA is funded by the Fund dr. Daniël De Coninck, which is managed by the King Baudouin Foundation (Belgium). This research paper received no additional external funding, and the funder was not involved in this research. Authors’ contributions LT, DB and MS conducted the interviews. In-depth analysis was performed by the main researcher LT, in close collaboration wit BS. LT wrote the manuscript. All authors read and approved the final manuscript. Conclusion This qualitative paper highlights the importance of the SILCQ-elements – Supporting, Involving, Listening, Coordinating and Questioning – when reinforcing patients’ self-management. In providing good care, healthcare professionals are expected to prioritize these actions. The model should be further explored by all stakeholders to support the development of self-management interventions in primary care practice. Abbreviations WHO: World health organization; WISE: Whole system informing self-management engagement; PCA: Primary care academy; GP: General practitioner; SILCQ: Supporting, involving, listening, coordinating and questioning Consent for publication Not applicable. Availability of data and materials Availability of data and materials The datasets used and/or analysed during the current study available from the corresponding author on reasonable request. The interview guide is made available at: The datasets used and/or analysed during the current study available from the corresponding author on reasonable request. The interview guide is made available at: https://kuleuven- my.sharepoint.com/:b:/g/personal/lotte_timmermans_kuleuven_be/EVzVbmXRf3JEqRHNMVNEYFMBygAE0Vc380kvRrz2rRD0Iw? e=sfo6gI https://kuleuven my.sharepoint.com/:b:/g/personal/lotte_timmermans_kuleuven_be/EVzVbmXRf3JEqRHNMVNEYFMBygAE0Vc380kvRrz2rRD0Iw? e=sfo6gI Ethics approval and consent to participate Ethical approval for the study was obtained from the Ethical Committee of University of Antwerp (B300201942302). The study was in accordance with the principles outlined in the Declaration of Helsinki. Participants in the interviews were pseudonymised using identification numbers to ensure their confidentiality. Informed consent was obtained by all participants in this study. All methods were carried out in accordance with relevant guidelines. Competing interests The authors have no conflicts of interest to declare. References A systematic review of care management interventions targeting multimorbidity and high care utilization. BMC Health Serv Res. 2018;18(1):65. 9. Barlow JH, Sturt J, Hearnshaw H. Self-management interventions for people with chronic conditions in primary care: Examples from arthritis asthma and diabetes Health Educ J 2002;61(4):365–78 9. Barlow JH, Sturt J, Hearnshaw H. Self-management interventions for people with chronic conditions in primary care: Examples from arthritis, asthma and diabetes. Health Educ J. 2002;61(4):365–78. 0. Zapka JG, Lemon SC. Interventions for patients, providers, and health care organizations. V 10. Zapka JG, Lemon SC. Interventions for patients, providers, and health care organizations. Vol. 101, Cancer. 2004. p. 1165–87. 11. Dwamena F, Holmes-Rovner M, Gaulden CM, Jorgenson S, Sadigh G, Sikorskii A, et al. Interventions for providers to promote a patient-centred approach in clinical consultations. Vol. 12, Cochrane Database of Systematic Reviews. 2012. 12. Latkin CA, Knowlton AR. Social Network Assessments and Interventions for Health Behavior Change: A Critical Review. Behav Med. 2015 Jul 3;41(3):90–7. 13. Coster S, Norman I. Cochrane reviews of educational and self-management interventions to guide nursing practice: A review. Vol. 46, International Journal of Nursing Studies. 2009. p. 508–28. 14. Alvarez C, Greene J, Hibbard J, Overton V. The role of primary care providers in patient activation and engagement in self- management: A cross-sectional analysis. BMC Health Serv Res. 2016 Mar 11;16(1):85. 15. Coulter A, Ellins J. Patient-focused interventions A review of the evidence [Internet]. 2006 [cited 2021 Mar 30]. Available from: 16. Funnell MM. Helping Patients Take Charge of Their Chronic Illnesses. Fam Pract Manag. 2000;7(3):47. 5. Coulter A, Ellins J. Patient-focused interventions A review of the evidence [Internet]. 2006 [c ins J. Patient-focused interventions A review of the evidence [Internet]. 2006 [cited 2021 Mar 3 15. Coulter A, Ellins J. Patient focused interventions A review of the evidence [Internet]. 2006 [cited 2021 Mar 30]. Available from: 16. Funnell MM. Helping Patients Take Charge of Their Chronic Illnesses. Fam Pract Manag. 2000;7(3):47. 6. Funnell MM. Helping Patients Take Charge of Their Chronic Illnesses. Fam Pract Manag. 2 16. Funnell MM. Helping Patients Take Charge of Their Chronic Illnesses. Fam Pract Manag. 2000;7(3):47. 17. McWilliam CL. Patients, persons or partners? Involving those with chronic disease in their care. Vol. 5, Chronic Illness. 2009. p. 277–92. 18. Glasgow RE, Wagner EH, Schaefer J, Mahoney LD, Reid RJ, Greene SM. Development and validation of the Patient Assessment of Chronic Illness Care (PACIC). Acknowledgements Page 12/15 This paper was written on behalf of the Primary Care Academy (PCA). The PCA is a research and teaching network of four universities and six university colleges in Flanders, the White-Yellow Cross (Flemish home care organization) and patient This paper was written on behalf of the Primary Care Academy (PCA). The PCA is a research and teaching network of four universities and six university colleges in Flanders, the White-Yellow Cross (Flemish home care organization) and patient Page 12/15 representatives in Belgium. The purpose of the consortium is to reinforce knowledge about primary care and to develop interventions, optimal roadmaps, and hands-on toolkits for primary care policies, practice, and education, built upon the principles of goal-oriented care, interprofessional collaboration and self-management. We expressly thank our PCA colleagues Dominique Van de Velde and Patricia De Vriendt for their expertise and guidance while preparing, performing, and analysing the dyadic interviews. We also would like to acknowledge all the participating participants and their informal caregivers. representatives in Belgium. The purpose of the consortium is to reinforce knowledge about primary care and to develop interventions, optimal roadmaps, and hands-on toolkits for primary care policies, practice, and education, built upon the principles of goal-oriented care, interprofessional collaboration and self-management. We expressly thank our PCA colleagues Dominique Van de Velde and Patricia De Vriendt for their expertise and guidance while preparing, performing, and analysing the dyadic interviews. We also would like to acknowledge all the participating participants and their informal caregivers. References 1. Kuluski K, Ho JW, Hans PK, La Nelson M. Community care for people with complex care needs: Bridging the gap between health and social care. Int J Integr Care. 2017 Jul 1;17(4). 1. Kuluski K, Ho JW, Hans PK, La Nelson M. Community care for people with complex care needs: Bridging the gap between health and social care. Int J Integr Care. 2017 Jul 1;17(4). 2. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: A cross-sectional study. Lancet. 2012;380(9836):37–43. 2. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care research, and medical education: A cross-sectional study. Lancet. 2012;380(9836):37–43. 3. Hajat C, Stein E. The global burden of multiple chronic conditions: A narrative review. Prev Med Reports. 2018;12:284. 3. Hajat C, Stein E. The global burden of multiple chronic conditions: A narrative review. Prev 4. Van Der Heide I, Melchiorre MG, Quattrini S, Boerma W. Innovating care for people with multiple chronic conditions in Europe An overview [Internet]. [cited 2020 Aug 14]. Available from: www.nivel.eu 4. Van Der Heide I, Melchiorre MG, Quattrini S, Boerma W. Innovating care for people with multiple chronic conditions in Europe An overview [Internet]. [cited 2020 Aug 14]. Available from: www.nivel.eu 5. Chronic disease & Policy - European chronic disease alliance [Internet]. [cited 2020 Oct 12]. Available from: https://alliancechronicdiseases.org/chronic-disease-policy/ 5. Chronic disease & Policy - European chronic disease alliance [Internet]. [cited 2020 Oct 12]. Available from: https://alliancechronicdiseases.org/chronic-disease-policy/ 6. Devins GM, Binik YM, Hutchinson TA, Hollomby DJ, Barré PE, Guttmann RD. The emotional impact of end-stage renal disease: Importance of patients’ perceptions of intrusiveness and control. Int J Psychiatry Med. 1983;13(4):327–43. 6. Devins GM, Binik YM, Hutchinson TA, Hollomby DJ, Barré PE, Guttmann RD. The emotional impact of end-stage renal disease: Importance of patients’ perceptions of intrusiveness and control. Int J Psychiatry Med. 1983;13(4):327–43. 7. Megari K. Quality of life in chronic disease patients. Heal Psychol Res. 2013;1(3). 7. Megari K. Quality of life in chronic disease patients. Heal Psychol Res. 2013;1(3). 7. Megari K. Quality of life in chronic disease patients. Heal Psychol Res. 2013;1(3). 8. Baker JM, Grant RW, Gopalan A. A systematic review of care management interventions targeting multimorbidity and high care utilization. BMC Health Serv Res. 2018;18(1):65. 8. Baker JM, Grant RW, Gopalan A. References Effectiveness of self-management interventions in young adults with type 1 and 2 diabetes: a systematic review and meta-analysis. Diabet Med. 2020;37(2):229–41. 30. Wong SKW, Smith HE, Chua JJS, Griva K, Cartwright EJ, Soong AJ, et al. Effectiveness of self-management interventions in young adults with type 1 and 2 diabetes: a systematic review and meta-analysis. Diabet Med. 2020;37(2):229–41. 31. Jonkman NH, Westland H, Trappenburg JCA, Groenwold RHH, Bischoff EWMA, Bourbeau J, et al. Characteristics of effective self-management interventions in patients with COPD: individual patient data meta-analysis. Eur Respir J. 2016;48(1):55–68. 31. Jonkman NH, Westland H, Trappenburg JCA, Groenwold RHH, Bischoff EWMA, Bourbeau J, et al. Characteristics of effective self-management interventions in patients with COPD: individual patient data meta-analysis. Eur Respir J. 2016;48(1):55–68. 32. Smith SM, Wallace E, Clyne B, Boland F, Fortin M. Interventions for improving outcomes in patients with multimorbidity in primary care and community setting: a systematic review. Syst Rev. 2021;10:271. 32. Smith SM, Wallace E, Clyne B, Boland F, Fortin M. Interventions for improving outcomes in patients with multimorbidity in primary care and community setting: a systematic review. Syst Rev. 2021;10:271. 33. Franklin M, Lewis S, Willis K, Bourke-Taylor H, Smith L. Patients’ and healthcare professionals’ perceptions of self- management support interactions: Systematic review and qualitative synthesis. Chronic Illn. 2018;14(2):79–103. 33. Franklin M, Lewis S, Willis K, Bourke-Taylor H, Smith L. Patients’ and healthcare professionals’ perceptions of self- management support interactions: Systematic review and qualitative synthesis. Chronic Illn. 2018;14(2):79–103. 34. Chertoff J. The evolving physician-patient relationship: equal partnership, more responsibility. Qual Prim Care. 2015;23(1):1– 3. 35. Graneheim UH, Lindgren BM, Lundman B. Methodological challenges in qualitative content analysis: A discussion paper. Nurse Educ Today. 2017 Sep 1;56:29–34. 36. Iglesias FH, Celada CA, Navarro CB, Morales LP, Visus NA, Valverde CC, et al. Complex care needs in multiple chronic conditions: Population prevalence and characterization in primary care. A study protocol. Int J Integr Care. 2018;18(2). 37. Slev VN, Pasman HRW, Eeltink CM, van Uden-Kraan CF, Verdonck-deLeeuw IM, Francke AL. Self-management support and eHealth for patients and informal caregivers confronted with advanced cancer: an online focus group study among nurses. BMC Palliat Care. 2017;16(1). 38. Fu Y, McNichol E, Marczewski K, Closs SJ. Exploring the Influence of Patient-Professional Partnerships on the Self- Management of Chronic Back Pain: A Qualitative Study. Pain Manag Nurs. 2016;17(5):339–49. 39. Phillips RL, Short A, Kenning A, Dugdale P, Nugus P, McGowan R, et al. References Med Care. 2005 May;43(5):436–44. 19. Barlow J, Wright C, Sheasby J, Turner A, Hainsworth J. Self-management approaches for people with chronic conditions: A review. Patient Educ Couns. 2002;48(2):177–87. 20. Glasgow RE, Davis CL, Funnell MM, Beck A. Implementing practical interventions to support chronic illness self-management. Jt Comm J Qual Saf. 2003 Nov 1;29(11):563–74. 21. Lorig KR, Sobel DS, Ritter PL, Laurent D, Hobbs M. Effect of a self-management program on patients with chronic disease. Eff Clin Pract. 2001; 21. Lorig KR, Sobel DS, Ritter PL, Laurent D, Hobbs M. Effect of a self-management program on patients with chronic disease. Eff Clin Pract. 2001; Page 13/15 22. Samuelson M, Tedeschi P, Aarendonk D, De La Cuesta C, Groenewegen P. Position Paper of the European Forum for Primary care Title: Improving inter-professional collaboration in Primary Care. 2012. 23. Dent M, Pahor M. Patient involvement in Europe – a comparative framework. Vol. 29, Journal of Health, Organisation and Management. 2015. p. 546–55. T. Patient and public involvement in chronic illness: beyond the expert patient. BMJ. 2009;338 24. Greenhalgh T. Patient and public involvement in chronic illness: beyond the expert patient. 25. Salemonsen E, Førland G, Hansen BS, Holm AL. Understanding beneficial self-management support and the meaning of user involvement in lifestyle interventions: A qualitative study from the perspective of healthcare professionals. Vol. 20, BMC Health Services Research. 2020. p. 1–12. 26. Thompson DG, O’Brien S, Kennedy A, Rogers A, Whorwell P, Lovell K, et al. The WISE model of self-management support. 2018 [cited 2021 Jul 27]; Available from: https://www.ncbi.nlm.nih.gov/books/NBK487594/ 27. Mills SL, Brady TJ, Jayanthan J, Ziabakhsh S, Sargious PM. Toward consensus on self-management support: the international chronic condition self-management support framework. Health Promot Int. 2017;32(6):942–52. 27. Mills SL, Brady TJ, Jayanthan J, Ziabakhsh S, Sargious PM. Toward consensus on self-management support: the international chronic condition self-management support framework. Health Promot Int. 2017;32(6):942–52. 28. O’connell S, Mc Carthy VJC, Savage E. Frameworks for self-management support for chronic disease: a cross-country comparative document analysis. BMC Heal Serv Res. 2018;18(1). 29. Lindsay S, Kingsnorth S, Mcdougall C, Keating H. A systematic review of self-management interventions for children and youth with physical disabilities. Disabil Rehabil. 2014;36(4):276–88. 29. Lindsay S, Kingsnorth S, Mcdougall C, Keating H. A systematic review of self-management interventions for children and youth with physical disabilities. Disabil Rehabil. 2014;36(4):276–88. 30. Wong SKW, Smith HE, Chua JJS, Griva K, Cartwright EJ, Soong AJ, et al. References The role of the general practitioner in multidisciplinary teams: a qualitative study in elderly care. BMC Fam Pract. 2018;19(1). 50. Lake AJ, Staiger PK. Seeking the views of health professionals on translating chronic disease self-management models into practice. Patient Educ Couns. 2010;79(1):62–8. 50. Lake AJ, Staiger PK. Seeking the views of health professionals on translating chronic disease self-management models into practice. Patient Educ Couns. 2010;79(1):62–8. References Achieving patient-centred care: the potential and challenge of the patient-as-professional role. Heal Expect. 2015;18(6):2616–28. 39. Phillips RL, Short A, Kenning A, Dugdale P, Nugus P, McGowan R, et al. Achieving patient-centred care: the potential and challenge of the patient-as-professional role. Heal Expect. 2015;18(6):2616–28. 40. Vahdat S, Hamzehgardeshi L, Hessam S, Hamzehgardeshi Z. Patient Involvement in Health Care Decision Making: A Review. Iran Red Crescent Med J. 2014;16(1):12454. 40. Vahdat S, Hamzehgardeshi L, Hessam S, Hamzehgardeshi Z. Patient Involvement in Health Care Decision Making: A Review. Iran Red Crescent Med J. 2014;16(1):12454. 41. Lenzen SA, van Dongen JJJ, Daniëls R, van Bokhoven MA, van der Weijden T, Beurskens A. What does it take to set goals for self-management in primary care? A qualitative study. Fam Pract. 2016;33(6):698–703. 42. Langford AT, Sawyer DR, Gioimo S, Brownson CA, O’Toole ML. Patient-centered goal setting as a tool to improve diabetes self- management. Diabetes Educ. 2007;33 Suppl 6(SUPPL. 6). 43. Hughes S, Lewis S, Willis K, Rogers A, Wyke S, Smith L. Goal setting in group programmes for long-term condition self- management support: experiences of patients and healthcare professionals. Psychol Health. 2020;35(1):70–86. Page 14/15 Page 14/15 44. How to Incorporate Self-Management Support into your Practice (Video Transcript). Content last reviewed February 2016. Agency for Healthcare Research and Quality, Rockville, MD. [Internet]. [cited 2021 Jul 29]. Available from: https://www.ahrq.gov/ncepcr/tools/self-mgmt/how-script.html 45. Miles C, Arden-Close E, Thomas M, Bruton A, Yardley L, Hankins M, et al. Barriers and facilitators of effective self-management in asthma: systematic review and thematic synthesis of patient and healthcare professional views. npj Prim Care Respir Med 2017 271. 2017;27(1):1–21. 46. Tocchi C, McCorkle R, Knobf MT. Multidisciplinary Specialty Teams: A Self-Management Program for Patients With Advanced Cancer. J Adv Pract Oncol. 2015;6(5):408. 47. Stange KC, Jaén CR, Flocke SA, Miller WL, Crabtree BF, Zyzanski SJ. The value of a family physician. J Fam Pract. 1998 46(5) 363 8 47. Stange KC, Jaén CR, Flocke SA, Miller WL, Crabtree BF, Zyzanski SJ. The value of a family physician. J Fam Pract. 1998;46(5):363–8. 48. Birgitte S, Frank B, Jan D. What is the role of the general practitioner towards the family caregiver of a community-dwelling demented relative? A systematic literature review. Scand J Prim Health Care. 2009;27(1):31–40. 49. Grol SM, Molleman GRM, Kuijpers A, Sande R van der, Fransen GAJ, Assendelft WJJ, et al. Figure 1 Conceptual model on the support of self-management Icon made by Freepik from www.flaticon.com Conceptual model on the support of self-management Icon made by Freepik from www.flaticon.com Page 15/15
https://openalex.org/W4366266260	https://pure.knaw.nl/ws/files/941071302/1_s2.0_S0165178123001646_main.pdf	English	null	Objective polysomnography-based sleep features and major depressive disorder subtypes in the general population	Psychiatry research	2,023	cc-by	10,343	Objective polysomnography-based sleep features and major depressive disorder subtypes in the general population Objective polysomnography-based sleep features and major depressive disorder subtypes in the general population Solelhac, Geoffroy; Berger, Mathieu; Strippoli, Marie-Pierre F; Marchi, Nicola Andrea; Stephan, Aurélie; Petit, Jean-Marie; Bayon, Virginie; Imler, Théo; Haba-Rubio, Jose; Raffray, Tifenn; Vollenweider, Peter; Marques-Vidal, Pedro; Waeber, Gerard; Siclari, Francesca; Léger, Damien; Geoffroy, Pierre A; Preisig, Martin; Heinzer, Raphaël yp g p p Solelhac, Geoffroy; Berger, Mathieu; Strippoli, Marie-Pierre F; Marchi, Nicola Andrea; Stephan, Aurélie; Petit, Jean-Marie; Bayon, Virginie; Imler, Théo; Haba-Rubio, Jose; Raffray, Tifenn; Vollenweider, Peter; Marques-Vidal, Pedro; Waeber, Gerard; Siclari, Francesca; Léger, Damien; Geoffroy, Pierre A; Preisig, Martin; Heinzer, Raphaël 2023 DOI (link to publisher) 10.1016/j.psychres.2023.115213 document version Publisher's PDF, also known as Version of record document license CC BY Link to publication in KNAW Research Portal 2023 DOI (link to publisher) 10.1016/j.psychres.2023.115213 document version Publisher's PDF, also known as Version of record document license CC BY Link to publication in KNAW Research Portal citation for published version (APA) Solelhac, G., Berger, M., Strippoli, M.-P. F., Marchi, N. A., Stephan, A., Petit, J.-M., Bayon, V., Imler, T., Haba- Rubio, J., Raffray, T., Vollenweider, P., Marques-Vidal, P., Waeber, G., Siclari, F., Léger, D., Geoffroy, P. A., Preisig, M., & Heinzer, R. (2023). Objective polysomnography-based sleep features and major depressive disorder subtypes in the general population. Psychiatry research, 324, 115213. https://doi.org/10.1016/j.psychres.2023.115213 citation for published version (APA) Solelhac, G., Berger, M., Strippoli, M.-P. F., Marchi, N. A., Stephan, A., Petit, J.-M., Bayon, V., Imler, T., Haba- Rubio, J., Raffray, T., Vollenweider, P., Marques-Vidal, P., Waeber, G., Siclari, F., Léger, D., Geoffroy, P. A., Preisig, M., & Heinzer, R. (2023). Objective polysomnography-based sleep features and major depressive disorder subtypes in the general population. Psychiatry research, 324, 115213. https://doi.org/10.1016/j.psychres.2023.115213 General rights C i h d General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. ghts for the publications made accessible in the public portal are retained by the authors and/or other copyright owners accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the KNAW public portal for the purpose of private stu ch • Users may download and print one copy of any publication from the KNAW public portal for the purpose of private study or research. You may not further distribute the material or use it for any profit making activity or commercial gain esearch. • You may not further distribute the material or use it for any profit-making activity or commercial gain. • You may freely distribute the URL identifying the publication in the KNAW public portal. Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim down policy believe that this document breaches copyright please contact us providing details, and we will remove access to the wo vestigate your claim. Download date: 24. Oct. 2024 Objective polysomnography-based sleep features and major depressive disorder subtypes in the general population Geoffroy Solelhac a,, Mathieu Berger a, Marie-Pierre F. Strippoli b, Nicola Andrea Marchi a, Aur´elie Stephan a, Jean-Marie Petit c, Virginie Bayon a, Th´eo Imler a, Jose Haba-Rubio a,d, Tifenn Raffray a,d, Peter Vollenweider e, Pedro Marques-Vidal e, Gerard Waeber e, Damien L´eger f, Francesca Siclari a,g,h,l, Pierre A. Geoffroy i,j,k, Martin Preisig b,1, Rapha¨el Heinzer a,1 a Center for Investigation and Research in Sleep (CIRS), Lausanne University Hospital (CHUV), University of Lausanne, Lausanne, Switzerland b Center for research in Psychiatric Epidemiology and Psychopathology (CEPP), Department of Psychiatry, Lausanne University Hospital and Un Prilly, Switzerland chiatric Neuroscience (CNP), Department of Psychiatry, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland ep Center, Lausanne, Switzerland c Center for Psychiatric Neuroscience (CNP), Department of Psychiatry, Lausanne University Hospital and University of Lausanne, Lausanne, Sw d Florimont Sleep Center, Lausanne, Switzerland Center for Psychiatric Neuroscience (CNP), Department of Psychiatry, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland d Florimont Sleep Center, Lausanne, Switzerland e Department of Medicine, Internal Medicine, Lausanne University Hospital (CHUV) and University of Lausanne, Lausanne, Switzerland f Universit´e Paris Cit´e, VIFASOM, AP-HP, Hˆotel-Dieu, Centre du Sommeil et de la Vigilance, Paris, France g The Sense Innovation and Research Center, Lausanne and Sion, Switzerland h Department of Clinical Neurosciences Lausanne University Hospital (CHUV), Lausanne, Switzerland i GHU Paris - Psychiatry & Neurosciences, Paris, France j Universit´e de Paris, NeuroDiderot, Inserm, Paris, France k D´epartement de Psychiatrie et d’Addictologie, AP-HP, GHU Paris Nord, DMU Neurosciences, Hopital Bichat - Claude Bernard, Paris, France l Netherlands Institute for Neuroscience, Amsterdam, Netherlands Contents lists available at ScienceDirect Contents lists available at ScienceDirect E-mail address: pure@knaw.nl E-mail address: pure@knaw.nl Download date: 24. Oct. 2024 Psychiatry Research 324 (2023) 115213 A R T I C L E I N F O Keywords: Major depressive disorder subtypes Melancholic Atypical Major depressive disorder Sleep Polysomnography Depression Slow wave sleep Rapid eye movement sleep General population Insomnia and its opposite hypersomnia are part of the diagnostic criteria for major depressive disorder (MDD). However, no study has investigated whether the postulated sleep alterations in clinical subtypes of MDD are reflected in polysomnography (PSG)-derived objective sleep measures. The objective of this study was to establish associations between the melancholic, atypical and unspecified subtypes of MDD and objective PSG- based sleep features. This cross-sectional analysis included 1820 community-dwelling individuals who under went PSG and a semi-structured psychiatric interview to elicit diagnostic criteria for MDD and its subtypes. Adjusted robust linear regression was used to assess associations between MDD subtypes and PSG-derived objective sleep measures. Current melancholic MDD was significantly associated with decreased absolute delta power and sleep efficiency and with increased wake after sleep onset. Remitted unspecified MDD was signifi cantly associated with increased rapid eye movements density. No other significant associations were identified. Our findings reflect that some PSG-based sleep features differed in MDD subtypes compared with no MDD. The largest number of significant differences were observed for current melancholic MDD, whereas only rapid eye movements density could represent a risk factor for MDD as it was the only sleep measure that was also asso ciated with MDD in remitted participants. Corresponding author at: Center d’Investigation et de Recherche sur le Sommeil (CIRS), center Hospitalier Universitaire Vaudois (CHUV), Rue du Bugnon 46, 1011 Lausanne, Switzerland. E-mail addresses: Geoffroy.Solelhac@chuv.ch (G. Solelhac), Mathieu.Berger@chuv.ch (M. Berger), Marie-Pierre.Strippoli@chuv.ch (M.-P.F. Strippoli), nicola. marchi@unil.ch (N.A. Marchi), Jean-Marie.Petit@chuv.ch (J.-M. Petit), Virginie.Bayon@chuv.ch (V. Bayon), Jose.Haba-Rubio@chuv.ch (J. Haba-Rubio), Tifenn. Raffray@chuv.ch (T. Raffray), Peter.Vollenweider@chuv.ch (P. Vollenweider), Pedro-Manuel.Marques-Vidal@chuv.ch (P. Marques-Vidal), Gerard.Waeber@chuv. ch (G. Waeber), damien.leger@aphp.fr (D. L´eger), Francesca.Siclari@chuv.ch (F. Siclari), Martin.Preisig@chuv.ch (M. Preisig), Raphael.Heinzer@chuv.ch (R. Heinzer). 1 Co last authors: These authors contributed equally to the work 0165-1781/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Available online 18 April 2023 0165-1781/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Received 1 December 2022; Received in revised form 12 April 2023; Accepted 14 April 2023 1. Introduction Sleep disturbances including insomnia and hypersomnia are part of the diagnostic criteria for major depressive disorder (MDD), and more than 90% of patients experiencing depression report sleep complaints (Ford and Cooper-Patrick, 2001; Geoffroy, 2018; Geoffroy et al., 2018; Geoffroy and Palagini, 2021; Tsuno et al., 2005). According to longi tudinal studies, which suggest a bidirectional relationship between sleep complaints and MDD, sleep problems may not only be a marker of depressive episodes but could also represent a predisposing factor for depressive episodes (Baglioni et al., 2011; Blanken et al., 2020; Buysse et al., 2008; Chang et al., 1997; Franzen and Buysse, 2008; Jaussent et al., 2011; Paunio et al., 2015; Riemann and Voderholzer, 2003; Sai toh et al., 2022; Suh et al., 2013). Moreover, sleep complaints frequently persist after a depressive episode and increase the risk of recurrence (Lee et al., 2013). To the best of our knowledge, no study has yet investigated whether PSG-derived objective sleep measures, including SWS, REM pressure and sleep continuity, could be documented in association with sleep alterations in patients with different MDD subtypes. We hypothesized that 1) the increase in REM sleep pressure (increase in REM density and percentage of REM sleep, decrease in REM latency), 2) the decrease in sleep stage N3 and delta power, 3) the alterations of sleep continuity measures are greater in participants with melancholic MDD subtype than in participants without MDD, and 4) certain characteristics, in particular decrease in delta power and increase in REM density, persist after the remission of melancholic episodes. Therefore, the aim of the present study was to investigate associations between MDD and melancholic, atypical and unspecified MDD subtypes with objective PSG-based sleep features relating to SWS, REM pressure and sleep continuity. To further determine whether PSG findings were dependent on the MDD state, associations were assessed for both current and remitted depressive episodes at the time of the PSG investigation. 2.2. Polysomnography (PSG)-based sleep features Participants performed a full night PSG at home (Titanium, Embla® Flaga, Reykjavik, Iceland). PSG were performed according to the American Academy of Sleep Medicine (AASM) 2007 recommendations (Iber, 2007) and included: electroencephalography (EEG) leads (F3, F4, C1, C2, O1 and O2, 256 Hz sampling rate); electrooculography (EOG, left and right); electromyography (EMG, chin and anterior tibialis mus cle); electrocardiography (ECG, one lead); oxygen saturation (SpO2); airflow (nasal cannula); abdominal and thoracic respiratory efforts; snoring; and body position. PSG data were visually scored according to the AASM guidelines 2007 (Iber, 2007) for sleep. l In a recent systematic review on electroencephaography (EEG) characteristics in melancholic patients, 10 studies were identified that included sleep EEG data (Bruun et al., 2021). In this review, four studies found increased REM latency in melancholic patients compared to non-melancholic people or healthy controls (Giles et al., 1990; Hubain et al., 1995; Rush et al., 1982; Sitaram et al., 1984) and two studies found higher REM density in melancholic MDD compared with healthy controls (Dippel et al., 1987; Sitaram et al., 1984). Two studies found a larger REM proportion in melancholic patients than in healthy controls (Iorio et al., 1994; Sitaram et al., 1984). However, not all studies found significant differences in REM sleep measures. Melancholic patients also exhibited decreased TST, decreased sleep efficiency, and earlier morning awakenings compared to control groups (Feinberg et al., 1982; Frank et al., 1992; Iorio et al., 1994; Rush et al., 1982). A recently published meta-analysis showed that patients with MDD and hypersomnia (close REM density was analyzed using an algorithm from the YASA library (Agarwal et al., 2005; Vallat and Walker, 2021; Yetton et al., 2016) in Python software (Version 3.9) to automatically detect rapid eye move ments on the two EOG (right and left eyes) channels in REM sleep. The following specifications were used for REM detection: between 50 and 325 microV for the amplitude; between 0.3 and 1.2 s for the duration; and between 0.5 and 5 Hz for the frequency. The variable calculated and used in the present analysis was the REM density, corresponding to the number of REMs per minute of REM sleep. The power spectral sleep EEG analysis has been previously described (Lecci et al., 2020). https://doi.org/10.1016/j.psychres.2023.115213 Available online 18 April 2023 Received 1 December 2022; Received in revised form 12 April 2023; Accepted 14 April 2023 0165-1781/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 0165-1781/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http:/ G. Solelhac et al. G. Solelhac et al. Psychiatry Research 324 (2023) 115213 to the atypical MDD subtype) had an increased TST on ad libitum PSG without altered sleep efficiency compared to healthy controls (Plante et al., 2017). 1. Introduction During depressive episodes, objective changes in sleep patterns assessed by polysomnography (PSG) have primarily been observed in three domains: 1) slow wave sleep (SWS), illustrated by a reduction in the proportion of N3 sleep and delta wave activity (Borb´ely et al., 1984; L´eger et al., 2018; Murphy and Peterson, 2015); 2) rapid eye movement sleep (REM) pressure, shown by a shorter REM sleep latency, and in creases in the proportion of REM sleep and REM density (Baglioni et al., 2016; Benca et al., 1992; Kupfer and Foster, 1972; Palagini et al., 2013); and 3) sleep continuity, seen as increases in sleep onset latency and wake after sleep onset and decreases in total sleep time (TST) and sleep effi ciency (Baglioni et al., 2016). A meta-analysis including data from nineteen studies per sleep feature confirmed the alterations in percent age of N3 sleep, REM latency, REM density, percentage of REM sleep, TST, and sleep efficiency in patients with MDD during depressive epi sodes (Pillai et al., 2011). The same meta-analysis also showed that there was an increase in REM latency, a decreased percentage of REM sleep, an increase in TST and sleep efficiency, and, surprisingly, a further decrease in percentage of N3 sleep after MDD remission (Pillai et al., 2011). A decreased proportion of N3 sleep and increased REM density were also reported in individuals at high risk for depression (Pillai et al., 2011). 2.1. Design and participants This study used data from the prospective CoLaus\|PsyColaus cohort study, which assessed associations between mental disorders and car diovascular risk factors in the community. The cohort was randomly selected from 35- to 75-year-old residents of Lausanne, Switzerland, from 2003 to 2006 (Firmann et al., 2008). Participants underwent thorough physical and psychiatric evaluations at baseline and at three follow-up visits (Preisig et al., 2009). The PSG investigation took place between 2009 and 2013 (Heinzer et al., 2015), i.e. between the first and second follow-up of the study. This analysis included patients who had a PSG investigation with subsequent psychiatric evaluation to determine the MDD status at the time of the PSG. Individuals with any lifetime diagnoses of bipolar or schizoaffective disorders and/or schizophrenia were excluded. The CoLaus\|PsyColaus and HypnoLaus studies were approved by the Ethics Committee of the Vaud Canton, and written informed consent was obtained from all individuals. Given the well-known heterogeneity of MDD, which may partially explain inconsistent results, attempts were made to subtype MDD, e.g. using the classic melancholic and atypical subtypes according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) (American Psychiatric Association 1994) or the DSM-5 (American Psychiatric Association, 2013; Bell, 1994). These subtypes have been shown to differ in course and responses to treatment and are postulated to differ with respect to sleep characteristics. Indeed, early-morning awakening is one of the diagnostic criteria for melan cholic, whereas hypersomnia is part of the criteria for atypical depres sive episodes. Moreover, both clinical and epidemiological studies have revealed that they are associated with different biological mechanisms. (Antonijevic, 2006; Musliner et al., 2016) including metabolic features, inflammation and stress hormones, which are also associated with sleep. 2.5. Statistical analysis Univariate comparisons were performed using Pearson’s chi-square test for categorical variables and one-way ANOVA for continuous vari ables as appropriate. Associations between MDD and its subtypes (independent variables) and PSG-derived sleep characteristics (dependent variables) were assessed using robust linear regression models for each PSG-based sleep feature. These robust linear regressions were preferred to multiple linear regression models because of the lack of normality of the residuals observed for multiple linear regression models and to reduce the influ ence of outliers. Models were adjusted for variables associated with MDD subtypes at a significance level of p<0.20 on univariate analysis. The models were adjusted for age, sex, BMI, current smoking status, use of antidepressants and anxiolytics (including hypnotics), AHI, PLMSI, anxiety disorders and substance use disorders. MDD was subdivided according to the lifetime history of episodes with atypical or melancholic features into three subtypes: 1) MDD with atypical features only; 2) MDD with melancholic features only; or 3) unspecified MDD with neither atypical nor melancholic features or with combined atypical and melancholic features. The terminology “any MDD” was used when referring to analyses ignoring the subtypes of MDD. The DIGS also elicits information on the timing of depressive episodes, which allowed us to determine whether episodes were current or remitted at the time of PSG. Information about severity (depression Global Assessment of Functioning (GAF), percentage of recurrence and time in remission at PSG were collected during the DIGS interview. An additional analysis was performed regarding the association be tween AHI and MDD subtypes adjusted for age, sex and BMI using robust linear regression model. R (Version 4.0) software was used for all analyses. The ‘lmRob’ function from R package ‘robust’ (Version 0.7–0) was used for the robust linear regression model. Table 1 Objective polysomnography (PSG)-based sleep features. 2.2. Polysomnography (PSG)-based sleep features After an automatic rejection procedure, sleep EEG recordings were re-referenced to the average of the two mastoid chan nels and band pass filtered between 0.5 and 35 Hz with a finite impulse 2 G. Solelhac et al. Psychiatry Research 324 (2023) 115213 2.3. Psychiatric evaluation Diagnostic information on mental disorders was collected using the French version (Leboyer et al., 1995) of the semi-structured Diagnostic Interview for Genetic Studies (DIGS) (Nurnberger et al., 1994). In terviews were carried out by trained master-level psychologists. A senior psychologist reviewed all interviews and assessments. Psychiatric di agnoses and subtyping of major depressive episodes relied on DSM-IV criteria. The specifier for depressive episodes with atypical features required at least two of the following symptoms: (1) increased appetite or significant weight gain; (2) hypersomnia; (3) leaden paralysis; and (4) interpersonal rejection sensitivity. The specifier for melancholic features required either a loss of pleasure in activities or a lack of reactivity to pleasurable stimuli and at least three out of the following five symptoms: (1) depression regularly worse in the morning; (2) early morning awakening; (3) psychomotor retardation or agitation; (4) anorexia or weight loss; and (5) excessive guilt. For the diagnosis of melancholic episodes, we could not take into account the criterion ‘distinct quality of depressed mood’ because it was not assessed in the DIGS. 2.4. Covariates response. Absolute power spectral density in delta frequency band (1–4 Hz) was calculated on the F3 electrode using the Welch’s method on artifact free consecutive, non-overlapping 6-second epochs (Hamming windows, 8 segments, 50% overlap). Absolute delta power spectral density was analyzed for non-REM (NREM) sleep stages N1, N2 and N3 together. The following objective PSG-based sleep features were analyzed, as defined in Table 1 SWS; REM pressure; and sleep continuity. Socio-demographic characteristics, education level and smoking history were assessed using a standardized interview during the physical investigation at the first follow-up. This investigation also included anthropometric measurements. Height was measured using a vertical stadiometer (Seca) and weight was measured using a calibrated scale to within 0.1 kg (Seca); these values were used to calculate the body mass index (BMI in kg/m2). Respiratory events were scored according to the 2012 AASM criteria (Berry et al., 2012) and periodic leg movements during sleep (PLMS) were scored according to the official World Asso ciation of Sleep Medicine (WASM) standards 2006 (Zucconi et al., 2006). The average number of apneas/hypopneas and PLMS per hour of sleep (apnea-hypopnea index [AHI] and PLMS index [PLMSI]) were calculated. Information on specific anxiety (generalized anxiety disor der, agoraphobia, social phobia, panic disorder) and substance use dis orders (alcohol or drug abuse or dependence) was collected during the DIGS interview at the psychiatric investigation. Information on psy chotropic treatment including anxiolytics, hypnotics and antidepres sants was collected during the DIGS interview at the psychiatric investigation and during the interview the evening before the PSG. Table 1 Table 1 Objective polysomnography (PSG)-based sleep features. Objective polysomnography (PSG)-based sleep features. 2.5. Statistical analysis Objective PSG-based sleep features Definition (1) Slow wave sleep Sleep stage 3 (N3) Proportion of total sleep time spent in the N3 sleep stage Delta power Absolute delta spectral power (µV2/Hz) in non- REM sleep (2) Rapid eye movement pressure Rapid eye movement sleep Proportion of total sleep time spent in the REM sleep stage Rapid eye movements density Number of rapid eye movements per minute of REM sleep Rapid eye movement latency Latency (min) to REM sleep after sleep onset (3) Sleep continuity Sleep onset latency Duration (min) between light off and first sleep onset Wake after sleep onset Duration (min) spent in wake after the first epoch of sleep Total sleep time Duration (min) spent in sleep during the recording Sleep efficiency Sleep efficiency (%) (TST100/total recording time) REM: rapid eye movement; TST, total sleep time. REM: rapid eye movement; TST, total sleep time. 3.1. Study population PSG investigation was performed in 2162 patients, 1908 of whom had a subsequent psychiatric evaluation. After exclusion of patients meeting the exclusion criteria, the final sample included 1820 partici pants (Fig. 1). There were significant differences between different MDD subtypes with respect to age, sex, BMI, current tobacco consumption, use of antidepressants, use of anxiolytics and current anxiety disorders (Table 2). PSG-based sleep measures by MDD subtype status are shown in Table 3. For participants with remitted MDD, the time between remission and PSG was 5.30[1.17–14.05] years (median[IQR]). The course characteristics of MDD subtypes including severity assessed by GAF score, percentage of recurrence for current MDD and time in remission at PSG are described in Table 2. We did not find a significant association between the subtypes of MDD and AHI when adjusted for age, sex and BMI (Table 6). 3 Psychiatry Research 324 (2023) 11 Fig. 1. Study flowchart. Solelhac et al. G. Solelhac et al. Psychiatry Research 324 (2023) 115213 Fig. 1. Study flowchart. 3.2. Slow wave sleep 3.3. Rapid eye movement pressure Remitted MDD participants showed an increase in rapid eye move ment density compared to no MDD participants. According to MDD subtypes, the rapid eye movement density was significantly increased in unspecified remitted MDD compared to no MDD participants (Fig. 2, Table 4). No differences were found for rapid eye movement sleep percentage and rapid eye movement latency. 4. Discussion Current MDD participants showed a decreased in delta power compared to no MDD. To the best of our knowledge, this is the first study assessing objec tive PSG-based sleep features in a large population-based study utilizing semi-structured diagnostic interviews to determine MDD subtypes, allowing determination of associations between sleep features and MDD subtypes as a function of the remission status of depressive episodes. Key findings were as follows: 1) the melancholic MDD subtype was associ ated with objective sleep alterations, including decreased absolute delta power and sleep efficiency and increased wake after sleep onset; 2) the atypical subtype was not associated with any changes in polysomno graphic features; 3) the unspecified subtype was only associated with an increase in REM density; and 4) associations were not entirely explained by MDD state, given that the associations between the unspecified subtype and REM density were observed in individuals with remitted episodes. Regarding the subtypes of MDD, the decreased delta power was only significant for current melancholic MDD compared to no MDD partici pants (Fig. 2, Table 4). No differences were found between MDD sub types compared to no MDD participants for sleep stage 3 percentage. 3.4. Sleep continuity Regarding sleep macrostructure, the lack of change in N3 sleep proportion in individuals with versus without MDD contrasts with meta- analytical findings that reported a lower proportion of N3 sleep stage in patients with MDD compared with controls (Pillai et al., 2011). For sleep microstructure, the reduced absolute delta power associated with the current MDD melancholic subtype is partially in line with previous clinical data in patients with any MDD (Armitage et al., 2000; Plante et al., 2012). According to clinical studies, we observe an association between absolute delta power and any MDD (Pillai et al., 2011). The heterogeneity of this association in clinical samples could be due to the No difference was found regarding sleep continuity for current or remitted MDD compared to no MDD participants without taking regardless of MDD subtypes into account. According to MDD subtypes, a significant increase wake after sleep onset and decrease sleep efficiency were found for participants with current melancholic MDD compared to no MDD (Fig. 2, Table 5). No differences were found for total sleep time and sleep onset latency features. 4 Psychiatry Research 324 (2023) 115213 G. Solelhac et al. Table 2 Table 2 Characteristics of participants according to major depressive disorder subtype (N = 1820). 3.4. Sleep continuity Major depressive disorder subtype status No MDD Remitted Unspecified Remitted Melancholic Remitted Atypical Current Unspecified Current Melancholic Current Atypical Stat p N = 993 N = 417 N = 189 N = 104 N = 57 N = 35 N = 25 Socio-demographic characteristics Age, mean (SD), y 59.7 (11.3) 57.2 (10.3) 58.5 (10.5) 55.7 (10.6) 51.8 (8.9) 54.0 (8.3) 53.5 (8.8) F6=9.8 <0.001 Women, n (%) 416 (41.9) 261 (62.6) 118 (62.4) 78 (75.0) 38 (66.7) 21 (60.0) 19 (76.0) Х2=101.2 <0.001 Education level, n (%) Х2=11.2 0.512 Low 516 (52.0) 195 (46.8) 85 (45.0) 52 (50.0) 25 (43.9) 20 (57.1) 14 (56.0) Middle 262 (26.4) 123 (29.5) 66 (34.9) 29 (27.9) 17 (29.8) 8 (22.9) 8 (32.0) High 214 (21.6) 99 (23.7) 38 (20.1) 23 (22.1) 15 (26.3) 7 (20.0) 3 (12.0) Behavioral characteristics n (%) Current smokers 149 (15.0) 88 (21.1) 40 (21.2) 14 (13.5) 12 (21.1) 12 (34.3) 7 (28.0) Х2=20.0 0.003 Metabolic characteristics BMI, mean (SD), kg/m2 26.2 (4.2) 25.8 (4.3) 25.6 (4.1) 27.0 (4.7) 25.9 (5.6) 25.2 (4.3) 28.3 (7.1) F6=2.9 0.008 Sleep events, mean (SD) AHI (events/h) 16.5 (16.6) 12.7 (13.0) 13.6 (14.9) 14.9 (15.2) 10.8 (13.2) 14.7 (15.8) 13.6 (19.4) F6=4.1 <0.001 AHI ≥15/h, n (%) 405 (40.8) 116 (27.8) 59 (31.2) 34 (32.7) 11 (19.3) 12 (34.3) 10 (40.0) Х2=31.5 <0.001 PLMSI (events/h) 14.6 (25.2) 11.1 (18.6) 13.8 (21.7) 11.8 (22.9) 7.4 (15.5) 9.3 (13.5) 10.5 (18.5) F6=2.1 0.052 PLMSI ≥15/h, n (%) 295 (29.7) 103 (24.7) 57 (30.2) 28 (26.9) 9 (15.8) 9 (25.7) 5 (20.0) Х2=9.3 0.158 Concomitant treatments, n (%) Antidepressants 34 (3.4) 35 (8.4) 30 (15.9) 18 (17.3) 20 (35.1) 11 (31.4) 11 (44.0) Х2=167.9 <0.001 Anxiolytics 65 (6.5) 25 (6.0) 24 (12.7) 11 (10.6) 20 (35.1) 10 (28.6) 5 (20.0) Х2=84.5 <0.001 Psychiatric comorbidities, n (%) Current anxiety disorders† 21 (2.1) 29 (7.0) 13 (6.9) 6 (5.8) 5 (8.8) 4 (11.4) 4 (16.0) Х2=35.5 <0.001 Current substance use disorders 32 (3.2) 19 (4.6) 8 (4.2) 1 (1.0) 2 (3.5) 0 (0.0) 3 (12.0) Х2=10.2 0.117 Course characteristics GAF score, mean (SD) 79.7 (10.3) 46.8 (10.1) 44.3 (11.0) 46.2 (10.1) 51.3 (7.0) 49.7 (8.7) 49.3 (6.6) F6=799.5 <0.001 Recurrent MDE, n (%) – 145 (34.8) 85 (45.0) 48 (46.2) 28 (49.1) 18 (51.4) 14 (56.0) Х2=521.6 <0.001 Time in remission at PSG, median (IQR), y – 7.5 (3.2–17.1) 7.4 (2.5–18.4) 5.0 (2.6–10.5) – – – kw Х2=521.62 <0.001 Descriptive statistics are presented as mean (standard deviation) for continuous variables and n (percentage) for categorical variables. 3.4. Sleep continuity Surprisingly, in our study, positive associations with REM density were observed only in individuals with any remitted or any unspecified MDD rather than current depressive episodes. Although the cross-sectional nature of our data did not allow us to determine the temporal relationship between the onset of depressive episodes and the beginning of alterations in REM density, we hypothesize that elevated REM density is a premorbid trait in individuals with unspecified MDD (Pillai et al., 2011). 4.2. REM sleep pressure Any remitted MDD and remitted unspecified MDD were significantly associated with increased rapid eye movements density, which is considered as the most robust biomarker of elevated REM pressure in MDD (Wichniak et al., 2000). This finding is consistent with the result of clinical studies conducted in the 1970s supporting an increase in REM pressure in patients with depression (Kupfer and Foster, 1972). Although usual sleep characteristics such as REM latency, sleep effi ciency, and proportion of N3 decrease with age and vary according to 3.4. Sleep continuity Pearson’s chi-square test for categorical variables and one-way ANOVA for continuous variables were used as appropriate. (Х2, Fdegrees of freedom). Kruskal Wallis non-parametric test for Time in remission (kw Х2). † Generalized anxiety disorder, agoraphobia, social phobia, panic disorder;. * Table 2 Characteristics of participants according to major depressive disorder subtype (N = 1820). ants according to major depressive disorder subtype (N = 1820 Descriptive statistics are presented as mean (standard deviation) for continuous variables and n (percentage) for categorical variables. Pearson’s chi-square test for categorical variables and one-way ANOVA for continuous variables were used as appropriate. (Х2, Fdegrees of freedom). Kruskal Wallis non-parametric test for Time in remission (kw Х2). † † Generalized anxiety disorder, agoraphobia, social phobia, panic disorder;. * Alcohol and drug abuse or dependence * Alcohol and drug abuse or dependence Alcohol and drug abuse or dependence BMI: body mass index, MDD: major depressive disorder, SD: standard deviation, y: years, AHI: apnea-hypopnea index, PLMSI: periodic leg movement during sleep index; GAF depression Global Assessment of Functioning; MDE: major depressive disorder; IQR: interquartile range.; PSG: Polysomnography. g p BMI: body mass index, MDD: major depressive disorder, SD: standard deviation, y: years, AHI: apnea-hypopnea index, PLMSI: periodic leg movement during sleep index; GAF depression Global Assessment of Functioning; MDE: major depressive disorder; IQR: interquartile range.; PSG: Polysomnography. proportions of individuals with the melancholic MDD subtype. The re striction of the association between absolute delta-power and the melancholic MDD subtype or any MDD to individuals with a current episode suggests that this is dependent on the MDD state. MDD severity, REM density has been shown to be stable across age and MDD severity (Lauer et al., 1991; Wichniak et al., 2000). Surprisingly, in our study, positive associations with REM density were observed only in individuals with any remitted or any unspecified MDD rather than current depressive episodes. Although the cross-sectional nature of our data did not allow us to determine the temporal relationship between the onset of depressive episodes and the beginning of alterations in REM density, we hypothesize that elevated REM density is a premorbid trait in individuals with unspecified MDD (Pillai et al., 2011). MDD severity, REM density has been shown to be stable across age and MDD severity (Lauer et al., 1991; Wichniak et al., 2000). 4.4. Limitations Our results need to be viewed in the light of several limitations. First, the cross-sectional nature of the data did not allow us to draw conclu sions regarding the direction of the observed associations between MDD subtypes and sleep features. Second, we did not control for multiple comparison explaining that our analyses have an exploratory character and should first be confirmed in other samples. Thirdly, because the psychiatric assessment took place after the PSG investigation, inaccurate recall bias may have influenced the establishment of remission status at the time of the sleep investigation. However, this would have been most likely to underestimate the number of cases of MDD and reduce levels of statistical significance. Forth, despite the large size of the total sample, the number of individuals with current melancholic (n = 35) and current atypical (n = 25) MDD was small, raising the possibility that associations may not have been detected because of limited statistical power. Fifth, some participants were treated with psychotropic drugs that may in fluence sleep, but the models were adjusted for these medications. Sixth, we did not collect information on psychotherapy for sleep disorders. However, it seems unlikely to us that a significant proportion of the cohort was undergoing such treatment at the time of the PSG exami nation. Similarly, we do not have data on physical activity on the day of the PSG examination, which could have influenced the results. Fig. 2. Associations between major depressive disorder subtypes and poly somnography (PSG)-based sleep features according to multiple-adjusted robust linear regression models. Fig. 2. Associations between major depressive disorder subtypes and poly somnography (PSG)-based sleep features according to multiple-adjusted robust linear regression models. Adjustment for age, sex, body mass index, current smoking status, use of psy chotropic drugs (antidepressants and anxiolytics), anxiety and substance use disorders. Adjustment for age, sex, body mass index, current smoking status, use of psy chotropic drugs (antidepressants and anxiolytics), anxiety and substance use disorders. The color scale (between –1 and 1) is a visual representation of the unstan dardized beta coefficients for each PSG-based sleep features with respect to the mathematical signs. More intense red color indicates more positive unstan dardized beta coefficient value for the corresponding PSG-based sleep feature; more intense blue color indicates more negative unstandardized beta coefficient value for the PSG-based sleep feature. p<0.05. 4.4. Limitations N3 (%): proportion of total sleep time spent in the N3 sleep stage, SOL: sleep onset latency, WASO: wake after sleep onset, TST: total sleep time, SE: sleep efficiency; REML: rapid eye movements latency; REMD: rapid eye movements density; REM: rapid eye movement sleep. 4.3. Sleep continuity MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements density, REML: rapid eye movements latency, SE: sleep efficiency, SOL: sleep onset latency, TST: total sleep time, WASO: wake after sleep onset. may not be the optimal tool for assessing TST because it was performed over a limited time period (i.e., during the night) rather than over 24 h, thus potentially underestimating TST (e.g. in case of a daytime nap). Other potential explanations for the absence of associations between TST and MDD in our study were the milder severity of MDD in our population-based sample compared with clinical studies, and previously described discordance between subjective complaints of insomnia and objective EEG abnormalities (Andrillon et al., 2020; Castro et al., 2013; Zhang and Zhao, 2007). Fig. 2. Associations between major depressive disorder subtypes and poly somnography (PSG)-based sleep features according to multiple-adjusted robust linear regression models. Adjustment for age, sex, body mass index, current smoking status, use of psy chotropic drugs (antidepressants and anxiolytics), anxiety and substance use di d may not be the optimal tool for assessing TST because it was performed over a limited time period (i.e., during the night) rather than over 24 h, thus potentially underestimating TST (e.g. in case of a daytime nap). Other potential explanations for the absence of associations between TST and MDD in our study were the milder severity of MDD in our population-based sample compared with clinical studies, and previously described discordance between subjective complaints of insomnia and objective EEG abnormalities (Andrillon et al., 2020; Castro et al., 2013; Zhang and Zhao, 2007). 4.3. Sleep continuity We observed a significant increase in wake after sleep onset and a decrease in sleep efficiency in individuals with melancholic MDD, such an association was found only for current patients at the time of the PSG. 5 Psychiatry Research 324 (2023) 115213 G. Solelhac et al. G. Solelhac et al. Table 3 Objective polysomnography (PSG)-based sleep features according to major depressive disorder subtype. Table 3 Objective polysomnography (PSG)-based sleep features according to major depressive disorder subtype. Major depressive disorder subtype status No MDD Remitted Unspecified Remitted Melancholic Remitted Atypical Current Unspecified Current Melancholic Current Atypical N = 993 N = 417 N = 189 N = 104 N = 57 N = 35 N = 25 Slow wave sleep N3 (%) 19.3 (8.3) 21.1 (7.9) 20.1 (8.3) 21.1 (8.5) 19.9 (9.1) 19.0 (7.5) 21.6 (11.0) Delta power (µV2/Hz) 117.5 (82.0) 124.5 (63.7) 133.8 (158.1) 134.8 (75.4) 147.1 (111.6) 105.6 (67.4) 123.8 (63.7) Sleep continuity SOL (min) 16.5 (21.4) 17.4 (24.2) 21.4 (23.0) 15.7 (24.1) 20.7 (28.3) 15.9 (21.4) 18.1 (13.6) WASO (min) 76.9 (57.5) 67.0 (50.1) 70.3 (53.3) 56.8 (50.0) 61.9 (55.5) 79.7 (59.7) 51.0 (47.7) TST (min) 395.1 (67.4) 407.9 (71.0) 407.4 (71.0) 421.9 (80.8) 395.6 (78.9) 409.2 (76.1) 401.1 (67.1) SE (%) 84.1 (10.9) 86.1 (9.6) 85.6 (10.3) 88.4 (9.6) 86.5 (13.2) 84.5 (9.8) 88.7 (10.2) REM pressure REML (min) 89.5 (52.1) 92.6 (58.8) 96.0 (63.3) 97.1 (63.8) 118.7 (88.1) 127.1 (81.0) 103.5 (57.3) REMD (/min) 3.0 (2.3) 3.6 (2.9) 3.4 (3.1) 3.1 (2.2) 3.2 (2.7) 4.0 (2.8) 3.1 (2.2) REM (%) 21.7 (6.1) 22.4 (5.8) 22.9 (5.9) 22.0 (6.7) 20.9 (7.6) 21.2 (5.7) 21.2 (6.3) Values are mean (standard deviation). MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements density, REML: rapid eye movements latency, SE: sleep efficiency, SOL: sleep onset latency, TST: total sleep time, WASO: wake after sleep onset. Table 3 Objective polysomnography (PSG)-based sleep features according to major depressive disorder subtype. Values are mean (standard deviation). %): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements latency, SE: sleep efficiency, SOL: sleep onset latency, TST: total sleep time, WASO: wake after sleep onset. Table 5 Table 5 Associations between major depressive disorder subtypes and sleep continuity (multiple-adjusted robust linear regression models). SOL (min) WASO (min) TST (min) SE (%) β 95% CI β 95% CI β 95% CI β 95% CI Current MDD (any) 0.12 (−2.71,2.95) 6.40 (−2.10,14.91) −6.54 (−20.93,7.85) −1.40 (−3.18,0.38) Current atypical MDD 4.44 (−0.13,9.01) −7.72 (−23.95,8.50) −12.04 (−39.80,15.71) 1.13 (−3.00,5.26) Current melancholic MDD −2.10 (−6.06,1.85) 22.43* (8.51,36.34) 10.09 (−14.47,34.64) ¡3.60* (¡7.10,¡0.11) Current unspecified MDD −0.56 (−3.88,2.75) 4.43 (−6.96,15.82) −11.33 (−30.76,8.10) −1.03 (−3.91,1.84) Remitted MDD (any) −0.50 (−1.86,0.86) −1.94 (−6.08,2.20) 6.21 (−0.78,13.19) 0.34 (−0.53,1.20) Remitted atypical MDD −2.03 (−4.38,0.32) −5.46 (−13.72,2.81) 13.04 (−1.42,27.49) 1.24 (−0.83,3.31) Remitted melancholic MDD 1.73 (−0.14,3.60) −2.70 (−9.24,3.83) 4.62 (−6.41,15.65) 0.32 (−1.29,1.94) Remitted unspecified MDD −1.02 (−2.38,0.33) −0.96 (−5.71,3.80) 5.39 (−2.77,13.56) 0.12 (−1.06,1.31) No MDD (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) Adjustment for age, sex, body mass index, current smoking status, apnea-hypopnea index, periodic leg movements during sleep index, use of psychotropic drugs (antidepressants and anxiolytics), anxiety and substance use disorders. 95% CI: 95% confidence interval; MDD: major depressive disorder, SE: sleep efficiency, SOL: sleep onset latency, TST: total sleep time, WASO: wake after sleep onset. * p <0.05;. ** epressive disorder subtypes and sleep continuity (multiple-adjusted robust linear regression models). Adjustment for age, sex, body mass index, current smoking status, apnea-hypopnea index, periodic leg movements during sleep index, use of psychotropic drugs (antidepressants and anxiolytics), anxiety and substance use disorders. ii p y y 95% CI: 95% confidence interval; MDD: major depressive disorder, SE: sleep efficiency, SOL: sleep onset latency, TST: total sleep time, WASO: wake after sleep onset. * p <0.05;. Declaration of Competing Interest The authors have no conflicts of interest to declare. Table 6 Associations between major depressive disorder subtypes and apnea hypopnea index (multiple-adjusted robust linear regression models). The HypnoLaus and the CoLaus\|PsyCoLaus studies were and are supported by research grant from GlaxoSmithKline, the Faculty of Biology and Medicine of Lausanne, the Swiss National Science Foun dation (grants 3200B0–105993, 3200B0–118308, 33CSCO-122661, 33CS30–139468, 33CS30–148401, 33CS30_177535 and 3247730_204523), the Swiss Personalized Health Network (project: Swiss Ageing Citizen Reference, (grant 2018DRI01)), Leenaards Foun dation and Vaud Pulmonary League (Ligue Pulmonaire Vaudoise). The contents of this research are solely the responsibility of the authors. FS is supported by the European Research Council Starting Grant 101039782 DREAMSCAPE. AHI (events/h of sleep) β 95% CI Current MDD (any) −0.49 (−2.32,1.34) Current atypical MDD −1.07 (−5.02,2.88) Current melancholic MDD 1.77 (−1.69,5.23) Current unspecified MDD −0.60 (−3.27,2.07) Remitted MDD (any) −0.14 (−1.06,0.78) Remitted atypical MDD 1.46 (−0.56,3.48) Remitted melancholic MDD 0.08 (−1.63,1.46) Remitted unspecified MDD −0.41 (−1.55,0.72) No MDD (ref) 0 (ref) Adjustment for age, sex, body mass index. 95% CI: 95% confidence interval; MDD: major depressive disorder, AHI: apnea hypopnea index. p <0.05. Adjustment for age, sex, body mass index. 95% CI: 95% confidence interval; MDD: major depressive disorder, AHI: apnea hypopnea index. p <0.05. 4.5. Conclusion Slow wave sleep REM pressure N3 (%) Delta power (µV2/Hz) REML (min) REMD (/min) REM (%) β 95% CI β 95% CI β 95% CI β 95% CI β 95% CI Current MDD (any) −0.27 (−1.95,1.41) ¡16.74 (¡27.89,¡5.58) 3.14 (−5.05,11.32) −0.06 (−0.46,0.35) −0.87 (−2.12,0.38) Current atypical MDD 0.08 (−3.25,3.40) −2.61 (−23.66,18.44) −7.18 (−22.99,8.64) −0.20 (−1.01,0.60) −1.04 (−3.48,1.40) Current melancholic MDD −0.93 (−3.71,1.86) ¡30.05 (¡48.86,¡11.23) 13.68 (−0.33,27.68) 0.45 (−0.29,1.18) −0.96 (−3.06,1.13) Current unspecified MDD 0.02 (−2.25,2.30) −15.36 (−30.91,0.18) 3.97 (−7.17,15.11) −0.22 (−0.78,0.34) −0.69 (−2.41,1.03) Remitted MDD (any) 0.43 (−0.37,1.24) 1.80 (−3.41,7.01) −0.51 (−4.32,3.30) 0.23* (0.03,0.42) 0.13 (−0.47,0.74) Remitted atypical MDD 0.31 (−1.35,1.97) 3.65 (−7.12,14.43) 3.32 (−4.40,11.03) 0.07 (−0.34,0.48) 0.01 (−1.23,1.25) Remitted melancholic MDD 0.00 (−1.27,1.28) 0.10 (−8.17,8.38) −2.65 (−8.73,3.44) 0.08 (−0.24,0.40) 0.52 (−0.43,1.48) Remitted unspecified MDD 0.65 (−0.29,1.59) 2.36 (−3.69,8.42) −0.36 (−4.80,4.08) 0.34** (0.11,0.57) 0.00 (−0.71,0.70) No MDD (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) Adjustment for age, sex, body mass index, current smoking status, apnea-hypopnea index, periodic leg movements during sleep index, use of psychotropic drugs (antidepressants and anxiolytics), anxiety and substance use disorders. 95% CI: 95% confidence interval; MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements density, REML: rapid eye movements latency. * p <0.05;. p <0.01. Table 5 Associations between major depressive disorder subtypes and sleep continuity (multiple-adjusted robust linear regression models). SOL (min) WASO (min) TST (min) SE (%) β 95% CI β 95% CI β 95% CI β 95% CI Current MDD (any) 0.12 (−2.71,2.95) 6.40 (−2.10,14.91) −6.54 (−20.93,7.85) −1.40 (−3.18,0.38) Current atypical MDD 4.44 (−0.13,9.01) −7.72 (−23.95,8.50) −12.04 (−39.80,15.71) 1.13 (−3.00,5.26) C t l h li MDD 2 10 ( 6 06 1 85) 22 43 (8 51 36 34) 10 09 ( 14 47 34 64) 3 60* ( 7 10 0 11) Table 4 Associations between major depressive disorder subtypes and slow wave sleep and rapid-eye movement pressure (multiple-adjusted robust linear regression models). p y , y 95% CI: 95% confidence interval; MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements density, REML: rapid eye movements latency. * <0 05 4.5. Conclusion Slow wave sleep REM pressure N3 (%) Delta power (µV2/Hz) REML (min) REMD (/min) REM (%) β 95% CI β 95% CI β 95% CI β 95% CI β 95% CI Current MDD (any) −0.27 (−1.95,1.41) ¡16.74 (¡27.89,¡5.58) 3.14 (−5.05,11.32) −0.06 (−0.46,0.35) −0.87 (−2.12,0.38) Current atypical MDD 0.08 (−3.25,3.40) −2.61 (−23.66,18.44) −7.18 (−22.99,8.64) −0.20 (−1.01,0.60) −1.04 (−3.48,1.40) Current melancholic MDD −0.93 (−3.71,1.86) ¡30.05 (¡48.86,¡11.23) 13.68 (−0.33,27.68) 0.45 (−0.29,1.18) −0.96 (−3.06,1.13) Current unspecified MDD 0.02 (−2.25,2.30) −15.36 (−30.91,0.18) 3.97 (−7.17,15.11) −0.22 (−0.78,0.34) −0.69 (−2.41,1.03) Remitted MDD (any) 0.43 (−0.37,1.24) 1.80 (−3.41,7.01) −0.51 (−4.32,3.30) 0.23* (0.03,0.42) 0.13 (−0.47,0.74) Remitted atypical MDD 0.31 (−1.35,1.97) 3.65 (−7.12,14.43) 3.32 (−4.40,11.03) 0.07 (−0.34,0.48) 0.01 (−1.23,1.25) Remitted melancholic MDD 0.00 (−1.27,1.28) 0.10 (−8.17,8.38) −2.65 (−8.73,3.44) 0.08 (−0.24,0.40) 0.52 (−0.43,1.48) Remitted unspecified MDD 0.65 (−0.29,1.59) 2.36 (−3.69,8.42) −0.36 (−4.80,4.08) 0.34** (0.11,0.57) 0.00 (−0.71,0.70) No MDD (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) Adjustment for age, sex, body mass index, current smoking status, apnea-hypopnea index, periodic leg movements during sleep index, use of psychotropic drugs (antidepressants and anxiolytics), anxiety and substance use disorders. 95% CI: 95% confidence interval; MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements density, REML: rapid eye movements latency. * p <0 05; Table 4 Associations between major depressive disorder subtypes and slow wave sleep and rapid-eye movement pressure (multiple-adjusted robust linear regression models). Associations between major depressive disorder subtypes and slow wave sleep and rapid-eye movement pressure (multiple-adjusted robust linear regression models). Slow wave sleep REM pressure Adjustment for age, sex, body mass index, current smoking status, apnea-hypopnea index, periodic leg movements during sleep index, use of psychotropic drugs (antidepressants and anxiolytics), anxiety and substance use disorders. 95% CI: 95% confidence interval; MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements density, REML: rapid eye movements latency. * p <0.05;. ** p <0.01. Table 5 Associations between major depressive disorder subtypes and sleep continuity (multiple-adjusted robust linear regression models). 95% CI: 95% confidence interval; MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement sleep, REMD: rapid eye movements density, REML: rapid eye movements latency. * 0 05 4.5. Conclusion Our findings revealed that even in individuals with a current depressive episode, there was only a small number of associations be tween MDD and PSG-based sleep features. In addition, these associations varied by MDD subtype and were most common in those with the melancholic subtype, providing some additional support to the perti nence of subtyping MDD. Interestingly, REM density was not dependent on MDD state (current vs. remitted) and could be a risk factor for the onset of MDD or a residual symptom of depressive episodes. In the future, prospective studies that analyze the predictive value of poly somnography on incident MDD subtypes are needed. These results are in line with previous studies investigating sleep and MDD (Pillai et al., 2011). Interestingly, these objective insomnia char acteristics typical of the complaint of patients with melancholic MDD are no longer present in the remitted patients. No other associations between MDD subtypes and sleep continuity features were found, although we expected an increase in TST in individuals with atypical MDD and a decrease in those with other MDD subtypes. However, PSG 6 Psychiatry Research 324 (2023) 1152 Funding The HypnoLaus and the CoLaus\|PsyCoLaus studies were and ar supported by research grant from GlaxoSmithKline, the Faculty o Biology and Medicine of Lausanne, the Swiss National Science Foun dation (grants 3200B0–105993, 3200B0–118308, 33CSCO-122661 33CS30–139468, 33CS30–148401, 33CS30_177535 an 3247730_204523), the Swiss Personalized Health Network (projec Swiss Ageing Citizen Reference, (grant 2018DRI01)), Leenaards Foun dation and Vaud Pulmonary League (Ligue Pulmonaire Vaudoise). Th contents of this research are solely the responsibility of the authors. FS supported by the European Research Council Starting Grant 10103978 DREAMSCAPE. 4.5. Conclusion D l i f C i I Table 4 Associations between major depressive disorder subtypes and slow wave sleep and rapid-eye movement pressure (multiple-adjusted robust linear regression models Slow wave sleep REM pressure N3 (%) Delta power (µV2/Hz) REML (min) REMD (/min) REM (%) β 95% CI β 95% CI β 95% CI β 95% CI β 95% CI Current MDD (any) −0.27 (−1.95,1.41) ¡16.74 (¡27.89,¡5.58) 3.14 (−5.05,11.32) −0.06 (−0.46,0.35) −0.87 (−2.12,0.38) Current atypical MDD 0.08 (−3.25,3.40) −2.61 (−23.66,18.44) −7.18 (−22.99,8.64) −0.20 (−1.01,0.60) −1.04 (−3.48,1.40) Current melancholic MDD −0.93 (−3.71,1.86) ¡30.05 (¡48.86,¡11.23) 13.68 (−0.33,27.68) 0.45 (−0.29,1.18) −0.96 (−3.06,1.13) Current unspecified MDD 0.02 (−2.25,2.30) −15.36 (−30.91,0.18) 3.97 (−7.17,15.11) −0.22 (−0.78,0.34) −0.69 (−2.41,1.03) Remitted MDD (any) 0.43 (−0.37,1.24) 1.80 (−3.41,7.01) −0.51 (−4.32,3.30) 0.23* (0.03,0.42) 0.13 (−0.47,0.74) Remitted atypical MDD 0.31 (−1.35,1.97) 3.65 (−7.12,14.43) 3.32 (−4.40,11.03) 0.07 (−0.34,0.48) 0.01 (−1.23,1.25) Remitted melancholic MDD 0.00 (−1.27,1.28) 0.10 (−8.17,8.38) −2.65 (−8.73,3.44) 0.08 (−0.24,0.40) 0.52 (−0.43,1.48) Remitted unspecified MDD 0.65 (−0.29,1.59) 2.36 (−3.69,8.42) −0.36 (−4.80,4.08) 0.34** (0.11,0.57) 0.00 (−0.71,0.70) No MDD (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) Adjustment for age, sex, body mass index, current smoking status, apnea-hypopnea index, periodic leg movements during sleep index, use of psychotropic drug (antidepressants and anxiolytics), anxiety and substance use disorders. 95% CI: 95% confidence interval; MDD: major depressive disorder, N3 (%): proportion of total sleep time spent in the N3 sleep stage, REM: rapid eye movement slee REMD: rapid eye movements density, REML: rapid eye movements latency. * p <0.05;. p <0.01. Table 5 Associations between major depressive disorder subtypes and sleep continuity (multiple-adjusted robust linear regression models). 4.5. Conclusion SOL (min) WASO (min) TST (min) SE (%) β 95% CI β 95% CI β 95% CI β 95% CI Current MDD (any) 0.12 (−2.71,2.95) 6.40 (−2.10,14.91) −6.54 (−20.93,7.85) −1.40 (−3.18,0.38) Current atypical MDD 4.44 (−0.13,9.01) −7.72 (−23.95,8.50) −12.04 (−39.80,15.71) 1.13 (−3.00,5.26) Current melancholic MDD −2.10 (−6.06,1.85) 22.43 (8.51,36.34) 10.09 (−14.47,34.64) ¡3.60* (¡7.10,¡0.11) Current unspecified MDD −0.56 (−3.88,2.75) 4.43 (−6.96,15.82) −11.33 (−30.76,8.10) −1.03 (−3.91,1.84) Remitted MDD (any) −0.50 (−1.86,0.86) −1.94 (−6.08,2.20) 6.21 (−0.78,13.19) 0.34 (−0.53,1.20) Remitted atypical MDD −2.03 (−4.38,0.32) −5.46 (−13.72,2.81) 13.04 (−1.42,27.49) 1.24 (−0.83,3.31) Remitted melancholic MDD 1.73 (−0.14,3.60) −2.70 (−9.24,3.83) 4.62 (−6.41,15.65) 0.32 (−1.29,1.94) Remitted unspecified MDD −1.02 (−2.38,0.33) −0.96 (−5.71,3.80) 5.39 (−2.77,13.56) 0.12 (−1.06,1.31) No MDD (ref) 0 (ref) 0 (ref) 0 (ref) 0 (ref) Adjustment for age, sex, body mass index, current smoking status, apnea-hypopnea index, periodic leg movements during sleep index, use of psychotropic drug (antidepressants and anxiolytics), anxiety and substance use disorders. 95% CI: 95% confidence interval; MDD: major depressive disorder, SE: sleep efficiency, SOL: sleep onset latency, TST: total sleep time, WASO: wake after sleep onse * p <0.05;. ** p <0.01. Table 6 Associations between major depressive disorder subtypes and apnea hypopnea index (multiple-adjusted robust linear regression models). AHI (events/h of sleep) β 95% CI Current MDD (any) −0.49 (−2.32,1.34) Current atypical MDD −1.07 (−5.02,2.88) Current melancholic MDD 1.77 (−1.69,5.23) Current unspecified MDD −0.60 (−3.27,2.07) Remitted MDD (any) −0.14 (−1.06,0.78) Remitted atypical MDD 1.46 (−0.56,3.48) Remitted melancholic MDD 0.08 (−1.63,1.46) Remitted unspecified MDD −0.41 (−1.55,0.72) No MDD (ref) 0 (ref) Adjustment for age sex body mass index 95% CI: 95% confidence interval; G. Solelhac et al. Psychiatry Research 324 (2023) 115213 G. Solelhac et al. Table 4 Table 4 Associations between major depressive disorder subtypes and slow wave sleep and rapid-eye movement pressure (multiple-adjusted robust linear regression models). Table 4 Associations between major depressive disorder subtypes and slow wave sleep and rapid-eye movement pressure (multiple-adjusted robust linear regression models). References Agarwal, R., Takeuchi, T., Laroche, S., Gotman, J., 2005. Detection of Rapid-Eye Movements in Sleep Studies. IEEE Trans. Biomed. Eng. 52 (8), 1390–1396. https:// doi.org/10.1109/TBME.2005.851512. Giles, D.E., Roffwarg, H.P., Rush, A.J., Guzick, D.S., 1990. Age-adjusted threshold values for reduced REM latency in unipolar depression using ROC analysis. Biol. Psychiatry 27 (8), 841‑853. https://doi.org/10.1016/0006-3223(90)90465-e. American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders 5th ed. Heinzer, R., Vat, S., Marques-Vidal, P., Marti-Soler, H., Andries, D., Tobback, N., Mooser, V., Preisig, M., Malhotra, A., Waeber, G., Vollenweider, P., Tafti, M., Haba- Rubio, J., 2015. Prevalence of sleep-disordered breathing in the general population : the HypnoLaus study. Lancet Respir. Med. 3 (4), 310‑318. https://doi.org/10.1016/ S2213-2600(15)00043-0. Andrillon, T., Solelhac, G., Bouchequet, P., Romano, F., Le Brun, M.-.P., Brigham, M., Chennaoui, M., L´eger, D., 2020. Revisiting the value of polysomnographic data in insomnia : more than meets the eye. Sleep Med. 66, 184‑200. https://doi.org/ 10.1016/j.sleep.2019.12.002. Hubain, P.P., Souery, D., J¨onck, L., Staner, L., Van Veeren, C., Kerkhofs, M., Mendlewicz, J., Linkowski, P., 1995. Relationship between the Newcastle scale and sleep polysomnographic variables in major depression : a controlled study. Eur. Neuropsychopharmacol. 5 (2), 129‑134. https://doi.org/10.1016/0924-977X(95) 00011-D. Antonijevic, I.A., 2006. Depressive disorders—Is it time to endorse different pathophysiologies? Psychoneuroendocrinology 31 (1), 1‑15. https://doi.org/ 10.1016/j.psyneuen.2005.04.004. Armitage, R., Hoffmann, R., Trivedi, M., Rush, A.J., 2000. Slow-wave activity in NREM sleep : sex and age effects in depressed outpatients and healthy controls. Psychiatry Res. 95 (3), 201‑213. https://doi.org/10.1016/s0165-1781(00)00178-5. Iber, C., 2007. The AASM Manual for the Scoring of Sleep and Associated Events : Rules, Terminology and Technical Specifications. American Academy of Sleep Medicine. i Iorio, G., Marciano, F., Martino, M., Kemali, D., 1994. Statistical comparison of transition sleep variables in depressed and normal subjects. Eur. Psychiatry 9 (2), 95‑100. https://doi.org/10.1017/S0924933800001826. Baglioni, C., Battagliese, G., Feige, B., Spiegelhalder, K., Nissen, C., Voderholzer, U., Lombardo, C., Riemann, D., 2011. Insomnia as a predictor of depression : a meta- analytic evaluation of longitudinal epidemiological studies. J. Affect. Disord. 135 (1‑3), 10‑19. https://doi.org/10.1016/j.jad.2011.01.011. p g Jaussent, I., Bouyer, J., Ancelin, M.-.L., Akbaraly, T., P´er`es, K., Ritchie, K., Besset, A., Dauvilliers, Y., 2011. Insomnia and daytime sleepiness are risk factors for depressive symptoms in the elderly. Sleep 34 (8), 1103‑1110. https://doi.org/10.5665/ SLEEP.1170. Baglioni, C., Nanovska, S., Regen, W., Spiegelhalder, K., Feige, B., Nissen, C., Reynolds, C.F., Riemann, D., 2016. Sleep and mental disorders : a meta-analysis of polysomnographic research. Psychol. Bull. References Musliner, K.L., Munk-Olsen, T., Laursen, T.M., Eaton, W.W., Zandi, P.P., Mortensen, P.B., 2016. Heterogeneity in 10-year course trajectories of moderate to severe major depressive disorder. JAMA Psychiatry 73 (4), 346‑353. https://doi.org/10.1001/ jamapsychiatry.2015.3365. Buysse, D.J., Angst, J., Gamma, A., Ajdacic, V., Eich, D., R¨ossler, W., 2008. Prevalence, course, and comorbidity of insomnia and depression in young adults. Sleep 31 (4), 473‑480. https://doi.org/10.1093/sleep/31.4.473. i j p y y Nurnberger Jr., J.I., Blehar, M.C., Kaufmann, C.A., York-Cooler, C., Simpson, S.G., Harkavy-Friedman, J., Severe, J.B., Malaspina, D., Reich, T, 1994. Diagnostic interview for genetic studies : rationale, unique features, and training. Arch. Gen. Psychiatry 51 (11), 849‑859. https://doi.org/10.1001/ archpsyc.1994.03950110009002. p g p Castro, L.S., Poyares, D., Leger, D., Bittencourt, L., Tufik, S., 2013. Objective prevalence of insomnia in the S˜ao Paulo, Brazil epidemiologic sleep study. Ann. Neurol. 74 (4), 537‑546. https://doi.org/10.1002/ana.23945. p g Chang, P.P., Ford, D.E., Mead, L.A., Cooper-Patrick, L., Klag, M.J., 1997. Insomnia in young men and subsequent depression. The Johns Hopkins Precursors Study. J. Epidemiol. 146 (2), 105‑114. https://doi.org/10.1093/oxfordjournals.aje. a009241. p y Palagini, L., Baglioni, C., Ciapparelli, A., Gemignani, A., Riemann, D., 2013. REM sleep dysregulation in depression : state of the art. Sleep Med. 17 (5), 377‑390. https:// doi.org/10.1016/j.smrv.2012.11.001. Paunio, T., Korhonen, T., Hublin, C., Partinen, M., Koskenvuo, K., Koskenvuo, M., Kaprio, J., 2015. Poor sleep predicts symptoms of depression and disability retirement due to depression. J. Affect. Disord. 172, 381‑389. https://doi.org/ 10.1016/j.jad.2014.10.002. Dippel, B., Lauer, C., Riemann, D., Majer-Trendel, K., Krieg, J.C., Berger, M., 1987. Sleep and dreams in eating disorders. Psychother. Psychosom. 48 (1‑4), 165‑169. https:// doi.org/10.1159/000288048. Feinberg, M., Gillin, J.C., Carroll, B.J., Greden, J.F., Zis, A.P., 1982. EEG studies of sleep in the diagnosis of depression. Biol. Psychiatry 17 (3), 305‑316. Pillai, V., Kalmbach, D.A., Ciesla, J.A., 2011. A meta-analysis of electroencephalographic sleep in depression : evidence for genetic biomarkers. Biol. Psychiatry 70 (10), 912‑919. https://doi.org/10.1016/j.biopsych.2011.07.016. Firmann, M., Mayor, V., Vidal, P.M., Bochud, M., P´ecoud, A., Hayoz, D., Paccaud, F., Preisig, M., Song, K.S., Yuan, X., Danoff, T.M., Stirnadel, H.A., Waterworth, D., Mooser, V., Waeber, G., Vollenweider, P., 2008. The CoLaus study : a population- based study to investigate the epidemiology and genetic determinants of cardiovascular risk factors and metabolic syndrome. BMC Cardiovasc. Disord. 8, 6. https://doi.org/10.1186/1471-2261-8-6. Plante, D.T., Cook, J.D., Goldstein, M.R., 2017. References 142 (9), 969‑990. https://doi.org/ 10.1037/bul0000053. Kupfer, D.J., Foster, F.G., 1972. Interval between onset of sleep and rapid-eye-movement sleep as an indicator of depression. Lancet 2 (7779), 684‑686. https://doi.org/ 10.1016/s0140-6736(72)92090-9. Bell, C.C., 1994. DSM-IV : diagnostic and Statistical Manual of Mental Disorders. JAMA 272 (10), 828‑829. https://doi.org/10.1001/jama.1994.03520100096046. Lauer, C.J., Riemann, D., Wiegand, M., Berger, M., 1991. From early to late adulthood. Changes in EEG sleep of depressed patients and healthy volunteers. Biol. Psychiatry 29 (10), 979‑993. https://doi.org/10.1016/0006-3223(91)90355-p. Benca, R.M., Obermeyer, W.H., Thisted, R.A., Gillin, J.C., 1992. Sleep and Psychiatric Disorders : a Meta-analysis. Arch. Gen. Psychiatry 49 (8), 651‑668. https://doi.org/ 10.1001/archpsyc.1992.01820080059010. Berry, R.B., Budhiraja, R., Gottlieb, D.J., Gozal, D., Iber, C., Kapur, V.K., Marcus, C.L., Mehra, R., Parthasarathy, S., Quan, S.F., Redline, S., Strohl, K.P., Ward, S.L.D., Tangredi, M.M., 2012. Rules for Scoring Respiratory Events in Sleep : update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. J. Clin. Sleep Med. 08 (05), 597‑619. https://doi.org/10.5664/jcsm.2172. Leboyer, M., Barbe, B., Gorwood, P., Teherani, M., Allilaire, J.F., Preisig, M., Matthey, M. L., Poyetton, V., Ferrero, F., 1995. Interview diagnostique pour les ´etudes g´en´etiques. Inserm, Paris. g q Lecci, S., Cataldi, J., Betta, M., Bernardi, G., Heinzer, R., Siclari, F., 2020. Electroencephalographic changes associated with subjective under- and overestimation of sleep duration. Sleep 43 (11). https://doi.org/10.1093/sleep/ zsaa094. Blanken, T.F., Borsboom, D., Penninx, B.W.J.H., Van Someren, E.J.W., 2020. Network outcome analysis identifies difficulty initiating sleep as a primary target for prevention of depression : a 6-year prospective study. Sleep 43 (5), 1‑6. https://doi. org/10.1093/sleep/zsz288. Lee, E., Cho, H.J., Olmstead, R., Levin, M.J., Oxman, M.N., Irwin, M.R., 2013. Persistent sleep disturbance : a risk factor for recurrent depression in community-dwelling older adults. Sleep 36 (11), 1685‑1691. https://doi.org/10.5665/sleep.3128. g Borb´ely, A.A., Tobler, I., Loepfe, M., Kupfer, D.J., Ulrich, R.F., Grochocinski, V., Doman, J., Matthews, G., 1984. All-night spectral analysis of the sleep EEG in untreated depressives and normal controls. Psychiatry Res. 12 (1), 27‑33. https:// doi.org/10.1016/0165-1781(84)90135-5. L´eger, D., Debellemaniere, E., Rabat, A., Bayon, V., Benchenane, K., Chennaoui, M., 2018. Slow-wave sleep : from the cell to the clinic. Sleep Med. 41, 113‑132. https:// doi.org/10.1016/j.smrv.2018.01.008. Murphy, M.J., Peterson, M.J., 2015. Sleep disturbances in depression. Sleep Med. Clin. 10 (1), 17‑23. https://doi.org/10.1016/j.jsmc.2014.11.009. Bruun, C.F., Arnbjerg, C.J., Kessing, L.V., 2021. Electroencephalographic parameters differentiating melancholic depression, non-melancholic depression, and healthy controls. A systematic review. Front. Psychiatry 12, 648713. https://doi.org/ 10.3389/fpsyt.2021.648713. freelance medical writer, funded by the University Hospital of Lausanne. freelance medical writer, funded by the University Hospital of Lausanne. freelance medical writer, funded by the University Hospital of Lausanne. sociodemographic characteristics and psychiatric comorbidity in a population-based study. J. Affect. Disord. 226, 132‑141. https://doi.org/10.1016/j.jad.2017.09.032. Geoffroy, P.A., Palagini, L., 2021. Biological rhythms and chronotherapeutics in depression. Prog. Neuropsychopharmacol. 106, 110158 https://doi.org/10.1016/j. pnpbp.2020.110158. Acknowledgments GS, MB, MPS, MP and RH designed the study. GS performed statis tical analysis. GS wrote the first draft of manuscript. All authors inter preted the data, critically reviewed the manuscript and approved the final version. GS is the guarantor of this work and, as such, had (with RH) full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The authors thank Professor G´erard Waeber, Professor Mehdi Tafti, Professor Vincent Mooser, Daniela Andries, and Nadia Tobback for their important contributions to the HypnoLaus and CoLaus\|PsyCoLaus co horts, the Lausanne population that volunteered to participate in the CoLaus\|PsyCoLaus and HypnoLaus studies, and the entire CoLaus\|Psy CoLaus team. English writing assistance was provided by Nicola Ryan, a 7 G. Solelhac et al. Psychiatry Research 324 (2023) 115213 References Objective measures of sleep duration and continuity in major depressive disorder with comorbid hypersomnolence : a primary investigation with contiguous systematic review and meta-analysis. J. Sleep Res. 26 (3), 255‑265. https://doi.org/10.1111/jsr.12498. p g Ford, D.E., Cooper-Patrick, L., 2001. Sleep disturbances and mood disorders : an epidemiologic perspective. Depress. Anxiety 14 (1), 3‑6. https://doi.org/10.1002/ da.1041. Plante, D.T., Landsness, E.C., Peterson, M.J., Goldstein, M.R., Riedner, B.A., Wanger, T., Guokas, J.J., Tononi, G., Benca, R.M., 2012. Sex-related differences in sleep slow wave activity in major depressive disorder : a high-density EEG investigation. BMC Psychiatry 12, 146. https://doi.org/10.1186/1471-244X-12-146. Frank, E., Kupfer, D.J., Hamer, T., Grochocinski, V.J., McEachran, A.B., 1992. Maintenance treatment and psychobiologic correlates of endogenous subtypes. J. Affect. Disord. 25 (3), 181‑189. https://doi.org/10.1016/0165-0327(92)90004-p Preisig, M., Waeber, G., Vollenweider, P., Bovet, P., Rothen, S., Vandeleur, C., Guex, P., Middleton, L., Waterworth, D., Mooser, V., Tozzi, F., Muglia, P., 2009. The PsyCoLaus study : methodology and characteristics of the sample of a population- based survey on psychiatric disorders and their association with genetic and cardiovascular risk factors. BMC Psychiatry 9, 9. https://doi.org/10.1186/1471- 244X-9-9. Franzen, P.L., Buysse, D.J., 2008. Sleep disturbances and depression : risk relationships for subsequent depression and therapeutic implications. Dialogues Clin. Neurosci. 10 (4), 473‑481. Geoffroy, P.A., 2018. Clock genes and light signaling alterations in bipolar disorder : when the biological clock is off. Biol. Psychiatry 84 (11), 775‑777. https://doi.org/ 10.1016/j.biopsych.2018.09.006. Riemann, D., Voderholzer, U., 2003. Primary insomnia : a risk factor to develop depression? J. Affect. Disord. 76 (1‑3), 255‑259. https://doi.org/10.1016/s0165- 0327(02)00072-1. Geoffroy, P.A., Hoertel, N., Etain, B., Bellivier, F., Delorme, R., Limosin, F., Peyre, H., 2018. Insomnia and hypersomnia in major depressive episode : prevalence, 8 8 G. Solelhac et al. Psychiatry Research 324 (2023) 115213 Rush, A.J., Giles, D.E., Roffwarg, H.P., Parker, C.R., 1982. Sleep EEG and dexamethasone suppression test findings in outpatients with unipolar major depressive disorders. Biol. Psychiatry 17 (3), 327‑341. Wichniak, A., Riemann, D., Kiemen, A., Voderholzer, U., Jernajczyk, W., 2000. Comparison between eye movement latency and REM sleep parameters in major depression. Eur. Arch. Psychiatry Clin. Neurosci. 250 (1), 48‑52. https://doi.org/ 10.1007/s004060050009. Saitoh, K., Yoshiike, T., Kaneko, Y., Utsumi, T., Matsui, K., Nagao, K., Otsuki, R., Aritake- Okada, S., Kadotani, H., Kuriyama, K., Suzuki, M., 2022. Associations of nonrestorative sleep and insomnia symptoms with incident depressive symptoms over 1–2 years : longitudinal results from the Hispanic Community Health Study/ Study of Latinos and Sue˜no Ancillary Study. References Depress. Anxiety 39 (5), 419‑428. https://doi.org/10.1002/da.23258. Yetton, B.D., Niknazar, M., Duggan, K.A., McDevitt, E.A., Whitehurst, L.N., Sattari, N., Mednick, S.C., 2016. Automatic detection of rapid eye movements (REMs) : a machine learning approach. J. Neurosci. Methods 259, 72‑82. https://doi.org/ 10.1016/j.jneumeth.2015.11.015. Zhang, L., Zhao, Z.-X., 2007. Objective and subjective measures for sleep disorders. Neurosci. Bull. 23 (4), 236‑240. https://doi.org/10.1007/s12264-007-0035-9. p g Sitaram, N., Dube, S., Jones, D., Pohl, R., Gershon, S., 1984. Acetylcholine and alpha 1- adrenergic sensitivity in the separation of depression and anxiety. Psychopathology 17 (Suppl 3), 24‑39. https://doi.org/10.1159/000284129. Zucconi, M., Ferri, R., Allen, R., Baier, P.C., Bruni, O., Chokroverty, S., Ferini-Strambi, L., Fulda, S., Garcia-Borreguero, D., Hening, W.A., Hirshkowitz, M., H¨ogl, B., Hornyak, M., King, M., Montagna, P., Parrino, L., Plazzi, G., Terzano, M.G., International Restless Legs Syndrome Study Group (IRLSSG), 2006. The official World Association of Sleep Medicine (WASM) standards for recording and scoring periodic leg movements in sleep (PLMS) and wakefulness (PLMW) developed in collaboration with a task force from the International Restless Legs Syndrome Study Group (IRLSSG). Sleep Med. 7 (2), 175‑183. https://doi.org/10.1016/j. sleep.2006.01.001. Suh, S., Kim, H., Yang, H.-.C., Cho, E.R., Lee, S.K., Shin, C., 2013. Longitudinal course of depression scores with and without insomnia in non-depressed individuals : a 6-year follow-up longitudinal study in a Korean cohort. Sleep 36 (3), 369‑376. https://doi. org/10.5665/sleep.2452. Tsuno, N., Besset, A., Ritchie, K., 2005. Sleep and Depression. J. Clin. Psychiatry 66 (10), 1254‑1269. Vallat, R., Walker, M.P., 2021. An open-source, high-performance tool for automated sleep staging. Elife 10, e70092. https://doi.org/10.7554/eLife.70092.
https://openalex.org/W4385198651	https://www.researchsquare.com/article/rs-3171585/latest.pdf	English	null	The Effect of Two Different Types of Music Played to Cancer Patients During Chemotherapy on Anxiety, Nausea, and Satisfaction Levels	Research Square (Research Square)	2,023	cc-by	7,765	The Effect of Two Different Types of Music Played to Cancer Patients During Chemotherapy on Anxiety, Nausea, and Satisfaction Levels Öznur ERBAY DALLI (  oznurerbay@uludag.ed Uludağ University Derya AKÇA DOĞAN Uludağ University Seda PEHLİVAN Uludağ University Yasemin YILDIRIM Ege University Türkkan EVRENSEL Uludağ University Öznur ERBAY DALLI (  oznurerbay@uludag.edu.tr ) Derya AKÇA DOĞAN Uludağ University Seda PEHLİVAN Uludağ University Yasemin YILDIRIM Ege University Türkkan EVRENSEL Uludağ University Research Article Keywords: Chemothreapy, Cancer, Music, Nursing Posted Date: July 24th, 2023 Version of Record: A version of this preprint was published at Supportive Care in Cancer on November 20th, 2023. See the published version at https://doi.org/10.1007/s00520-023-08165-9. Page 1/16 Page 1/16 Page 1/16 Abstract Purpose: To investigate the effect of two different types of music on anxiety, nausea, and satisfaction levels in cancer patients receiving chemotherapy (CT) for the first time. Methods: The study was conducted as a single-blind, pre-test, post-test, three-group randomized controlled trial in an outpatient CT unit between August 2022 and February 2023. A simple (computer-based) and stratified (age and gender) randomization method was used to assign 75 patients to the relaxing music group (RMG), Turkish classical music group (TCMG), and control group (CG) (n=25 each). The primary outcome was the change in anxiety levels measured by Spielberger's State Anxiety Inventory before (T0) and after (T1) CT session. Secondary outcomes were the change in the severity of nausea from T0 to T1 and the level of satisfaction at T1. Results: The groups were similar in terms of baseline sociodemographic and health-related characteristics. Anxiety levels were lower than the baseline in RMG and TCMG in comparison to CG, and repeated measures analysis showed a significant group × time interaction (p=0.001, F=210.221, η2=0.745). Nausea severity increased from T0 to T1 for CG but decreased for RMG and TCMG with a significant group × time interaction (p=0.001, F=100.785, η2=0.583). The satisfaction level was significantly higher in TCMG than in CG and RMG (8.64±0.95 vs. 7.88±0.72 and 7.00±0.70, respectively). Conclusion: Music may be an effective non-pharmacologic option to relieve patients' anxiety and nausea during first- time CT, irrespective of music type. Larger, multicenter studies evaluating the long-term effect of music are needed to confirm these findings. Trial registration number/date: clinicaltrials.gov (NCT05687838) / 2022-13/18 Introduction Cancer is a major global health problem that has been on the rise, especially in the last 10 years, and is one of the leading causes of death worldwide [1, 2]. One of the most effective treatment methods for cancer is chemotherapy (CT) [3]. CT is a form of treatment with natural or synthetic chemicals and biological agents that have selective lethal effects, especially against rapidly proliferating cells [4]. Chemotherapeutic drugs prevent the growth and proliferation of cancer cells but also affect rapidly proliferating healthy cells such as intestinal and oral mucosal epithelium, bone marrow cells, and hair follicle cells [5]. Likewise, CT has several side effects such as myelosuppression, mucositis, nausea, vomiting, diarrhea, alopecia, fatigue, and pain, which altogether significantly affect patient comfort [6]. This situation may cause patients to worry about the development of such side effects and to experience anxiety before CT due to the procedures performed [7–9]. Studies have shown that the prevalence of CT-related anxiety in cancer patients ranges from 16.2– 26.7% [10, 11]. High levels of chemotherapy-induced anxiety can cause various physiological (rapid heartbeat, chest tightness, shortness of breath, etc.) and psychological (restlessness, inability to concentrate, etc.) distress by stimulating the stress response in patients [12, 13]. This can also trigger digestive symptoms such as nausea, vomiting, or diarrhea, leading to amplified side effects during and after CT [14–16]. Healthcare professionals and nurses who implement CT sessions must take precautions before CT to reduce the anxiety of cancer patients, provide comfort by minimizing the possible side effects of chemotherapy, and increase the effectiveness of treatment. Pharmacological interventions (antidepressants, benzodiazepines, antipsychotic drugs) in the management of CT- induced anxiety can also control anxiety-related symptoms such as nausea/vomiting [17, 18]. However, these drugs have side effects such as respiratory suppression, increased fatigue, decreased concentration, and causing sedation and Page 2/16 confusion [18]. Therefore, non-pharmacological interventions (music, relaxation, aromatherapy, etc.) are becoming preferable in managing the anxiety of cancer patients who have to cope with several other side effects related to CT [18, 19]. Compared to other non-pharmacologic approaches, music is a noninvasive and low-cost tool that is frequently used in every field and has been shown to be effective in the management of several adverse symptoms [20]. Music activates brain areas associated with memory, cognitive function, and emotions, helping to maintain cognitive function and reduce anxiety and stress levels [21]. Sample size calculation A priori power analysis was performed using the Power Analysis and Sample Size (PASS) program to determine the sample size needed in the study. The effect size of the study groups was calculated using the results of the anxiety level measured by STAI in a similar study conducted by Al-Jubouri et al. (2021) [25] and found to be d = 0.39. Using the calculated effect size, the sample size required to reach 80% power at the 0.05 significance level was calculated as 66 participants (22 in each group). After adjusting for a 10% attrition rate [26], the final sample size required was established as 74. Participants The inclusion criteria were (a) being 18 years old or older and (b) being diagnosed with any stage or type of cancer and scheduled to receive CT for the first time. Patients with known hearing or vision problems, an education level lower than middle school [the American Psychological Association reports that the State-Trait Anxiety Inventory (STAI) is appropriate to be administered for patients with an education level of sixth grade and above], a diagnosis of psychiatric or neurological disease, and previous cancer treatment, such as surgery or radiotherapy before CT treatment, were excluded from the study. Study design and setting This study was conducted in the outpatient CT unit of a university hospital in Bursa between August 2022 and February 2023 as a parallel, three-group randomized controlled trial with a prospective, single-blind, pre-test, post-test experimental design. The study was conducted according to the Consolidated Standards of Reporting Trials (CONSORT). The outpatient CT unit where the study was conducted has a capacity of 37 beds. A total of eight nurses worked in the unit, including one nurse-in-charge, one assistant nurse-in-charge, and six CT nurses. In the unit, an average of 50–60 patients are treated daily, including supportive therapies (blood transfusion, etc.). Introduction In addition to its overall positive effects, different types of music (cultural, melodic/harmonic, and rhythmic) may also modulate the individual's response to music [22]. Although there are studies showing the positive effect of music on the management of symptoms in patients receiving CT [14, 15, 23, 24], studies examining the effect of different music genres are limited. In this context, this study aimed to examine the effect of two different types of music on anxiety, nausea, and satisfaction levels in patients receiving CT for the first time. Relaxing music group In the RM group (RMG), "MusiCure® 5 Seasons" compositions, which have been specially developed by Niels Eje (Gefion Records, Copenhagen, Denmark), were preferred [31]. The compositions included melodies with harp, cello, strings, and sound elements from nature (rain, birds, forest sounds) in soft rhythm in the range of 60–80 bpm for relaxation purposes based on the research of the acoustic environment of hospitals and shown to have a positive effect on symptoms such as pain, well-being, and anxiety of patients [31]. Randomization and allocation A simple and stratified randomization method was used in the study. Stratified randomization aimed to control the differential effect of age and gender on the way of perceiving and responding to music, as reported in the literature [27, 28]. In stratified randomization, which also increases the reliability of the study, patients were randomized according to age (18–40, 41–63, 64–85 years) and gender (male-female) and assigned to study groups. In simple randomization, the groups to which patients were assigned were determined by a web-based computer program at Randomization.com (http://www.jerrydallal.com/random/randomize.htm). After the assignments were obtained, a bag of sealed envelopes Page 3/16 Page 3/16 containing the numbers was created, and the envelopes were drawn from the bag to determine which patient group was selected at that moment. containing the numbers was created, and the envelopes were drawn from the bag to determine which patient group was selected at that moment. Intervention Participants were randomly assigned to one of the following groups: relaxing music group (RMG), Turkish classical music group (TCMG), and control group (CG). Two researchers, nurse-in-charge and assistant-nurse-in-charge in the outpatient CT unit worked together to identify patients who met the inclusion criteria, prepare the necessary materials for the intervention groups, and conduct the pre-test/post-test administration. The researchers recorded information on adherence with the interventions, including start/end times of music, volume, and reasons for interruptions, if any. Patients in the intervention groups were asked open-ended questions at the end of the intervention about the acceptability of the choice and application of music, the appropriateness of the materials used, and how the music made them feel; all groups were asked open-ended questions about their expectations from the healthcare professionals and nurses during the CT process. Care was taken to select referenced headphones for the proper transmission of music and the proper utilization of neuro- acoustic therapies of music. For this purpose, neuroscience and music therapy-based recommendations were taken into account to choose headphones designed to isolate the subject from external noise to reduce distraction [29, 30]. Sennheiser HD 280 Pro over-ear headphones with the specified features were preferred. For hygienic purposes, disposable covers were placed around the headphones and replaced for each patient. The music selected by the researchers was transferred to digital media (mp3 format). The duration of the music was determined as at least 60 minutes, taking into account that the average duration of CT treatment is 1.5 to 2 hours. If the participants wanted to continue listening to the music after 60 minutes were completed, no intervention was made to stop the music. All patients received the standard premedication treatment (2x8 mg dexamethasone, 1x45.5 mg pheniramine, and 1x3 mg granisetron) before the CT. Blinding Due to the nature of the intervention, blinding the participants and the nursing researchers who conducted the intervention was not possible. However, the fact that the data analysis was carried out by a statistician independent of the study ensured that the results were blinded. Outcome measures After obtaining written consent from the participants, data were collected by two researchers using data collection tools in face-to-face interviews and through observation. The data collection tools consisted of Patient Identification Form, Patient Follow-up Form, STAI Turkish version, Nausea Assessment Form, and Satisfaction Assessment Form. The data collection tools were administered twice: before CT (T0) and after CT (T1). Patient Identification Form included a total of eight questions about the demographic characteristics (age, gender, education level, marital status) and health status (cancer type, comorbidities, CT protocol) of the patients. Patient Identification Form included a total of eight questions about the demographic characteristics (age, gender, education level, marital status) and health status (cancer type, comorbidities, CT protocol) of the patients. Patient Follow-up Form was used to record the changes in the patient's condition during the music application, the volume of the music, and the duration of listening to the music. Patient Follow-up Form was used to record the changes in the patient's condition during the music application, the volume of the music, and the duration of listening to the music. Spielberger's STAI was used to assess the level of anxiety. The STAI self-administered scale with a total of 40 four-point Likert items; 20 items assess the state anxiety (STAI-S) (a transient state affected by the current situation in which the respondent reports how he/she currently feels), and 20 items assess the trait anxiety (STAI-T) (a general tendency to worry in which the respondent reports how he/she feels "in general") [34]. The possible scores range from a minimum of 20 points to a maximum of 80 points on both the STAI-T and STAI-S subscales. The scores are categorized as "no or low anxiety" (20–37), "moderate anxiety" (38–44), and "high anxiety" (45–80) [34]. Both STAI-S and STAI-T were assessed at T0; only STAI-S was assessed at T1. Nausea Assessment Form was used to assess the nausea level of the participants at T0 and T1 based on a Visual Analog Scale (VAS) within a range of 0–10 points, where "0" represents no nausea and "10" represents the most severe nausea. Satisfaction Assessment Form was used to assess the satisfaction of patients about the interventions at T1 based on a VAS within a range of 0–10 points, where "0" represents "not at all satisfied" and "10" represents "very satisfied". Turkish classical music group Turkish classical music (TCM), which has a special place in Turkish culture, was utilized for patients in the TCMG. TCM has a musical structure that includes maqam (melodic structure), which involves different chords and microtonal intervals that vary according to geographical region and artist. Most of the melodic aspects of TCM are described by the term maqam [32]. It has been indicated that maqam music, which matured during the supremacy of the Ottoman Empire and whose effects have been studied since the time of Farabi, a Turkish scientist who lived in the 8th century, has been used effectively in the treatment of several conditions and negative symptoms [32, 33]. The Turkish Music Research and Promotion Group (TÜMATA), which was established in Turkey in 1976, has continued its research on the effects of Page 4/16 Page 4/16 different maqams. This study used TCM in the "Rast" maqam, which is one of the four main maqams with high therapeutic effect and included in the "Music and Health Series 2" album specially created by TÜMATA. The "Rast" maqam involves musical instruments such as tambour, bağlama, ney, and oud in the range of 60–66 bpm, which provides feelings of peace, comfort, and happiness to the individual. different maqams. This study used TCM in the "Rast" maqam, which is one of the four main maqams with high therapeutic effect and included in the "Music and Health Series 2" album specially created by TÜMATA. The "Rast" maqam involves musical instruments such as tambour, bağlama, ney, and oud in the range of 60–66 bpm, which provides feelings of peace, comfort, and happiness to the individual. Control group Participants in the control group (CG) received standard care, including pre-CT instructions (e.g., lying down to rest, reporting any discomfort at the infusion site) and premedication treatment, without any music or distraction intervention. Outcome measures Participants in the control group (CG) received standard care, including pre-CT instructions (e.g., lying down to rest, reporting any discomfort at the infusion site) and premedication treatment, without any music or distraction intervention. Outcome measures This form also included open-ended questions about the acceptability of interventions and the expectations during the CT process. Analysis The data obtained from the study were analyzed in SPSS 28 (IBM, Statistical Package for Social Sciences). Before proceeding to the data analysis, the normality of the data was checked with skewness/kurtosis values and the Shapiro- Wilk test. Appropriate descriptive statistics were used to summarize the characteristics of the participants and outcome variables. Continuous variables were expressed as means and standard deviation; categorical variables were expressed as frequencies and percentages. Sociodemographic data and key outcome variables for different groups were compared using the chi-square test or one-way ANOVA test. Mixed model ANOVA with post hoc Bonferroni adjustment was used to determine the effects of time, group, and group × time interactions. Mauchly's test was used to test for the sphericity of the variance-covariance matrix. If the variance-covariance matrix lacked sphericity, the Greenhouse-Geisser technique Page 5/16 Page 5/16 was used to correct it. The effect sizes for the mixed measures ANOVAs were calculated as partial eta squared (η2) and classified as small 0.02–0.13, medium 0.13–0.26, and large > 0.26 [35]. The significance level was set as p < 0.05 (2- tailed). was used to correct it. The effect sizes for the mixed measures ANOVAs were calculated as partial eta squared (η2) and classified as small 0.02–0.13, medium 0.13–0.26, and large > 0.26 [35]. The significance level was set as p < 0.05 (2- tailed). Recruitment, attrition, and adherence The flowchart for the recruitment and group assignment stages of the study is shown in Fig. 1. Initially, 328 patients were assessed for eligibility. Of these, 218 did not meet the inclusion criteria, 19 eligible patients refused to participate, and 15 were excluded for other reasons, such as not being available in time for the start of CT or limited intervention materials (3 on-ear headphones + 3 mp3 players). Thus, the remaining 76 patients were randomly divided into three groups. One participant in the TCMG could not continue the music intervention due to dyspnea and low oxygen saturation at the eighth minute after the initiation of CT. Apart from this, all participants in the intervention groups completed a single session of music intervention for at least 60 minutes (Table 1). As a result, 75 participants (25 in each group) were included in the final analysis. No music-related adverse events occurred during the interventions. Across participants in the intervention groups, the mean volume of self-tuned music was 20.22 ± 1.07 and there was no significant difference in the volume of music between the groups. The average duration of the music across the intervention groups was 70.60 ± 6.88 minutes. The duration of listening to the music was significantly longer in the TCMG than that in the RMG (p = 0.011) (Table 1). Table 1 Descriptions of interventions Adherence variables RMG (n = 25) TCMG (n = 26) p value Number of patients receiving eligible intervention (60 minutes), n (%) 25 (100) 25 (96.2) 0.998 Number of patients discontinued intervention, n (%) 0 (0) 1 (3.8) Music volume, mean ± SD 20.36 ± 1.15 20.08 ± 0.99 0.362 Duration of music (minutes), mean ± SD 68.16 ± 5.38 73.04 ± 6.61 0.011* SD: standart deviation; RMG: relaxing music group; TCMG: Turkish classical music group; *: p < 0.05 Participants' baseline characteristics Participants' baseline characteristics The socio-demographic and clinical characteristics of the patients across the whole sample and related differences between the groups are shown in Table 2. The mean age was 62.18 ± 9.71 years, 52% of the patients were female, 41.3% were high school graduates, and a great majority were married (88%). No significant difference was found between the groups in terms of age, gender, educational status, and marital status (p > 0.05). The most common comorbidities in the patients included in the study were diabetes mellitus (24%) and hypertension (21.3%). More than half of the patients had lung or breast cancer (25.3% for both). The most common CT protocols received by patients were carboplatin (29.3%) and adriamycin + cyclophosphamide (AC) (25.3%). The average baseline STAI-T score across the whole sample was 41.13 ± 6.42, indicating moderate anxiety. There was no significant difference between the groups in terms of comorbidities, cancer type, CT protocol, and baseline STAI-T (p > 0.05). Primary and secondary outcomes Page 6/16 Table 3 shows the change in anxiety and nausea levels in the study groups from the start (T0) to the end (T1) of the CT sessions. A plot of the changes in primary and secondary outcomes from T0 to T1 is presented in Fig. 2. No significant difference was found among the baseline anxiety and nausea levels in the intervention and control groups (p > 0.05). Mixed-design repeated measures ANOVA analyses revealed a significant time effect in the RMG, TCMG, and CG in terms of anxiety and nausea levels (p = 0.001, F = 270.602, η2 = 0.790, and p = 0.001, F = 145.497, η2 = 0.669, respectively). Repeated measures group x time interaction showed significant differences in both anxiety and nausea levels among the groups (p = 0.001, F = 210.221, η2 = 0.745, and p = 0.001, F = 100.785, η2 = 0.583, respectively). In CG, there was no change in the anxiety levels from T0 to T1, whereas there was an increase in nausea. In the RMG and TCMG, the anxiety and nausea levels decreased significantly compared to pre-intervention, but there was no significant difference between the two groups (Fig. 2). Post-intervention satisfaction level was significantly higher in the TCMG compared to the CG and RMG (8.64 ± 0.95 vs. 7.88 ± 0.72 and 7.00 ± 0.70, respectively). Participants' baseline characteristics Forty-two patients (13 in the RMG, 18 in the TCMG, and 11 in the CG) completed open-ended questions about the acceptability of interventions and their expectations after the study was completed. Approximately 77% of the patients in the RMG reported that they enjoyed listening to the music and felt generally peaceful as if they were taking a walk in nature. Approximately 83% of the patients in the TCMG reported that they liked music and felt nostalgic, reminiscent of memories, and peaceful while listening to music. In both groups, some of the participants indicated that it would be good to be given the option to choose their favorite music. Participants did not provide any negative feedback about the on-ear headphones. Participants, including those in the CG, commonly reported their expectations from healthcare professionals during the CT treatment process as meeting their information needs more comprehensively, especially during the first CT application, and offering different nonpharmacological options (e.g. watching movies, providing daily newspapers or books, etc.) to make this process more comfortable. Participants' baseline characteristics Page 7/16 Table 2 Baseline characteristics of participants P 8/16 Baseline characteristics of participants Characteristics Total (n = 75) RMG (n = 25) TCMG (n = 25) CG (n = 25) χ2/F p value Age, mean ± SD 62.18 ± 9.71 61.84 ± 10.30 62.80 ± 9.31 61.92 ± 9.86 0.073 0.929 Age, n (%) 18–40 years 4 (5.3) 2 (8.0) 1 (4.0) 1 (4.0) 0.716 0.949 41–63 years 39 (52.0) 12 (48.0) 13 (52.0) 14 (56.0) 64–85 years 32 (42.7) 11 (44.0) 11 (44.0) 10 (40.0) Sex, n (%) Female 39 (52.0) 13 (52.0) 12 (48.0) 14 (56.0) 0.321 0.852 Male 36 (48.0) 12 (48.0) 13 (52.0) 11 (44.0) Education, n (%) Secondary 31 (41.3) 9 (36.0) 11 (44.0) 11 (44.0) 2.561 0.634 High 31 (41.3) 13 (52.0) 10 (40.0) 8 (32.0) Bachelor 13 (17.4) 3 (12.0) 4 (16.0) 6 (24.0) Marital status, n (%) Married 66 (88.0) 21 (84.0) 21 (84.0) 24 (96.0) 2.273 0.321 Unmarried/Widowed/Divorced/Separated 9 (12.0) 4 (16.0) 4 (16.0) 1 (4.0) Comorbiditiesa, n (%) Diabetes 18 (24.0) 5 (20.0) 6 (24.0) 7 (28.0) 0.439 0.803 Hypertension 16 (21.3) 4 (16.0) 5 (20.0) 7 (28.0) 1.112 0.573 CAD 2 (2.7) 2 (8.0) 0 (0) 0 (0) 0.411 0.128 COPD 3 (4.0) 2 (8.0) 1 (4.0) 0 (0) 2.083 0.353 Kidney failure 2 (2.7) 0 1 (4.0) 1 (4.0) 1.027 0.598 Cancer type, n (%) Lung 19 (25.3) 5 (20.0) 9 (36.0) 5 (20.0) 9.256 0.902 Breast 19 (25.3) 6 (24.0) 7 (28.0) 6 (24.0) Endometrial 9 (12.0) 4 (16.0) 2 (8.0) 3 (12.0) Pancreatic 15 (20.0) 4 (16.0) 5 (20.0) 6 (26.0) Characteristics Total (n = 75) RMG (n = 25) TCMG (n = 25) CG (n = 25) χ2/F p value Skin 2 (2.7) 1 (4.0) 0 (0) 1 (4.0) Sarcoma 2 (2.7) 1 (4.0) 0 (0) 1 (4.0) Ovarian 5 (6.7) 2 (8.0) 1 (4.0) 2 (8.0) Thyroid 1 (1.3) 0 (0) 0 (0) 1 (4.0) Colon 3 (4.0) 2 (8.0) 1 (4.0) 0 (0) Chemotherapy, n (%) AC (Adriamycin + Cyclophosphamide) 19 (25.3) 6 (24.0) 7 (28.0) 6 (24.0) 4.242 0.936 Carboplatin 22 (29.3) 9 (36.0) 7 (28.0) 6 (24.0) FOLFOX (Oxaliplatin + 5-fluorouracil) 18 (24.0) 6 (24.0) 6 (24.0) 6 (24.0) GC (Gemcitabine + Cisplatin) 2 (2.7) 1 (4.0) 0 (0) 1 (4.0) Gemcitabine 2 (2.7) 1 (4.0) 0 (0) 1 (4.0) PC (Paclitaxel + carboplatin) 12 (16.0) 2 (8.0) 5 (20.0) 5 (20.0) STAI-T, mean ± SD 41.13 ± 6.42 41.20 ± 6.21 40.28 ± 5.20 41.92 ± 7.75 0.403 0.670 SD: standart deviation; RMG: relaxing music group; TCMG: Turkish classical music group; CG: control group; CAD: Coronary artery disease; COPD: Chronic obstructive pulmonary disease; STAI-T: State-Trait Anxiety Inventory – Trait; a: some patients had more than one comorbidities. Participants' baseline characteristics SD: standart deviation; RMG: relaxing music group; TCMG: Turkish classical music group; CG: control group; CAD: Coronary artery disease; COPD: Chronic obstructive pulmonary disease; STAI-T: State-Trait Anxiety Inventory – Trait; a: some patients had more than one comorbidities. Page 9/16 Page 9/16 Table 3 Repeated measures for primary and secondary outcomes in the groups at baseline and post-intervention Outcome variables RMG (mean ± SD) TCMG (mean ± SD) CG (mean ± SD) Between Group Comparisons Mixed design Repeated Measures ANOVA test RMG vs. TCMG RMG vs. CG TCMG vs. CG p value p value p value F p value Anxiety T0 38.44 ± 4.27 37.88 ± 4.46 39.52 ± 6.73 0.999 0.998 0.998 270.602 0.790 0.001a T1 28.12 ± 3.34 27.60 ± 3.12 39.84 ± 6.23 0.911 0.001 0.001 74.208 0.333 0.001b 210.221 0.745 0.001c Nausea T0 2.60 ± 1.04 2.56 ± 0.65 2.54 ± 0.64 0.996 0.996 0.998 145.497 0.669 0.001a T1 1.12 ± 0.33 1.08 ± 0.27 4.08 ± 0.78 0.994 0.001 0.001 28.368 100.785 0.173 0.583 0.001b 0.001c Satisfaction T1 7.88 ± 0.72 8.64 ± 0.95 7.00 ± 0.70 0.004 0.001 0.001 SD: standart deviation; RMG: relaxing music group; TCMG: Turkish classical music group; CG: control group; a: time effect; b: group effect; c: groupxtime effect; T0: baseline; T1: post-intervention; values highlighted in bold = p < 0.05 \varvecη2 Table 3 Discussion The findings of this three-arm, randomized controlled trial showed that music interventions reduced anxiety and nausea compared to standard care, regardless of the type of music intervention, in cancer patients receiving CT for the first time. This suggests that the positive neurological effects of music in areas such as cognitive function and emotion may be useful in controlling adverse symptoms of cancer patients. Therefore, the results obtained from our study suggest that the development of music algorithms and the implementation of music therapy protocols during CT sessions, especially starting from the initial CT process, may be authorized for cancer patients. We investigated the calming and anxiolytic effects of music in our study by choosing relaxing and peaceful compositions (60–66 and 60–80 bpm). We aimed to see the effect of both cultural and familiarity characteristics of music by choosing TCM, a more preferred music genre in Turkish culture, as an intervention arm of our study. Culturally familiar music is recommended as part of an effective music intervention since it is more likely to encourage the individual's initial engagement and evoke positive memories [36]. In the other intervention arm of the study, we preferred an RM based on nature sounds that are less familiar compared to TCM but have the ability to overcome language, social and cultural barriers. Although culturally unfamiliar music is perceived as strange by the individual at the initial stage, it is reported that it can affect different regions in the brain and provide stronger cortical activity and neurological Page 10/16 Page 10/16 interaction compared to more familiar music [37, 38]. Adherence to the music sessions of at least 60 minutes in length was high. According to the feedback we received from our patients who completed the music session properly, the acceptability of both music interventions was high. However, some patients also expressed a desire to listen to their preferred music, which suggests that future study designs may take into account the patient's favorite or preferred music genres. The results of the primary and secondary outcomes on anxiety, nausea, and satisfaction level indicate a prospect of benefiting from the music intervention regardless of the genre. The STAI (STAI-T and STAI-S), which assessed the patients' level of anxiety, showed homogeneous moderate anxiety in all groups at baseline. Anxiety was significantly lower in both RMG and TCMG compared to the CG (standard care) after the intervention. Discussion Previous studies reported that music effectively reduces anxiety in cancer patients [23, 24, 39, 40], which is in line with the results of our study. However, although the patients in the TCMG had higher levels of satisfaction and listening time, the TCMG and RMG did not differ significantly in controlling anxiety. This may have been because the sample consisted of patients newly diagnosed with cancer and patients receiving CT for the first time. A cancer diagnosis can be an unexpected life-changing event for patients and their families. Patients may experience various reactions such as shock, denial, confusion, sadness, anger, guilt, and resignation when informed about the diagnosis. In addition to emotional turmoil, patients often have to quickly acquire new information to cope with the new situation, form treatment plans together with healthcare professionals, and understand their care options [41, 42]. This process can further increase patients' anxiety levels. Therefore, patients might be inclined to need to relax and calm down regardless of the type of music. As a matter of fact, in open-ended questions in which patients reported their thoughts about music therapy during the CT process and their expectations from healthcare professionals, they stated that they were highly satisfied with the music, but they needed more information about the disease process and wanted different non-pharmacological options to be offered in addition to music. In addition to physiological stimuli, psychological stimuli, such as anxiety caused by the stress response, may cause some unpleasant GI symptoms such as nausea and vomiting [43]. In light of this theory, we aimed to evaluate the nausea that might be caused by anxiety associated with the new cancer diagnosis and the first CT experience. In the three groups, nausea, as measured by a 10-point VAS, was low at the baseline, and there was no significant difference between the groups. Compared to the baseline, there was a decrease in post-intervention nausea in the RMG and TCMG, whereas there was an increase in the CG. Considering that all patients in the study sample received premedication treatment with strong antiemetics before CT, the fact that patients still reported nausea, even at a low level, suggests that it may be related to anxiety. Therefore, there might have been a linear decrease in the level of nausea associated with the decrease in anxiety in the intervention groups. Discussion However, it should be kept in mind that the majority of patients in this study received CT protocols such as carboplatin and AC, which are known to cause moderate to high levels of nausea [44]. In order to better clarify the effect of music on anxiety-related nausea, it may be recommended that future studies continue music therapy in the subsequent treatment sessions and evaluate patients by increasing the number of music sessions. Although we did not examine the effects of multi-session music on the patients’ response, our findings suggest that at least one hour of music therapy may improve anxiety and nausea outcomes in patients receiving CT for the first time. However, further studies comparing session frequency, music duration, or music preferred by patients are needed to clarify the potential benefits of music. Funding No funds, grants, or other support was received. Acknowledgments The authors thank the patients who agreed to participate in this study and the CT nurses who contributed to the implementation of the study. Limitations This study has limitations that should be considered. The study was conducted in a single center, only with patients receiving CT for the first time. Therefore, it may not be generalized or representative. Further multicenter studies with a Page 11/16 larger sample size are needed to confirm the findings of the current study. The use of self-reported measures to assess changes in patient outcomes might have led to social desirability and recall bias. The study evaluated the short-term effects of listening to music and did not include any follow-up for ongoing CT cycles. Blinding of patients and practitioners was not possible due to the nature of the study. The music in the current study was selected by the researchers and was not intended to provide the patient's preferred music, which has been reported to be more effective in some studies. Future studies may offer participants more music options according to their preferences. Another limitation is the lack of selection criteria or stratification parameters based on other subgroups, such as being more or less anxious, CT protocol, or cancer stage, in case a different confounding effect was present. As in other trials for complementary therapies, adjunctive treatment designs (i.e., administration of complementary therapies with standard drugs) limit the ability to isolate the effects of the complementary intervention from the overlapping effects of pharmacological intervention. Conclusion The results of this study highlighted that relaxing and traditional music can reduce anxiety and nausea in patients receiving CT for the first time. It was also found that the patient's satisfaction level increased after listening to music, especially when symptoms were reduced with music originating from their cultural background. Music is safe, inexpensive, and effective, thus, providing a simple, complementary, and alternative therapy. Nurses and other healthcare professionals can easily utilize music to control adverse effects in cancer patients, starting with the first course of CT. Future study designs should consider including larger samples, using patient-preferred music, extending the time frame for data collection including the subsequent CT sessions, and conducting follow-ups to determine any long-term benefit of music interventions. Data Availability The data that support the findings of this study are available from the corresponding The data that support the findings of this study are available from the corresponding author, upon reasonable request. Ethics approval and consent to participate Authorship Contributions All authors contributed to the study conception and design. Material preparation, intervention and data collection were performed by ÖED and DAD. The full manuscript was written by ÖED, DAD, SP and YY and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Conflicts of interest The authors declare that they have no conflicts of interest. Ethics approval and consent to participate The study was conducted according to the Declaration of Helsinki. Ethics committee approval was obtained from the Clinical Research Ethics Committee at xxx University School of Medicine (Decision number: 2022-13/18); other required permissions were granted by the institution where the study was conducted. Before baseline measurements and group Page 12/16 assignments were done, each participant was informed about the study and written informed consent for participation was obtained. Participation was voluntary. Participants had the right to leave the study at any time without affecting their treatment or services in any way. Patients' privacy was protected by ensuring anonymity and withholding personal information throughout the research process. The study protocol was registered on ClinicalTrails.gov (Registration number: NCT05687838). References 1. Siegel RL, Miller KD, Wagle NS, Jemal A (2023). Cancer statistics, 2023. CA Cancer J Clin 73(1):17–48. https://doi.org/10.3322/caac.21763 1. Siegel RL, Miller KD, Wagle NS, Jemal A (2023). Cancer statistics, 2023. CA Cancer J Clin 73(1):17–48. https://doi.org/10.3322/caac.21763 2. Kutluk T, Ahmed F, Cemaloğlu M et al (2021). Progress in palliative care for cancer in Turkey: a review of the literature. Ecancermedicalscience 15:1321. https://doi.org/10.3332/ecancer.2021.1321 2. Kutluk T, Ahmed F, Cemaloğlu M et al (2021). Progress in palliative care for cancer in Turkey: a review of the literature. Ecancermedicalscience 15:1321. https://doi.org/10.3332/ecancer.2021.1321 3. Anand U, Dey A, Chandel AKS,et al (2022). Cancer chemotherapy and beyond: Current status, drug candidates, associated risks and progress in targeted therapeutics. Genes Dis 10(4):1367–1401. https://doi.org/10.1016/j.gendis.2022.02.007 3. Anand U, Dey A, Chandel AKS,et al (2022). Cancer chemotherapy and beyond: Current status, drug candidates, associated risks and progress in targeted therapeutics. Genes Dis 10(4):1367–1401. https://doi.org/10.1016/j.gendis.2022.02.007 4. Roe H, Lennan E (2014). Role of nurses in the assessment and management of chemotherapy-related side effects in cancer patients. Nursing: Research and Reviews 4, 103–115. https://doi.org/10.2147/NRR.S41845 4. Roe H, Lennan E (2014). Role of nurses in the assessment and management of chemotherapy-related side effects in cancer patients. Nursing: Research and Reviews 4, 103–115. https://doi.org/10.2147/NRR.S41845 5. Debela DT, Muzazu SG, Heraro KD et al (2021). New approaches and procedures for cancer treatment: Current perspectives. SAGE Open Med 9:20503121211034366. https://doi.org/10.1177/20503121211034366 5. Debela DT, Muzazu SG, Heraro KD et al (2021). New approaches and procedures for cancer treatment: Current perspectives. SAGE Open Med 9:20503121211034366. https://doi.org/10.1177/20503121211034366 6. Devlin EJ, Denson LA, Whitford HS (2017). Cancer Treatment Side Effects: A Meta-analysis of the Relationship Between Response Expectancies and Experience. J Pain Symptom Manage 54(2):245–258.e2. https://doi.org/10.1016/j.jpainsymman.2017.03.017 6. Devlin EJ, Denson LA, Whitford HS (2017). Cancer Treatment Side Effects: A Meta-analysis of the Relationship Between Response Expectancies and Experience. J Pain Symptom Manage 54(2):245–258.e2. https://doi.org/10.1016/j.jpainsymman.2017.03.017 7. Parás-Bravo P, Paz-Zulueta M, Boixadera-Planas E et al (2020). Cancer Patients and Anxiety: A Gender Perspective. Int J Environ Res Public Health 17(4):1302. https://doi.org/10.3390/ijerph17041302 7. Parás-Bravo P, Paz-Zulueta M, Boixadera-Planas E et al (2020). Cancer Patients and Anxiety: A Gender Perspective. Int J Environ Res Public Health 17(4):1302. https://doi.org/10.3390/ijerph17041302 8. Naser AY, Hameed AN, Mustafa N et al (2021). Depression and Anxiety in Patients With Cancer: A Cross-Sectional Study. Front Psychol 12:585534. https://doi.org/10.3389/fpsyg.2021.585534 8. Naser AY, Hameed AN, Mustafa N et al (2021). Depression and Anxiety in Patients With Cancer: A Cross-Sectional Study. References Front Psychol 12:585534. https://doi.org/10.3389/fpsyg.2021.585534 9. Charalambous A, Giannakopoulou M, Bozas E, Paikousis L (2015). A Randomized Controlled Trial for the Effectiveness of Progressive Muscle Relaxation and Guided Imagery as Anxiety Reducing Interventions in Breast and Prostate Cancer Patients Undergoing Chemotherapy. Evid Based Complement Alternat Med 2015:270876. https://doi.org/10.1155/2015/270876 10. Bakhiet TE, Ali SM, Bakhiet AM (2021). Prevalence of depression and anxiety among adult patients undergoing chemotherapy in Khartoum, Sudan: A cross-sectional study. Journal of Affective Disorders Reports 6:100218. https://doi.org/10.1016/j.jadr.2021.100218 11. Pandey M, Sarita GP, Devi N, Thomas BC, Hussain BM, Krishnan R (2006). Distress, anxiety, and depression in cancer patients undergoing chemotherapy. World J Surg Oncol 4:68. 26. https://doi.org/10.1186/1477-7819-4-68 cancer patients undergoing chemotherapy. World J Surg Oncol 4:68. 26. https://doi.org/10.1186/1477-7819-4-68 12. Baqutayan SM (2012). The effect of anxiety on breast cancer patients. Indian J Psychol Med 34(2):119–123. https://doi.org/10.4103/0253-7176.101774 12. Baqutayan SM (2012). The effect of anxiety on breast cancer patients. Indian J Psychol Med 34(2):119–123. https://doi.org/10.4103/0253-7176.101774 Page 13/16 Page 13/16 13. El Kheir DYM, Ibrahim AHM (2019). Epidemiological assessment of distress during chemotherapy: who is affected?. J Taibah Univ Med Sci 14(5):448–453. https://doi.org/10.1016/j.jtumed.2019.08.004 14. Dadkhah B, Anisi E, Mozaffari N, Amani F, Pourghasemian M (2019). Effect of Music Therapy with Periorbital Massage on Chemotherapy-Induced Nausea and Vomiting In Gastrointestinal Cancer: A Randomized Controlled Trail. J Caring Sci 8(3):165–171. https://doi.org/10.15171/jcs.2019.02415 15. Lin MF, Hsieh YJ, Hsu YY, Fetzer S, Hsu MC (2011). A randomised controlled trial of the effect of music therapy and verbal relaxation on chemotherapy-induced anxiety. J Clin Nurs 20(7–8):988–999. https://doi.org/10.1111/j.1365- 2702.2010.03525.x 16. Vichaya EG, Chiu GS, Krukowski K et al. (2015). Mechanisms of chemotherapy-induced behavioral toxicities. Front Neurosci 9:131. https://doi.org/10.3389/fnins.2015.00131 17. Sakamoto R, Koyama A (2020). Effective Therapy Against Severe Anxiety Caused by Cancer: A Case Report and Review of the Literature. Cureus 12(6):e8414. https://doi.org/10.7759/cureus.8414 18. Traeger L, Greer JA, Fernandez-Robles C, Temel JS, Pirl WF (2012). Evidence-based treatment of anxiety in patients with cancer. J Clin Oncol 30(11):1197–1205. https://doi.org/10.1200/JCO.2011.39.5632 19. Forbes E, Baker AL, Britton B et al (2020). Non-pharmacological approaches to procedural anxiety reduction for patients undergoing radiotherapy for cancer: systematic review protocol. BMJ Open 10(10):e035155. https://doi.org/10.1136/bmjopen-2019-035155 20. Bradt J, Dileo C, Myers-Coffman K, Biondo J (2021). Music interventions for improving psychological and physical outcomes in people with cancer. Cochrane Database Syst Rev 10(10):CD006911. https://doi.org/10.1002/14651858.CD006911.pub4 20. Bradt J, Dileo C, Myers-Coffman K, Biondo J (2021). Music interventions for improving psychological and physical outcomes in people with cancer. References Cochrane Database Syst Rev 10(10):CD006911. https://doi.org/10.1002/14651858.CD006911.pub4 21. Gupta A, Bhushan B, Behera L (2018).. Short-term enhancement of cognitive functions and music: A three-channel model. Sci Rep 8(1):15528. https://doi.org/10.1038/s41598-018-33618-1 21. Gupta A, Bhushan B, Behera L (2018).. Short-term enhancement of cognitive functions and music: A three-channel model. Sci Rep 8(1):15528. https://doi.org/10.1038/s41598-018-33618-1 22. Wilkins RW, Hodges DA, Laurienti PJ, Steen M, Burdette JH (2014). Network science and the effects of music preference on functional brain connectivity: from Beethoven to Eminem. Sci Rep 4:6130. https://doi.org/10.1038/srep06130 22. Wilkins RW, Hodges DA, Laurienti PJ, Steen M, Burdette JH (2014). Network science and the effects of music preference on functional brain connectivity: from Beethoven to Eminem. Sci Rep 4:6130. https://doi.org/10.1038/srep06130 23. Tang H, Chen L, Wang Y, Zhang Y, Yang N, Yang N (2021). The efficacy of music therapy to relieve pain, anxiety, and promote sleep quality, in patients with small cell lung cancer receiving platinum-based chemotherapy. Support Care Cancer 29(12):7299–7306. https://doi.org/10.1007/s00520-021-06152-6 23. Tang H, Chen L, Wang Y, Zhang Y, Yang N, Yang N (2021). The efficacy of music therapy to relieve pain, anxiety, and promote sleep quality, in patients with small cell lung cancer receiving platinum-based chemotherapy. Support Care Cancer 29(12):7299–7306. https://doi.org/10.1007/s00520-021-06152-6 24. Nguyen KT, Xiao J, Chan DNS, Zhang M, Chan CWH (2022). Effects of music intervention on anxiety, depression, and quality of life of cancer patients receiving chemotherapy: a systematic review and meta-analysis. Support Care Cancer 30(7):5615–5626. https://doi.org/10.1007/s00520-022-06881-2 24. Nguyen KT, Xiao J, Chan DNS, Zhang M, Chan CWH (2022). Effects of music intervention on anxiety, depression, and quality of life of cancer patients receiving chemotherapy: a systematic review and meta-analysis. Support Care Cancer 30(7):5615–5626. https://doi.org/10.1007/s00520-022-06881-2 25. Al-Jubouri MBA, Isam SR, Hussein SM, Machuca-Contreras F (2021). Recitation of quran and music to reduce chemotherapy-induced anxiety among adult patients with cancer: A clinical trial. Nurs Open 8(4):1606–1614. https://doi.org/10.1002/nop2.781 25. Al-Jubouri MBA, Isam SR, Hussein SM, Machuca-Contreras F (2021). Recitation of quran and music to reduce chemotherapy-induced anxiety among adult patients with cancer: A clinical trial. Nurs Open 8(4):1606–1614. https://doi.org/10.1002/nop2.781 26. Singh A, Soni A, Saluja S (2013). Essentials of sample size calculation. Curr Med Res Pract 3:97–99. Available from: http://www.cmrpjournal.org/text.asp?2013/3/2/97/300688 26. Singh A, Soni A, Saluja S (2013). Essentials of sample size calculation. Curr Med Res Pract 3:97–99. Available from: http://www.cmrpjournal.org/text.asp?2013/3/2/97/300688 27. Vieillard S, Gilet AL (2013). Age-related differences in affective responses to and memory for emotions conveyed by music: a cross-sectional study. References Front Psychol 4:711. https://doi.org/10.3389/fpsyg.2013.00711 27. Vieillard S, Gilet AL (2013). Age-related differences in affective responses to and memory for emotions conveyed by music: a cross-sectional study. Front Psychol 4:711. https://doi.org/10.3389/fpsyg.2013.00711 28. Feizpour A, Parkington HC, Mansouri FA (2020). Cognitive sex differences in effects of music in Wisconsin Card Sorting Test. Psychology of Music 48(2):252–265. https://doi.org/10.1177/0305735618795030 28. Feizpour A, Parkington HC, Mansouri FA (2020). Cognitive sex differences in effects of music in Wisconsin Card Sorting Test. Psychology of Music 48(2):252–265. https://doi.org/10.1177/0305735618795030 29. Garland T (2014). Self-Regulation Interventions and Strategies: Keeping the Body, Mind & Emotions on Task in Children with Autism, ADHD or Sensory Disorders. Pesi Publishing & Media, USA. 29. Garland T (2014). Self-Regulation Interventions and Strategies: Keeping the Body, Mind & Emotions on Task in Children with Autism, ADHD or Sensory Disorders. Pesi Publishing & Media, USA. Page 14/16 30. O'Kelly J, James L, Palaniappan R, Taborin J, Fachner J, Magee WL (2013). Neurophysiological and behavioral responses to music therapy in vegetative and minimally conscious States. Front Hum Neurosci 7:884. https://doi.org/10.3389/fnhum.2013.00884 31. MusiCure (2005). Evidence Based Music and Films by composer Niels Eje. https://musicure.dk/en/home Accessed 30 August 2022. 32. Bozkurt B, Ayangil R, Holzapfel A (2014). Computational Analysis of Turkish Makam Music: Review of State-of-the- Art and Challenges. Journal of New Music Research 43(1):3–23. https://doi.org/10.1080/09298215.2013.865760 33. Yıldırım M (2021). Music Healing: History, Development, Application in Addictions And Music Therapy Applications in Turkey. Turkish Academic Research Review 6(2):477–497. https://doi.org/10.30622/tarr.935092 34. Spielberger CD (1983). Manual for the State-Trait Inventory STAI (Form Y). Palo Alto, CA: Mind Garden. 34. Spielberger CD (1983). Manual for the State-Trait Inventory STAI (Form Y). Palo Alto 35. Bakeman R (2005). Recommended effect size statistics for repeated measures designs. Behav Res Methods 37(3):379–84. https://doi.org/10.3758/BF03192707. 35. Bakeman R (2005). Recommended effect size statistics for repeated measures designs. Behav Res Methods 37(3):379–84. https://doi.org/10.3758/BF03192707. 36. Jagiello R, Pomper U, Yoneya M et al (2019). Rapid Brain Responses to Familiar vs. Unfamiliar Music – an EEG and Pupillometry study. Sci Rep 9, 15570. https://doi.org/10.1038/s41598-019-51759-9 36. Jagiello R, Pomper U, Yoneya M et al (2019). Rapid Brain Responses to Familiar vs. Unfamiliar Music – an EEG and Pupillometry study. Sci Rep 9, 15570. https://doi.org/10.1038/s41598-019-51759-9 37. Meltzer B, Reichenbach CS, Braiman C, Schiff ND, Hudspeth AJ, Reichenbach T (2015). The steady-state response of the cerebral cortex to the beat of music reflects both the comprehension of music and attention. Front Hum Neurosci 9:436. References https://doi.org/10.3389/fnhum.2015.00436 37. Meltzer B, Reichenbach CS, Braiman C, Schiff ND, Hudspeth AJ, Reichenbach T (2015). The steady-state response of the cerebral cortex to the beat of music reflects both the comprehension of music and attention. Front Hum Neurosci 9:436. https://doi.org/10.3389/fnhum.2015.00436 38. Kumagai Y, Arvaneh M, Tanaka T (2017). Familiarity Affects Entrainment of EEG in Music Listening. Front Hum Neurosci 11:384. https://doi.org/10.3389/fnhum.2017.00384 39. Bradt J, Dileo C, Myers-Coffman K, Biondo J (2021). Music interventions for improving psychological and physical outcomes in people with cancer. Cochrane Database Syst Rev 10(10):CD006911. https://doi.org/10.1002/14651858.CD006911.pub4 40. Bro ML, Johansen C, Vuust P et al (2019). Effects of live music during chemotherapy in lymphoma patients: a randomized, controlled, multi-center trial. Support Care Cancer 27(10):3887–3896. https://doi.org/10.1007/s00520- 019-04666-8 41. Coronado AC, Tran K, Chadder J et al (2017). The experience of patients with cancer during diagnosis and treatment planning: a descriptive study of Canadian survey results. Curr Oncol 24(5):332–337. https://doi.org/10.3747/co.24.3782 42. Derry HM, Epstein AS, Lichtenthal WG, Prigerson HG (2019). Emotions in the room: common emotional reactions to discussions of poor prognosis and tools to address them. Expert Rev Anticancer Ther 19(8):689–696. https://doi.org/10.1080/14737140.2019.1651648 43. Pasyar N, Rambod M, Zahedi F, Ramzi M (2022). Pain, Fatigue, Nausea, and Vomiting as the Predictors of Anxiety in Patients Undergoing Hematopoietic Stem Cell Transplantation: A Prospective Cohort Study. Support Care Cancer 30(7):5871–5879. https://doi.org/10.1007/s00520-022-06997-5 44. Roila F, Molassiotis A, Herrstedt J et al (2016). 2016 MASCC and ESMO guideline update for the prevention of chemotherapy- and radiotherapy-induced nausea and vomiting and of nausea and vomiting in advanced cancer patients. Ann Oncol 27(suppl 5):v119-v133. https://doi.org/10.1093/annonc/mdw270 Figures Figures Page 15/16 Figure 1 The flowchart for the participant recruitment and group allocation procedures Figure 1 The flowchart for the participant recruitment and group allocation procedures Figure 2 The changes in levels of anxiety and nausea in overall time points according to the groups Figure 1 Figure 1 The flowchart for the participant recruitment and group allocation procedures Figure 2 The changes in levels of anxiety and nausea in overall time points according to the groups Figure 2 Figure 2 The changes in levels of anxiety and nausea in overall time points according to the groups Page 16/16 Page 16/16
https://openalex.org/W2606991277	https://mdsoar.org/bitstream/11603/12951/1/Rushing_et_al-2017-Ecology_and_Evolution.pdf	English	null	Incorporating breeding abundance into spatial assignments on continuous surfaces	Ecology and evolution	2,017	cc-by	8,901	Received: 22 July 2016 \| Revised: 11 October 2016 \| Accepted: 19 October 2016 Received: 22 July 2016 \| Revised: 11 October 2016 \| Accepted: 19 October 2016 Received: 22 July 2016 \| Revised: 11 October 2016 \| Accepted: 19 October 2016 DOI: 10.1002/ece3.2605 DOI: 10.1002/ece3.2605 Correspondence Correspondence Clark S. Rushing, Migratory Bird Center, Smithsonian Conservation Biology Institute, National Zoological Park, Washington DC, USA. Email: rushingc@si.edu Incorporating breeding abundance into spatial assignments on continuous surfaces Clark S. Rushing1 \| Peter P. Marra1 \| Colin E. Studds1,2 1Migratory Bird Center, Smithsonian Conservation Biology Institute, National Zoological Park, Washington, DC, USA 1Migratory Bird Center, Smithsonian Conservation Biology Institute, National Zoological Park, Washington, DC, USA Abstract Determining the geographic connections between breeding and nonbreeding popu- lations, termed migratory connectivity, is critical to advancing our understanding of the ecology and conservation of migratory species. Assignment models based on stable isotopes historically have been an important tool for studying migratory con- nectivity of small-bodied species, but the low resolution of these assignments has generated interest into combining isotopes with other sources in information. Abundance is one of the most appealing data sources to include in isotope-based assignments, but there are currently no statistical methods or guidelines for opti- mizing the contribution of stable isotopes and abundance for inferring migratory connectivity. Using known-origin stable-hydrogen isotope samples of six Neotropical migratory bird species, we rigorously assessed the performance of as- signment models that differentially weight the contribution of the isotope and abundance data. For two species with adequate sample sizes, we used Pareto opti- mality to determine the set of models that simultaneously minimized both assign- ment error rate and assignment area. We then assessed the ability of the top models from these two species to improve assignments of the remaining four species com- pared to assignments based on isotopes alone. We show that the increased preci- sion of models that include abundance is often offset by a large increase in assignment error. However, models that optimally weigh the abundance data rela- tive to the isotope data can result in higher precision and, in some cases, lower error than models based on isotopes alone. The top models, however, depended on the distribution of relative breeding abundance, with patchier distributions requiring stronger downweighting of abundance, and we present general guidelines for fu- ture studies. These results confirm that breeding abundance can be an important source of information for studies investigating broad-scale movements of migratory birds and potentially other taxa. Ecology and Evolution. 2017;7:3847–3855. \| 3847 www.ecolevol.org This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2017 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd. Abstract 2Department of Geography & Environmental Systems, University of Maryland Baltimore County, Baltimore, MD, USA 2Department of Geography & Environmental Systems, University of Maryland Baltimore County, Baltimore, MD, USA 1 \| INTRODUCTION 2013), morphometrics (Rushing, Ryder, Saracco, & Marra, 2014), and band recoveries (Hobson, Wunder, Van Wilgenburg, Clark, & Wassenaar, 2009; Van Wilgenburg & Hobson, 2011), or with information on abundance (Flockhart et al., 2013; Hallworth, Studds, Sillett, & Marra, 2013). In this study, we describe a new quantitative method for making geographic assignments to origin that differentially weights stable- hydrogen isotope and breeding abundance data to maximize assign- ment area and minimize assignment error. We performed Bayesian assignments of origin for six species of Neotropical-Nearctic migratory birds: Wood Thrush (Hylocichla mustelina), Northern Parula (Setophaga americana), Prairie Warbler (Setophaga discolor), Black-and-White Warbler (Mniotilta varia), American Redstart (Setophaga ruticilla), and Ovenbird (Seiurus aurocapilla). Using stable-hydrogen isotope data col- lected at known breeding sites across the breeding range of each spe- cies enabled us to assess the performance of different models. Using Pareto optimality, a method for multi-objective optimization, we show that weighting the isotope and abundance data can increase the per- formance of assignment models but that the distribution of breeding abundance plays a critical role in determining the proper weightings. Abundance is one of the most appealing data sources to include in migratory connectivity analyses because high-quality, range-wide data are often freely available. However, a recently noted problem with combing stable isotopes and abundance in migratory connectivity es- timates is that many species have patchy breeding distributions. This causes centers of high abundance to be overrepresented and areas of low abundance to be underrepresented in assignments to origin (Hobson et al., 2014). Despite their wide use, there is no analytical method and no guidelines for optimizing the contribution of stable isotopes and abundance for inferring migratory connectivity. The first formal method to include breeding abundance and in- trinsic markers in geographic assignments was outlined by Royle and Rubenstein (2004). In their model, stable isotopes are used to assign individuals to one of a finite number of discrete breeding populations, denoted by b = 1, 2, … B. Each population is defined by a probabil- ity distribution that describes the expected values of the marker for individuals originating in that population. From this distribution, it is straightforward to estimate the likelihood that each population is the origin of an individual with an observed marker value y, denoted f(y\|b). 1 \| INTRODUCTION isotope data (via the likelihood) and breeding abundance (via the prior). Although the original model outlined by Royle and Rubenstein (2004) was developed to make geographic assignments to a few dis- crete breeding regions, the increasing availability of global isoscape and abundance surfaces has enabled researchers to make assignments on nearly continuous landscapes (Hobson et al., 2009; Sullivan et al., 2012; Van Wilgenburg & Hobson, 2011). This introduces a critical complication that has gone largely unrecognized. The problem arises because most species are patchily distributed across their range, with a few areas of relatively high abundance and extensive areas of lower abundance. When abundance is included as a prior in assignments to continuous surfaces, the low abundance sites may receive low poste- rior support compared to high abundance locations. Although this out- come is consistent with the logic proposed by Royle and Rubenstein (2004), in practice, the posterior probabilities for each location may simply reflect relative abundance, thus limiting the contribution of the isotope data to assignments (see González-Prieto, Hobson, Bayly, & Gómez, 2011; Hobson et al., 2014). In extreme cases, the inclusion of breeding abundance may lead to inaccurate assignments and ob- scure estimates of migratory connectivity, the biological processes of interest. Understanding how migratory species redistribute themselves across the annual cycle, known as migratory connectivity, is essential for un- derstanding range dynamics, identifying key threats, and developing coordinated conservation actions. Satellite tracking has revolutionized migratory connectivity research for large-bodied species (>100 g) by enabling remote transmission of individual movements over broad spatial and temporal scales (Block et al., 2011). For species too small to carry these transmitters, morphological and chemical signatures of individual organisms or tissues, termed intrinsic markers, are essential tools for estimating migratory connectivity. Stable isotopes are arguably the most useful intrinsic marker for studying migratory connectivity because of their comparatively low cost and scale of inference (Hobson, 2008) and have supported im- portant advances for numerous taxa, including birds (Rubenstein & Hobson, 2004), mammals (Sullivan, Bump, Kruger, & Peterson, 2012), and insects (Hobson, Soto, Paulson, Wassenaar, & Matthews, 2012; Hobson, Van Wilgenburg, Wassenaar, & Larson, 2012). Although sta- ble isotopes offer lower spatial resolution compared to direct tracking, substantial progress has occurred by combining stable isotopes with other intrinsic markers such as genetic data (Kelly, Ruegg, & Smith, 2005; Rundel et al. 1 \| INTRODUCTION Because areas differ with regard to relative abundance, it may be reasonable to assume that individuals are more likely to originate from high abundance populations than low abundance populations. Using Bayes rule, the relative abundance can be incorporated into the assignment model as a prior probability. This allows researchers to ex- plicitly model the probability that an individual originated from each RUSHING et al. population, that is, f(b\|y), and to formally base assignments on both the isotope data (via the likelihood) and breeding abundance (via the prior). K E Y W O R D S Bayes rule, migratory connectivity, Neotropical migratory birds, Pareto optimality, probabilistic assignment, stable-hydrogen isotopes \| 3847 www.ecolevol.org 3848 2.1 \| Study species and feather sampling To determine whether the inclusion of abundance data improved the performance of assignment models that rely on isotope data alone, we used stable-hydrogen isotope data from six species of Neotropical migratory birds collected at known breeding locations (Tables 1 and S1–S6). The use of known-origin samples allowed us to test explicitly the performance of alternative assignment models. Feather samples were collected from 2009 to 2011 at six breeding sites in the eastern United States that span the geographic extent of the breeding range and include a wide range of breeding abundance for each species. Vegetation types included bottomland hardwoods, coastal plain for- est, northern hardwoods, and spruce-fir forest at elevations from 5 to 3849 RUSHING et al. Powder standard and duplicate samples run for one in five to eight feathers indicated that analytical error (±1 SD) was <2‰. TABLE 1 Summary of sampling data. Abundance range indicates the minimum and maximum predicted breeding abundance at the sampling locations for each species. For each species, the precipitation-based hydrogen isoscape from Bowen et al. (2005) was converted to expected feather values using the slope parameter for long-distance migrants from Van Wilgenburg, et al., 2012, with either the intercepts for ground or nonground foragers Due to the comparatively small number of samples for Ovenbird, Northern Parula, Black-and-White Warbler, and Prairie Warbler, we restricted our full analysis of assignment models to Wood Thrush and American Redstart. After determining the top assignment models for those two species, we used the remaining four species for indepen- dent validations. Species n Abundance range (Maximum)a Foraging height Wood Thrush 120 3.26–17.89 (33.43) Ground American Redstart 110 0.05–7.05 (40.4) Nonground Ovenbird 30 0.02–22.22 (57.92) Ground Northern Parula 27 0.85–10.73 (25.22) Nonground Black-and-White Warbler 20 0.79–3.57 (12.51) Nonground Prairie Warbler 27 0.09–2.21 (23.35) Nonground aAbundance is expressed as the predicted number of birds per BBS route estimated through inverse distancing (Sauer et al., 2015). 2.2 To assign individuals to potential breeding locations, we first created base maps describing the variation in hydrogen isotope abundance and relative abundance across the breeding range of each species. To estimate the hydrogen isoscape, we converted a map of expected amount-weighted growing-season precipitation δ2H values (δ2Hp; Bowen et al. 2005) to expected feather δ2H values (δ2Hf) using pub- lished corrections for either ground foraging or nonground foraging long-distance migratory birds (Hobson, Van Wilgenburg, et al., 2012). In their analysis, Hobson, Van Wilgenburg, et al., 2012 found no sup- port for age-based differences in hydrogen isotope discrimination, and therefore, we did not apply any age-specific correction to the δ2H values. aAbundance is expressed as the predicted number of birds per BBS route estimated through inverse distancing (Sauer et al., 2015). 1,000 m. Birds were captured using mist nets, aged and sexed using cri- teria from Pyle (1997), banded with a United States Geological Survey (USGS) aluminum leg band, and released. One tail feather was removed from each bird before release and stored in a paper envelope. Stable- hydrogen isotope values in feathers grown on or near breeding sites are strongly correlated with stable-hydrogen isotope values in growing- season precipitation and therefore provide information about breeding origin (Hobson, & Wassenaar, 1997). Each of the six species molts their tail feathers following reproduction, usually from late August to early September. We therefore restricted our analyses to adult tail feathers collected between 1 June and 31 July. We did not analyze feathers from immature birds because their isotope values can reflect natal origins. Next, we assigned each bird to potential breeding locations using isotope values only. To do this, we calculated the likelihood that each raster cell represented the breeding location for each individual using a normal probability density function: (1) f(y∗μi,σ)= 1 √ 2πσ exp −1 2σ2 (y∗−μi)2 (1) where f(y\|μi,σ) is the likelihood that an individual with δ2Hf =y∗ originated from cell i, μi is the predicted δ2Hf value for cell i, and σ is the standard deviation of δ2Hf values within a single breeding site, which was assumed to be 12 ‰ (Rushing et al., 2014). Next, we con- verted the likelihood values to a probability surface by dividing each likelihood by the sum of all of the likelihoods (Hobson et al., 2009; Van Wilgenburg & Hobson, 2011). 2.2 We then sorted this probability surface from minimum to maximum and used a smoothing spline function to estimate the probability value (i.e., cutoff) that separated the upper 67% of the cumulative probabilities from the lower 33% (Chabot, Hobson, Van Wilgenburg, McQuat, & Lougheed, 2012; Hobson et al., 2009). Finally, we reclassified any cell with probability greater than the cutoff value as a likely (1) breeding origin and any cell with probabil- ity less than the cutoff as an unlikely (0) origin (Chabot et al., 2012; Hobson et al., 2009; Rushing et al., 2014). For each species and each sampling location, we then estimated the proportion of individuals that were misclassified (i.e., true breeding origin classified as unlikely; hereafter referred to as the error rate) and the mean proportion of raster cells classified as likely (referred to as the assignment area). Because the goal of assignments is to correctly classify the breeding locations while minimizing the assignment area, these two metrics provide quantitative and intuitive measures of model performance. where f(y\|μi,σ) is the likelihood that an individual with δ2Hf =y∗ originated from cell i, μi is the predicted δ2Hf value for cell i, and σ is the standard deviation of δ2Hf values within a single breeding site, which was assumed to be 12 ‰ (Rushing et al., 2014). Next, we con- verted the likelihood values to a probability surface by dividing each likelihood by the sum of all of the likelihoods (Hobson et al., 2009; Van Wilgenburg & Hobson, 2011). We then sorted this probability surface from minimum to maximum and used a smoothing spline function to estimate the probability value (i.e., cutoff) that separated the upper 67% of the cumulative probabilities from the lower 33% (Chabot, Hobson, Van Wilgenburg, McQuat, & Lougheed, 2012; Hobson et al., 2009). Finally, we reclassified any cell with probability greater than the cutoff value as a likely (1) breeding origin and any cell with probabil- ity less than the cutoff as an unlikely (0) origin (Chabot et al., 2012; Hobson et al., 2009; Rushing et al., 2014). For each species and each sampling location, we then estimated the proportion of individuals that were misclassified (i.e., true breeding origin classified as unlikely; hereafter referred to as the error rate) and the mean proportion of raster cells classified as likely (referred to as the assignment area). 2.3 Initially, we incorporated breeding abundance into the assignment model following the method outlined in recent assignment studies (Hallworth et al., 2013; Hobson et al., 2014). We used data from the North American Breeding Bird Survey (BBS) to create base maps of breeding abundance for each species (Figures 1a, 2a and S1; Sauer et al., 2015). Raw abundance estimates were then converted into a probability surface by dividing each cell by the sum of all cells (Hallworth et al., 2013). These relative abundance estimates were in- corporated into the assignment model using Bayes rule: where f(bi\|y) is the posterior probability that an individual with δ2Hf = y* originated from cell i, f(y*\|bi) is the likelihood of assignment to breeding cell i, and f(bi) is the relative abundance (i.e., the prior prob- ability) of cell i. Because the abundance data in equation 2 were un- weighted (see below), we refer to this model as the “naive-abundance” model. As before, the posterior probabilities were converted to likely and unlikely origins using a 67% odds, and we quantified assignment performance using the assignment error rate and assignment area metrics described above. 2.2 Because the goal of assignments is to correctly classify the breeding locations while minimizing the assignment area, these two metrics provide quantitative and intuitive measures of model performance. Isotope analyses were performed at the Stable Isotope Mass Spectrometry Facility of the Smithsonian Institution in Suitland, MD. Feathers were washed in a 2:1 chloroform:methanol solution to re- move surface oils and air-dried under a fume hood for 48 hr. After transport to the laboratory, feathers were allowed to equilibrate with the local atmosphere for 72 hr. A small sample of each feather (0.30– 0.35 mg) was packed into a silver capsule, combusted at 1,350°C in an elemental analyzer (Thermo TC/EA), and introduced online to an iso- tope ratio mass spectrometer (Thermo Delta V Advantage) via a Conflo IV interface. Four previously calibrated keratin standards were run for every 10 unknowns to account for exchangeable and nonexchangeable H in feather samples (IAEA-CH-7: δ2H = −100.3‰ Vienna standard mean ocean water [VSMOW]; Caribou Hoof Standard: δ2H = −197‰ VSMOW; Kudu Horn Standard: δ2H = −54.1 ‰ VSMOW; Spectrum Keratin Fine Powder: δ2H = −121.6 ‰ VSMOW). The δ2H values reported include only nonexchangeable H as determined by linear regression with the IAEA-CH-7, CHS, and KHS keratin standards (Wassenaar & Hobson, 2003) and are expressed in per mil units (‰) relative to the VSMOW-Standard Light Antarctic Precipitation (VSMOW-SLAP) scale. Replicate samples of the Spectrum Keratin Fine 3850 RUSHING et al. 2.4 \| Data weighting and model comparison (2) f(biy∗)= f(y∗bi)f(bi) ∑B i=1 f(y∗bi)f(bi) (2) f(biy∗)= f(y∗bi)f(bi) ∑B i=1 f(y∗bi)f(bi) As described above, the combination of isotope data and abun- dance data described by equation 2 may be problematic if the prior (2) 0.00 0.25 0.50 0.75 1.00 Rel. abun. Relative abundance (a) Origin Unlikely Likely Isotope-only model (b) Relative abundance (a) Relative abundance (a) Isotope-only model (b) Origin Unlikely Likely Isotope-only model (b) Origin Unlikely Likely Naive-abundance model Origin Unlikely Likely Top abundance model (c) (d) Origin Unlikely Likely Naive-abundance model (c) Origin Unlikely Likely Naive-abundance model Origin Unlikely Likely Top abundance model (c) (d) Origin Unlikely Likely Top abundance model (d) Origin Unlikely Likely Naive-abundance model (c) Naive-abundance model (c) Top abundance model (d) FIGURE 1 (a) Wood Thrush breeding abundance and sampling locations; (b) likely origins based on stable-hydrogen isotopes for one individual originating in North Carolina; (c) likely origins based on unweighted isotope and breeding abundance (i.e., naive model) for the same individual; (d) likely origins based on the top Wood Thrush model (abundance weight = 100, isotope weight = 10−7) 0.25 0.50 0.75 1.00 Rel. abun. Relative abundance Origin Unlikely Likely Isotope-only model Origin Unlikely Likely Naive-abundance model Origin Unlikely Likely Top abundance model (a) (b) (c) (d) 0.25 0.50 0.75 1.00 Rel. abun. Relative abundance Origin Unlikely Likely Isotope-only model (a) (b) 0.25 0.50 0.75 1.00 Rel. abun. Relative abundance (a) Origin Unlikely Likely Isotope-only model (b) Origin Unlikely Likely Isotope-only model (b) Relative abundance (a) Isotope-only model (b) Origin Unlikely Likely FIGURE 2 (a) American Redstart breeding abundance and sampling locations; (b) likely origins based on stable-hydrogen isotopes for one individual originating in Maryland; (c) likely origins based on unweighted isotope and breeding abundance (i.e., naive model) for the same individual; (d) likely origins based on the top American Redstart model (abundance weight = 10−1, isotope weight = 100). Under this model, the likely origins are still biased toward high abundance locations but lower abundance sites, including the true origin, still receive moderately high posterior support Origin Unlikely Likely Naive-abundance model (c) Origin Unlikely Likely Naive-abundance model Origin Unlikely Likely Top abundance model (c) (d) Naive-abundance model (c) Top abundance model (d) n nlikely kely Origin Unlikely Likely Top abundance model (d) Origin RUSHING et al. 3851 probability imposed by the abundance data overwhelms the likeli- hood estimated from the isotope data. If this is the case, weighting the abundance and/or isotope data may be necessary to obtain unbiased estimates of geographic origins (Rundel et al., 2013). To determine whether weighting the two data sources improved the assignment model performance, we followed Rundel et al. Relative abundance (a) (2013) and weighted the likelihood and abundance prior in equation 2 by raising each to all powers from 10−1 to 10, respectively, resulting in 442 assignment models (222 = 441 abundance models + isotope-only model). Powers >1 sharpen the distribution of values, giving more weight to high val- ues relative to low values. Powers <1, in contrast, flatten the distri- bution and give more relative weight to low values. For each model, we estimated the assignment error and assignment area as described above. locations included in our analysis. To test whether our results were sensitive to the specific sampling locations included in the analysis, we iteratively removed all individuals from each site and re-estimated the top models by comparing the new Pareto optimal models to the isotope-only model (estimated using the same individuals). If the abundance and isotope weights remained constant across these sce- narios, we concluded that the top models were robust to any differ- ences in the sampling locations included in the analysis. Species also vary considerably in their geographic distributions and patterns of abundance, and it is possible that these differences may influence the performance of assignment models based on iso- topes and abundance. Unfortunately, sample sizes for four of the species included in our analysis (Ovenbird, Black-and-White Warbler, Northern Parula, and Prairie Warbler) were too small to obtain reliable estimates of top model weights using the Pareto method described above. Instead, we tested whether the top models identified for Wood Thrush and American Redstart outperformed assignments based on the isotope-only model for the remaining species. For each species, we compared the assignment performance of the top Wood Thrush and American Redstart models to the performance of the respective isotope-only model. If the top Wood Thrush and American Redstart models outperformed the isotope-only models for the other species, we concluded that those models provide a general solution for assign- ments of Neotropical migratory songbirds. After fitting the models for Wood Thrush and American Redstart, we used a two-step approach to determine the top-performing mod- el(s). First, we used an optimization method termed Pareto optimality to eliminate from consideration any model that did not minimize both error rate and assignment area relative to other models. Relative abundance (a) Pareto op- timality is widely used in economics and engineering (Censor, 1977; Steuer, 1986) but has also been applied to a number of optimization problems in ecology and evolution (Kennedy, Ford, Singleton, Finney, & Agee, 2008; Reynolds & Ford, 1999; Shoval et al., 2012). Briefly, Pareto optimality describes a situation where a change in the system (e.g., changing the abundance and/or isotope weights) cannot improve one performance metric without worsening the other. Pareto optimal models are said to “dominate” all other models. For example, in this analysis, Pareto optimal models have both lower assignment area and error rate than all nonoptimal models in the set. The subset of mod- els that are Pareto optimal form the Pareto frontier, along which one cannot improve assignment error without increasing assignment area and vice versa. By considering only models along the Pareto frontier, we were able to eliminate a large number of models from further con- sideration and to explicitly define the trade-offs between assignment error and assignment area for only a few models. 3 \| RESULTS For both Wood Thrush and American Redstart, the majority of the 442 assignment models performed poorly relative to the best mod- els, indicating only a small range of weightings provided reasonable solutions for incorporating abundance into assignments (Figure 3). The isotope-only models had moderate rates for both species (Wood Thrush: 52%; American Redstart: 23%) but also relatively large as- signment areas (Figures 1b and 2b; Wood Thrush: 35%; American Redstart: 22%; Table 2). In contrast, the naive-abundance models had low assignment areas (Figures 1c and 2c; Wood Thrush: 19%; American Redstart: 10%) but high error rates (Wood Thrush: 32%; American Redstart: 75%; Table 2). Although identifying the Pareto frontier allowed us to restrict our attention to models that cannot be strictly improved with regard to both assignment area and assignment error, these models do not nec- essarily represent acceptable solutions for incorporating abundance into assignments. In particular, some models along the Pareto frontier may have large error rates and are thus inappropriate for assignment of unknown-origin birds. Therefore, as a second step in determining the top model(s), we compared the assignment area and the error rate of each Pareto optimal model to the area and error rate of the isotope- only model, under the assumption that any model that includes abun- dance should at least improve upon the isotope-only model. Thus, the final model set included only Pareto optimal models that outper- formed the isotope-only model along both axes. The use of Pareto optimality provided an efficient means of elim- inating poorly performing models. For Wood Thrush, 405 models (92%) were not Pareto optimal, leaving 37 models along the frontier (Figure 3). Only 16 of the 37 Pareto optimal models (43%) outper- formed the isotope-only model with regard to both assignment area and error rate (Figure 3). In general, the best performing Wood Thrush models tended to weight the isotope data by a power <1 (Table S7), indicating that assignments performed best when the likelihoods, but not the prior, were slightly flattened compared to their original distribution. The sampling location validation procedure revealed that the Wood Thrush isotope weights were sensitive to the inclusion of the North Carolina site (Figure 4). Previous analysis of these data indi- cated that the isotope data performed poorly for these individuals, 2.5 \| Sampling location and multispecies validation Species Isotope-only model Naïve-abundance model Area (%) Error (%) Area (%) Error (% Wood Thrush 35 53 19 33 A i R d t t 22 23 10 75 FIGURE 3 Assignment area and error rate of the 442 (a) Wood Thrush and (b) American Redstart assignment models G GGG G G G G G G G G G G G G G G G GGG GGG G G G G G G G G G G G G G G G GGG GGGG G G G G G G G G G G G G G G G GG GGGG G G G G G G G G G G G G G G G GG GGGG G G G G G G G G G G G G G G G GG GGGG G G G G G G G G G G G G G G G G G GGGG G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G GGGG G G GG G GG G G G GG G G G G G GGGG G G GG G G G G G GGG G G G G G GGGG G GG G GG G G G G G GG G GG G G GGG G GG GGG G G G G GG G G GG G GGGGG G G G G G GGGG GGGGGG G GGGGGGGG G G G G GG GGGG GG G GGGGGGGGGGG GG G GGG GG G G GGGGGGGGGGGGG G GGGGGGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGG GGG G GGGGGG G G GGGGGG GGGGGGGGG GGGG G G G GG G GGG G GG GGG G G G G GG GGG GGGGGGGGG GGG G G G GGG GGG GGGGGGG G G G GGG GGG G GGGGGG G G G G Isotope-only model Naive-abundance model Pareto optimal model Top model 0.0 0.2 0.4 0.6 0.00 0.25 0.50 0.75 1.00 Assignment area (%) G G G G G G G G G G G G G G GGG G G GG G G G G G G G G G G G G G G GGG G G GG G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G GGG G G GGG G G G GG G GG G GG G G GG G G G GG G G GG G G G G G GGGG G G GGG G GGG GGGGG G G GG GGG GGG GG G GGG GGGGGGG GGGGG G GGGG GGGG G G G GG G GG G GG GG GG G GGGGG GGG GG G GGGGG GG GGGG G GGG GGGGG GG G GGGG GGGG GGGGG G GGG GGG G G G GG GGG G GGGGG GGG GGG GGG G GG GG GGGGG GG GGGGGGGG G GG GGGGGGGGGG GGGGGGGGGG G GGGGGGG GGGGGGGGGGGGGGG GGGGGGGG GG G G G G GG G G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGG G GG GGGG GG G GGGGGGGGG GG GGGGGGG GGGGGGGGGGGGGG GGGGG G G G GGGG GG G GGGGGGGGG GG GGGGGGG GGGGGGGGGGGGGG GGGGG GG GG 0.0 0.2 0.4 0.6 0.25 0.50 0.75 1.00 Error rate (%) Assignment area (%) Wood thrush American redstart (a) (b) G GGG G G G G G G G G G G G G G G G GGG GGG G G G G G G G G G G G G G G G GGG GGGG G G G G G G G G G G G G G G G GG GGGG G G G G G G G G G G G G G G G GG GGGG G G G G G G G G G G G G G G G GG GGGG G G G G G G G G G G G G G G G G G GGGG G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G GGGG G G GG G GG G G G GG G G G G G GGGG G G GG G G G G G GGG G G G G G GGGG G GG G GG G G G G G GG G GG G G GGG G GG GGG G G G G GG G G GG G GGGGG G G G G G GGGG GGGGGG G GGGGGGGG G G G G GG GGGG GG G GGGGGGGGGGG GG G GGG GG G G GGGGGGGGGGGGG G GGGGGGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGG GGG G GGGGGG G G GGGGGG GGGGGGGGG GGGG G G G GG G GGG G GG GGG G G G G GG GGG GGGGGGGGG GGG G G G GGG GGG GGGGGGG G G G GGG GGG G GGGGGG G G G G Isotope-only model Naive-abundance model Pareto optimal model Top model 0.0 0.2 0.4 0.6 0.00 0.25 0.50 0.75 1.00 Assignment area (%) Wood thrush (a) (a) Wood thrush Assignment area (%) In general, the multispecies validation indicated that the top mod- els for Wood Thrush and American Redstart performed better than the isotope-only model for the four other species included in our anal- ysis (Table 2). 2.5 \| Sampling location and multispecies validation One concern with our approach is that the top models for Wood Thrush and American Redstart may be specific to the sampling RUSHING et al. 3852 a result, including the North Carolina samples in the current analysis favored models that had lower isotope weights than when these indi- viduals were not included (Figure 4). Aside from the influence of the North Carolina site, the cross-validation procedure indicated that the top Wood Thrush models were not highly sensitive to the sampling locations included in the analysis. with only 40% (13/32) of the individuals correctly assigned to their breeding site (Rushing et al., 2014). Downweighting the isotope data likely improved the assignment of these individuals by flattening the likelihood distribution relative to the prior, thus giving more weight to locations with higher abundance (and in this case, the true origin). As For American Redstart, 395 models (90%) were not Pareto op- timal, leaving 47 models along the frontier (Figure 3). In contrast to the Wood Thrush models, the isotope-only model was very close to the Pareto frontier and as a result, only two of the 46 Pareto optimal abundance models outperformed the isotope-only model with regard to both assignment area and error rate (Figure 3). Both models heavily downweighted abundance (weight range = 10−1.0–10−0.9) and but did not weight the isotope data (Table S8). The isotope weights and the abundance weights were similar across all cross-validation models for American Redstart (Figure 4), indicating that the top Redstart models were unaffected by sample site location. 2.5 \| Sampling location and multispecies validation For Ovenbird, Black-and-White Warbler, and Northern Parula, 15 of the 16 of the top Wood Thrush models outperformed the respective isotope-only models with regard to both assignment area and error rate (Tables S9–S11). In contrast, the top American Redstart models performed poorly for these species. For Prairie Warbler, none of the top Wood Thrush models outperformed the isotope-only model (Table S12), although the top American Redstart models did provide slight improvements in assignment area with no increase in error rate. G G G G G G G G G G G G G G GGG G G GG G G G G G G G G G G G G G G GGG G G GG G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G GGG G G GGG G G G GG G GG G GG G G GG G G G GG G G GG G G G G G GGGG G G GGG G GGG GGGGG G G GG GGG GGG GG G GGG GGGGGGG GGGGG G GGGG GGGG G G G GG G GG G GG GG GG G GGGGG GGG GG G GGGGG GG GGGG G GGG GGGGG GG G GGGG GGGG GGGGG G GGG GGG G G G GG GGG G GGGGG GGG GGG GGG G GG GG GGGGG GG GGGGGGGG G GG GGGGGGGGGG GGGGGGGGGG G GGGGGGG GGGGGGGGGGGGGGG GGGGGGGG GG G G G G GG G G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGG G GG GGGG GG G GGGGGGGGG GG GGGGGGG GGGGGGGGGGGGGG GGGGG G G G GGGG GG G GGGGGGGGG GG GGGGGGG GGGGGGGGGGGGGG GGGGG GG GG 0.0 0.2 0.4 0.6 0.25 0.50 0.75 1.00 Error rate (%) Assignment area (%) American redstart (b) (b) American redstart American redstart Assignment area (%) As for Wood Thrush and American Redstart, the naive-abundance models for the remaining four species performed poorly, with a mean error rate of 56% (range = 30%–93%). 2.5 \| Sampling location and multispecies validation 3853 LA MD MI MO NC VT –1.0 –0.5 0.0 0.5 1.0 Isotope weight (log10) (d) LA MD MI MO NC VT −1.0 −0.5 0.0 0.5 1.0 Abundance weight (log10) Cross-validation location (c) LA MD MI MO NC VT −1.0 −0.5 0.0 0.5 1.0 Abundance weight (log10) Cross-validation location LA MD MI MO NC VT –1.0 –0.5 0.0 0.5 1.0 Isotope weight (log10) (c) (d) LA MD MI MO NC VT −1.0 −0.5 0.0 0.5 1.0 Abundance weight (log10) Cross-validation location (c) breeding areas for adult birds captured in the stationary nonbreeding period or while on migration. We emphasize that combining breeding abundance and stable isotopes in assignment models will yield valid estimates of migratory connectivity for adults but not for immature birds because breeding abundance can be a poor indicator of regional productivity (Van Horne, 1983). Studies aimed at determining the natal origins of immature birds could instead estimate range-wide variation in fecundity with data from the Monitoring Avian Productivity and Survival (MAPS) program (Desante, Burton, Saracco, & Walker, 1995).ii distribution of abundance. To test this hypothesis, we fit a negative exponential distribution to the abundance data for each species and compared the rate parameter across species. Patchy abundance distri- butions are expected to be characterized by steeply declining distribu- tions and large rate parameters. More even abundance distributions should be less steep and have lower rate parameters. Consistent with our hypothesis, the rate parameters for American Redstart (23.3, 95% CI = 22.8-23.9; Figure S2) and Prairie Warbler (19.2, 18.3-20.1) are much larger than the parameters for Wood Thrush (8.4, 8.2-8.7), Ovenbird (8.2, 8.0-8.4), Northern Parula (12.3, 11.9-12.7), and Black- and-White Warbler (8.1, 7.8-8.3). The use of Pareto optimality allowed us to define explicitly the trade-off between precision and error and in that way focus only on models that provided potential solutions to the proper weighting of abundance and isotope data. For the four species with relatively even abundance distributions (Wood Thrush, Ovenbird, Black-and-White Warbler, and Northern Parula), several models along the Pareto frontier had both lower error and lower assignment area than models based on isotopes alone. Across these four species, models that weighted the isotope data by 10−0.6–10−0.8 but left the abundance data unweighted provided the best performance. Further downweighting the isotope data resulted in a large increase in error rate for Northern Parula and Black-and-White Warbler, offsetting the improved assignment area for the other species. 2.5 \| Sampling location and multispecies validation Thus, although the naive- abundance models greatly reduced the assignment area relative to the isotope-only models (mean decrease in assignment area compared to isotope-only models = 46%), they performed poorly compared to both the isotope-only model and the top weighted abundance models for all species (Table 2). The divergent results for Wood Thrush, Ovenbird, Black-and- White Warbler, and Northern Parula, on the one hand, and American Redstart and Prairie Warbler, on the other hand, suggest that the performance of abundance models is conditional on the underlying FIGURE 3 Assignment area and error rate of the 442 (a) Wood Thrush and (b) American Redstart assignment models FIGURE 3 Assignment area and error rate of the 442 (a) Wood Thrush and (b) American Redstart assignment models Species Isotope-only model Naïve-abundance model Top abundance model Area (%) Error (%) Area (%) Error (%) Area (%) Error (%) Wood Thrush 35 53 19 33 23 6 American Redstart 22 23 10 75 22 21 Ovenbird 25 63 14 53 18 33 Northern Parula 35 37 18 30 23 20 Black-and-White Warbler 25 53 16 47 24 21 Prairie Warbler 41 56 22 93 39 56 \| 3853 RUSHING et al. FIGURE 4 Sampling site cross- validation results for Wood Thrush (a-b) and American Redstart (c-d). For each site shown on the y-axis, the bars show the range of abundance weights (a, c) and isotope weights (b, d) from the top models when all individuals from that site are removed from the analysis. Top models were determined by comparing the assignment area and error rate of the Pareto optimal models and the assignment area and error rate of the isotope-only models IN MI NC VA VT –1.0 −0.5 0.0 0.5 1.0 Cross-validation location IN MI NC VA VT −1.0 −0.5 0.0 0.5 1.0 LA MD MI MO NC VT −1.0 −0.5 0.0 0.5 1.0 Abundance weight (log10) Cross-validation location LA MD MI MO NC VT –1.0 –0.5 0.0 0.5 1.0 Isotope weight (log10) (a) (b) (c) (d) \| 3853 IN MI NC VA VT –1.0 −0.5 0.0 0.5 1.0 Cross-validation location IN MI NC VA VT −1.0 −0.5 0.0 0.5 1.0 (a) (b) RUSHING et al. increase in error rate for these species. Therefore, this combination appears to provide a reasonable solution for these two species. curve-fitting approach described above. For species with patchy distributions (rate parameter > ~15–20), abundance data should only be included if it is heavily downweighted (10−1–10−0.9). increase in error rate for these species. Therefore, this combination appears to provide a reasonable solution for these two species. These results indicate that the distribution of breeding abundance plays a critical role in the performance of assignment models. For species with breeding ranges that are characterized by large areas of very low abundance and a few small centers of high abundance, this patchy distribution likely magnifies the influence of abundance data in the assignment model, overwhelming the information provided by the isotope data and leading to high posterior probabilities for only the high abundance sites (Hobson et al., 2014). As a result, even models that heavily downweight the abundance data perform relatively poorly compared to the isotope-only model. For species with relatively even distributions, in contrast, downweighting of the abundance data was unnecessary to give appropriate weight to the isotope data. Thus, we suggest that including abundance in assignment models has the most potential to improve assignment of species with relatively less patchy and more even distribution of breeding abundance. 3. 3. Moderately downweight isotope data for species with even abun- dance: For species with more even distributions of breeding abun- dance (rate parameter < 12), the best models tended to moderately downweight the isotope but not the abundance data. Therefore, we recommend for species with rate parameters < 12, isotope data should be weighted 10−0.6–10−0.8 and abundance data should be unweighted. Even when researchers follow these guidelines, we recommend com- paring the results of the abundance model to the results of assignments based on stable isotopes only. As we show here, the models along the Pareto frontier define the upper bound of assignment performance using abundance and stable-hydrogen isotopes and, in some cases, the perfor- mance of abundance models can be poor compared to the isotope-only model. As a result, large discrepancies between the two models should be investigated, reported, and carefully interpreted. Ultimately, however, including abundance in assignment models should be viewed as a pre- liminary step to improving estimates of migratory connectivity for any species. Occupancy probabilities calculated from species distribution models offer a potentially promising alternative to abundance (Fournier et al., 2016); in part, this approach could lessen the influence of areas with particularly high or low abundance. However, the largest improve- ments in assignments likely will come from incorporating multiple intrin- sic markers (e.g., DNA, morphology, or other isotopes), each of which provides complimentary information about the origin of an individual. Given the low cost and large sample sizes associated with assaying intrin- sic markers, assignment models based on these methods will continue to provide important insights into the migratory movements of many species. One of the primary assumptions, often made implicitly, behind add- ing abundance to assignment models is that at any given winter site, individuals mix in frequencies relative to their breeding abundance (González-Prieto et al., 2011). In other words, the logic behind this ap- proach assumes that there is no migratory connectivity. However, sev- eral decades of research on migratory connectivity have shown that complete mixing is the exception rather than the rule. Most migratory species show some degree of migratory connectivity, with different breeding populations migrating to different winter locations. Because the abundance model ignores this nonrandom mixing of individuals, the model will tend to obscure true patterns of migratory connectivity by biasing all assignments toward high abundance locations. Thus, we caution researchers to carefully consider adding abundance to assign- ment models when they know or expect some degree of migratory connectivity. Of course, for species for which there is no a priori information about migratory connectivity, it will be difficult to determine whether or how to include abundance in assignment models. When known- origin samples are available, researchers should always test model per- formance before assignment of unknown-origin individuals. In cases where known-origin samples are not available, we suggest researchers consider the following recommendations: ACKNOWLEDGMENTS C.S.R. was supported by a James Smithson and a Didden Postdoctoral Fellowship from the Smithsonian Institution, and C.E.S. was sup- ported by a Smithsonian Postdoctoral Fellowship. Grants from the U.S Department of Defense’s Legacy Resources Management Program to P.P.M. and C.E.S. (10-427) and Strategic Environmental Research Development Program (#RC-2121) to T. B. Ryder and P.PM funded feather collection. Thanks to the many BBS volunteers that collected data on breeding abundance, the Monitoring Avian Productivity and Survival (MAPS) program for providing additional Wood Thrush feather samples, and to C. France at the Smithsonian Stable Isotope Laboratory for assistance with sample preparation and analysis. 1. Do not use unweighted abundance data: For all six species in- cluded in our analysis, the naive-abundance model was not an optimal solution to model weighting and some cases provided unacceptably high error rates. Therefore, researchers should not use the unweighted abundance estimates in assignment models. 2. Downweight abundance for species with patchy distributions: Our results for American Redstart and Prairie Warbler, two species characterized by very patchy breeding distributions (exponential rate parameters > 19), suggest that including abundance in assign- ments provides only a small improvement over the isotope-only model and only when the abundance data were heavily down- weighted. Therefore, we recommend that researchers first charac- terize the patchiness of breeding distributions using the exponential 4 \| DISCUSSION Given the low resolution of many intrinsic markers, including abun- dance in assignment models is appealing because it can often greatly increase the precision of assignments. However, in all assignment mod- els that use intrinsic markers, there is an inherent trade-off between assignment error and precision. Using known-origin stable-hydrogen isotope samples of six Neotropical migratory bird species, we show that the increased precision of models that include abundance is often off- set by a large increase in assignment error. We demonstrate that proper weighting of the abundance and isotope data can result in models with higher precision and, in some cases, lower error than models based on isotopes alone. These results confirm that breeding abundance can be an important source of information for studies investigating large-scale movements of migratory birds and potentially other taxa. In contrast, including abundance did not provide an unequivocal improvement in assignment performance for the remaining two spe- cies (American Redstart and Prairie Warbler). For these species, the isotope-only model was close to the Pareto frontier. In general, this means that although including abundance in the model will decrease assignment area, it will also increase assignment error. Nevertheless, weighting abundance by 10−1 but leaving the isotope data unweighted did provide a slight decrease in assignment area without substantial Although our assignment models involved known-origin birds, their chief application will be estimating migratory connectivity to 3854 RUSHING et al. REFERENCES Kennedy, M. C., Ford, E. D., Singleton, P., Finney, M., & Agee, J. K. (2008). Informed multi-objective decision-making in environmental manage- ment using Pareto optimality. Journal of Applied Ecology, 45, 181–192.iii Block, B. A., Jonsen, I. D., Jorgensen, S. J., Winship, A. J., Shaffer, S. A., Bograd, S. J., … Ganong, J. E. (2011). Tracking apex marine predator movements in a dynamic ocean. Nature, 475, 86–90. Pyle, P. (1997). Identification guide to North American birds: Part I Columbidae to Ploceidae. Bolinas, Calif: Slate Creek Press. Bowen, G. J., Wassenaar, L. I. & Hobson, K. A. (2005). Global application of stable hydrogen and oxygen isotopes to wildlife forensics. Oecologia, 143(3), 337–348. Reynolds, J. H., & Ford, E. D. (1999). Multi-criteria assessment of ecological process models. Ecology, 80, 538–553. Royle, J. A., & Rubenstein, D. R. (2004). The role of species abundance in determining breeding origins of migratory birds with stable isotopes. Ecological Applications, 14, 1780–1788. Censor, Y. (1977). Pareto optimality in multiobjective problems. Applied Mathematics and Optimization, 4, 41–59. Chabot, A. A., Hobson, K. A., Van Wilgenburg, S. L., McQuat, G. J., & Lougheed, S. C. (2012). Advances in linking wintering migrant birds to their breeding-ground origins using combined analyses of genetic and stable isotope markers. PLoS ONE, 7, e43627. Rubenstein, D. R., & Hobson, K. A. (2004). From birds to butterflies: Animal movement patterns and stable isotopes. Trends in Ecology & Evolution, 19, 256–263. Rundel, C. W., Wunder, M. B., Alvarado, A. H., Ruegg, K. C., Harrigan, R., Schuh, A., … Novembre, J. (2013). Novel statistical methods for inte- grating genetic and stable isotope data to infer individual-level migra- tory connectivity. Molecular Ecology, 22, 4163–4176. Desante, D. F., Burton, K. M., Saracco, J. F., & Walker, B. L. (1995). Productivity indices and survival rate estimates from MAPS, a continent-wide programme of constant-effort mist-netting in North America. Journal of Applied Statistics, 22, 935–948. Rushing, C. S., Ryder, T. B., Saracco, J. F., & Marra, P. P. (2014). Assessing migratory connectivity for a long-distance migratory bird using multiple intrinsic markers. Ecological Applications, 24, 445–456. Flockhart, D. T., Wassenaar, L. I., Martin, T. G., Hobson, K. A., Wunder, M. B., & Norris, D. R. (2013). Tracking multi-generational colonization of the breeding grounds by monarch butterflies in eastern North America. Proceedings of the Royal Society of London B: Biological Sciences, 280, 20131087. Sauer, J. R., Hines, J. E., Fallon, J. E., Pardieck, K. REFERENCES L., Ziolkowski Jr, D. J., & Link, W. A. (2015). The North American Breeding Bird Survey, Results and Analysis 1966–2013. Version 01.30.2015 USGS Patuxent Wildlife Research Center, Laurel, MD. Fournier, A., Sullivan, A. R., Bump, J. K., Perkins, M., Shieldcastle, M. C., & King, S. L. (2016). Combining citizen science species distribution mod- els and stable isotopes reveals migratory connectivity in the secretive Virginia rail. Journal of Applied Ecology. doi:10.1111/1365-2664.12723 Fournier, A., Sullivan, A. R., Bump, J. K., Perkins, M., Shieldcastle, M. C., & King, S. L. (2016). Combining citizen science species distribution mod- els and stable isotopes reveals migratory connectivity in the secretive Virginia rail. Journal of Applied Ecology. doi:10.1111/1365-2664.12723 González-Prieto, A. M., Hobson, K. A., Bayly, N. J., & Gómez, C. (2011). Geographic origins and timing of fall migration of the Veery in northern Shoval, O., Sheftel, H., Shinar, G., Hart, Y., Ramote, O., Mayo, A., … Alon, U. (2012). Evolutionary trade-offs, Pareto optimality, and the geometry of phenotype space. Science, 336, 1157–1160. González-Prieto, A. M., Hobson, K. A., Bayly, N. J., & Gómez, C. (2011). Geographic origins and timing of fall migration of the Veery in northern Colombia. The Condor, 113, 860–868.t Steuer, R. E. (1986). Multiple criteria optimization: Theory, computation, and application. New York: Willey. Sullivan, A. R., Bump, J. K., Kruger, L. A., & Peterson, R. O. (2012). Bat-cave catchment areas: Using stable isotopes (δD) to determine the probable origins of hibernating bats. Ecological Applications, 22, 1428–1434. Hallworth, M. T., Studds, C. E., Sillett, T. S., & Marra, P. P. (2013). Do archival light-level geolocators and stable hydrogen isotopes provide compara- ble estimates of breeding-ground origin? The Auk, 130, 273–282. Van Horne, B. (1983). Density as a misleading indicator of habitat quality. The Journal of Wildlife Management, 47, 893–901. Hobson, K. A., & Wassenaar, L. I. (1997). Linking breeding and wintering grounds of neotropical migrant songbirds using stable hydrogen isoto- pic analysis of feathers. Oecologia, 109, 142–148. Van Wilgenburg, S. L., & Hobson, K. A. (2011). Combining stable-isotope (δD) and band recovery data to improve probabilistic assignment of mi- gratory birds to origin. Ecological Applications, 21, 1340–1351.ii Hobson, K. A. (2008). Isotopic tracking of migrant wildlife. In R. Michener, & K. Lajtha (Eds.), Stable isotopes in ecology and environmental science, 2nd ed. (pp. 155–175). Boston, MA: Blackwell Publishers. Wassenaar, L. I., & Hobson, K. A. (2003). None declared. Kelly, J. F., Ruegg, K. C., & Smith, T. B. (2005). Combining isotopic and ge- netic markers to identify breeding origins of migrant birds. Ecological Applications, 15, 1487–1494. DATA ARCHIVING Upon acceptance, the data used in this manuscript, along with all nec- essary code to reproduce the analyses, will be deposited permanently in the Smithsonian Digital Data Repository. Once archived, the data and code will be freely available to users. RUSHING et al. 3855 REFERENCES Comparative equilibration and online technique for determination of non-exchangeable hydrogen of keratins for use in animal migration studies. Isotopes in Environmental and Health Studies, 39, 211–217. Hobson, K. A., Soto, D. X., Paulson, D. R., Wassenaar, L. I., & Matthews, J. H. (2012). A dragonfly (δ2H) isoscape for North America: A new tool for determining natal origins of migratory aquatic emergent insects. Methods in Ecology and Evolution, 3, 766–772. SUPPORTING INFORMATION CONFLICT OF INTEREST of migratory birds using stable isotopes: Origins of harvested lesser scaup in North America. PLoS ONE, 4, e7915. of migratory birds using stable isotopes: Origins of harvested lesser scaup in North America. PLoS ONE, 4, e7915. None declared. SUPPORTING INFORMATION Hobson, K. A., Van Wilgenburg, S. L., Faaborg, J., Toms, J. D., Rengifo, C., Sosa, A. L., … Aguilar, R. B. (2014). Connecting breeding and wintering grounds of Neotropical migrant songbirds using stable hydrogen iso- topes: A call for an isotopic atlas of migratory connectivity. Journal of Field Ornithology, 85, 237–257. Additional Supporting Information may be found online in the support- ing information tab for this article. Additional Supporting Information may be found online in the support- ing information tab for this article. Hobson, K. A., Van Wilgenburg, S. L., Wassenaar, L. I., & Larson, K. (2012). Linking hydrogen (δ2H) isotopes in feathers and precipitation: Sources of variance and consequences for assignment to isoscapes. PLoS ONE, 7, e35137. How to cite this article: Rushing, C. S., Marra, P. P. and Studds, C. E. (2017), Incorporating breeding abundance into spatial assignments on continuous surfaces. Ecology and Evolution, 7: 3847–3855. https://doi.org/10.1002/ece3.2605 Hobson, K. A., Wunder, M. B., Van Wilgenburg, S. L., Clark, R. G., & Wassenaar, L. I. (2009). A method for investigating population declines
https://openalex.org/W4280637289	https://zenodo.org/records/6554955/files/%D0%90%D0%BD%D0%B2%D0%B0%D1%80%D0%B1%D0%B5%D0%BA%D0%BE%D0%B2%D0%B0%20%D0%A7%D0%B0%D1%80%D0%BE%D1%81%20%D0%90%D0%BD%D0%B2%D0%B0%D1%80%D0%B1%D0%B5%D0%BA%20%D2%9B%D0%B8%D0%B7%D0%B8.pdf	Russian	null	КАРТОШКАНИНГ МОРФОЛОГИК ХУСУСИЯТЛАРИ	Zenodo (CERN European Organization for Nuclear Research)	2,022	cc-by	841	https://doi.org/10.5281/zenodo.6554955 Анварбекова Чарос Анварбек қизи Эсонова Сурайѐ Илѐс қизи Норматова Малохатхон Тўйчибой қизи Тошкент давлат аграр университети Аннотатсия: Картошка ўсимлиги. Картошка пояси. Картошка гули. Барги. Меваси. Гулли навлардагина мева ҳосил қилиш хусусияти Калит сўзлар: Картошка, пояси, гули,барги,меваси, гуллаши, шохланиши . www.interonconf.com Аннотатсия: Картошка ўсимлиги. Картошка пояси. Картошка гули. Барги. Меваси. Гулли навлардагина мева ҳосил қилиш хусусияти Калит сўзлар: Картошка, пояси, гули,барги,меваси, гуллаши, шохланиши . Кириш Картошка томатдошлар оиласига (Solanаceae) мансуб бўлиб, Solanum авлодини ташкил этади. Бу авлод 200 дан зиѐд ѐввойи, ярим ѐввойи ва маданий турларни ўз ичига олиб, шундан фақат битта Sоlаnum tuberosum L. Маданий ҳолда кенг экилади. Палаги ва пояси. Картошка ўсимлиги палак бўлиб, поя ва баргдан иборат. Ҳар бир туп 3-5 та ва зиѐд поя ҳосил қилади. Поя сони нав белги бўлиб, уруғлик материалнинг вазнига ва ундаги нишлаган куртаклар сонига боғлиқ. Картошка пояси тик ва ѐтиб ўсувчан, шохланувчан бўлади. Поя кўндаланг кесими юмалоқ, кўпинча 3-4 қиррали бўлади. Ботаник уруғдан ўстирилган ўсимлик бир поя ҳосил қилади. Картошка навлари поясининг сони, қирралилиги, баландлиги, шохланиши ва рангланганлиги билан фарқланади. Поя сони бўйича навлар кўп ва кам пояли, уч ва кўп қиррали, шохланмайдиган, кучсиз, ўртача ва кучли шохланувчан пояли бўлади. Барги. Туганак ѐки уруғдан кўкариб чиққан ўсимлик барги оддий, четлари бутун бўлади. Ўсимликнинг кейинги ўсиши натижасида чуқур кесилган тоқ патсимон барглар ҳосил бўлади. У барг банди, ўқи, унда жойлашган охирги (учки) бўлак, 3-7 жуфт ѐн бўлаклар ва улар орасидаги оралиқ бўлакчалардан ташкил топган. Баргнинг охирги ва ѐн бўлаклари йириклиги, шакли, ѐн бўлаклар сони, уларнинг жойлашиши, рангланганлиги кабилар муҳим нав белги ҳисобланади. Барг бўлаклари йирик, майда ва ўрта йирикликда, шакли эса овал, узунчоқ, тухумсимон ва турли оралиқ кўринишларда бўлади. Охирги (учки) барг бўлаги асосининг шакли юраксимон, понасимон ва оралиқ кўринишда бўлади. Картошка навлари биринчи жуфт ѐн бўлаклар пластинкасининг эни бўйига нисбати (барг индекси) бўйича ҳам фарқланади. Барг бўлакчалари, йириклиги, шакли, жойлашиши бўйича навлар бир- www.interonconf.com 69 PAGE www.interonconf.com “INTERNATIONAL SCIENTIFIC RESEARCH CONFERENCE” BELARUS, International scientific-online conference катта бўлмаган яшил туганаклар пояларда, барг қўлтиқларида пайдо бўлади. Бунга сабаб, поя пастки қисмларини зараркунанда, касаллик ѐки бошқа шароитлар таъсирида шикастланиш ҳисобланади. Натижада барглардаги фотосинтез маҳсулотлари ер остки қисмлари туганакларга ўта олмайди. “INTERNATIONAL SCIENTIFIC RESEARCH CONFERENCE” BELARUS, International scientific-online conference “INTERNATIONAL SCIENTIFIC RESEARCH CONFERENCE” BELARUS, International scientific-online conference биридан ажратилади. Барг бўлакчалари йирик ва майда, кўриниши юмалоқ, чўзинчоқ ва оралиқ шаклда, жойлашиши ўққа ѐки ўтроқ бўлиб, барг ўқига нисбатан тўғри бурчакли ва аралаш бўлиб ўрнашган бўлади. Агар баргда бўлак ва бўлакчалар кўп бўлса, кучли кесилган, аксинча, кам бўлса кучсиз кесилган, ўта зич жойлашган бўлса, зич баргланган дейилади, аксинча эса сийрак баргланган деб юритилади. Барг ранги оч ва тўқ яшил рангда ялтироқ ѐки ялтироқсиз оддий бўлади. Картошканинг барг банди асоси баргчаларга эга бўлиб, унинг шакли ўроқсимон, баргсимон ва оралиқ кўринишда бўлади. Картошка гули–шингил гултўплам ҳисобланади. Гул тўплам ғуж ва сочма (тарқоқ ѐки шохланган) бўлади. Гулбанди узун ѐки қисқа, ингичка ѐки йўғон бўлиши мумкин. Ғунча – юмалоқ, овал ѐки узунчоқ шаклларда: кучсиз ѐки кучли тукланган ва тукланмаган бўлади. Рангланишининг тақсимланиши ғунчанинг ѐппасига, фақат юқори қисмида ва ташқи кўринадиган қисмида бўлиши билан фарқланади. Гул–гулкоса, гултож, уруғчи ва чангчидан ташкил топган. Гулкосанинг (рангланиши, тукланганлиги, шакли) навнинг характерли белгиларидир. Гулкоса рангланиши бўйича яшил, қисман ва тўлиқ рангланган, кучли ва кучсиз тукланган, унинг барг учлари бигизсимон ва баргсимон бўлади.Гултож ранги оқ, қизил-бинафша, кўк-бинафша, кўк ва ҳоказо бўлади. Гултожбаргнинг қатқатлиги ички ва ташқи бўлиши билан бир-биридан фарқланади. Картошка гулида 5 та чангчи бўлиб, ранги сариқ, оч сариқ, сарғиш-яшил ва тўқ сариқ ѐки апельсинга ўхшаш, шакли тўғри конус, цилиндр ѐки ноксимон, айрим навларда нотўғри шаклда, йириклиги бўйича йирик (катта) ва майда чангчили гулларга бўлинади. Уруғчи – тугунча, найча ва тумшуқчадан ташкил топган. Уруғчи тугунчаси шакли ва ранги бўйича картошка навлари кескин фарқланади. Ўсимликнинг гуллаши навига қараб турлича. Айрим навлар ғунчасини ташлаб юборади ва гулламайди, баъзилари эса гуллайди, лекин чангчиси стериль бўлгани учун мева ҳосил қилмайди. Фақат айрим фертиль гулли навлардагина мева ҳосил қилиш хусусияти яхши бўлади. Ҳаво ҳарорати ва намлиги картошканинг гуллаши ва уруғ ҳосил қилишига сезиларли таъсир этади. Картошка ўзидан чангланувчи ўсимлик. Меваси – резавор, икки уяли, кўп уруғли, сариқ-яшил рангли. Уруғи майда, ясси, сариқ рангда бўлиб, 1000 тасининг оғирлиги 0,5 грамм. Туганак – шакли ўзгарган (метаморфозлашган) поядир. Чунки у ер остки поянинг ѐн (қўлтиқ) куртакларидан ривожланган оқ поя (столон) учида озиқ моддаларнинг тўпланиши натижасида кенгайиб ҳосил бўлади. Столонлар навига қараб узун ва қисқа бўлиши мумкин. Туганак поя тузилишига эга бўлиб, унда куртаклар спирал жойлашган. Баъзан, www.interonconf.com 70 PAGE 70 PAGE ФОЙДАЛАНИЛГАН АДАБИЁТЛАР РЎЙХАТИ: 1. Hooker, W.J., 2001. Compendium of potato Diseases. The American Phytopathological Society, USA, p. 125. 1. Hooker, W.J., 2001. Compendium of potato Diseases. The American Phytopathological Society, USA, p. 125. 2. Ўзбекистон Республикаси ҳудудида экиш учун тавсия этилган қишлоқ хўжалик экинлари Давлат реестри. Тошкент: 2020. – б. 124. 2. Ўзбекистон Республикаси ҳудудида экиш учун тавсия этилган қишлоқ хўжалик экинлари Давлат реестри. Тошкент: 2020. – б. 124. 3. Остонақулов Т.Э., Зуев В.И., Қодирхўжаев О.К. Мевачилик ва сабзавотчилик (Сабзавотчилик): Дарслик. Наврўз. – Тошкент : 2018 (2020). – 552 б. 4. Остонақулов Т.Э. Ўзбекистонда туганак мевали экинлар. Монография: Наврўз. – Тошкент : 2020. – 324 б. 4. Остонақулов Т.Э. Ўзбекистонда туганак мевали экинлар. Монография: Наврўз. – Тошкент : 2020. – 324 б. www.interonconf.com
https://openalex.org/W2059574957	https://europepmc.org/articles/pmc4359443?pdf=render	English	null	Genome mining reveals unlocked bioactive potential of marine Gram-negative bacteria	BMC genomics	2,015	cc-by	12,631	RESEARCH ARTICLE Open Access Machado et al. BMC Genomics (2015) 16:158 DOI 10.1186/s12864-015-1365-z Henrique Machado1,2, Eva C Sonnenschein2, Jette Melchiorsen2 and Lone Gram2 Henrique Machado1,2, Eva C Sonnenschein2, Jette Melchiorsen2 and Lone Gram2 enrique Machado1,2, Eva C Sonnenschein2, Jette Melchiorsen2 and Lone Gram2 © 2015 Machado et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Correspondence: henma@biosustain.dtu.dk 1Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Kogle Allè 6, DK-2970 Hørsholm, Denmark 2Department of Systems Biology, Technical University of Denmark, Matematiktorvet bldg 301, DK-2800 Kgs Lyngby, Denmark Abstract Background: Antibiotic resistance in bacteria spreads quickly, overtaking the pace at which new compounds are discovered and this emphasizes the immediate need to discover new compounds for control of infectious diseases. Terrestrial bacteria have for decades been investigated as a source of bioactive compounds leading to successful applications in pharmaceutical and biotech industries. Marine bacteria have so far not been exploited to the same extent; however, they are believed to harbor a multitude of novel bioactive chemistry. To explore this potential, genomes of 21 marine Alpha- and Gammaproteobacteria collected during the Galathea 3 expedition were sequenced and mined for natural product encoding gene clusters. Results: Independently of genome size, bacteria of all tested genera carried a large number of clusters encoding different potential bioactivities, especially within the Vibrionaceae and Pseudoalteromonadaceae families. A very high potential was identified in pigmented pseudoalteromonads with up to 20 clusters in a single strain, mostly NRPSs and NRPS-PKS hybrids. Furthermore, regulatory elements in bioactivity-related pathways including chitin metabolism, quorum sensing and iron scavenging systems were investigated both in silico and in vitro. Genes with siderophore function were identified in 50% of the strains, however, all but one harboured the ferric-uptake-regulator gene. Genes encoding the syntethase of acylated homoserine lactones were found in Roseobacter-clade bacteria, but not in the Vibrionaceae strains and only in one Pseudoalteromonas strains. The understanding and manipulation of these elements can help in the discovery and production of new compounds never identified under regular laboratory cultivation conditions. High chitinolytic potential was demonstrated and verified for Vibrio and Pseudoalteromonas species that commonly live in close association with eukaryotic organisms in the environment. Chitin regulation by the ChiS histidine-kinase seems to be a general trait of the Vibrionaceae family, however it is absent in the Pseudomonadaceae. Hence, the degree to which chitin influences secondary metabolism in marine bacteria is not known. Conclusions: Utilizing the rapidly developing sequencing technologies and software tools in combination with phenotypic in vitro assays, we demonstrated the high bioactive potential of marine bacteria in an efficient, straightforward manner – an approach that will facilitate natural product discovery in the future. Keywords: AntiSMASH, Genome mining, Pseudoalteromonas, Secondary metabolites, Vibrionaceae Background Additionally, domains for tailoring the monomers can be present. In case of PKSs, such as in fatty acid synthesis, the monomers are acyl-CoAs, while NRPSs connect naturally occurring as well as unnatural amino acids to peptide chains. This wide range of possible sub- units and the possibilities of their combinations lead to the great diversity of polyketides (PKs) and non-ribosomal peptides (NRPs) [7]. For the last century, soil microorganisms have been isolated and screened intensively to discover novel anti- biotics and other drugs, and, in total, microorganisms have supplied more than 80.000 natural products [8]. Today, terrestrial Streptomyces is probably the best exploited genus with respect to secondary metabolites [9-13]. Streptomyces species produce a great diversity of compounds with antifungal (nystatin, natamycin, ampho- tericin), antibacterial (chloramphenicol, streptomycin, holo- mycin) and antiparasitic (ivermectin) activity [14]. Also new cultivation approaches are being used to culture new taxa, which potentially can be a source of novel compounds, as the recently described case of teixobactin [15]. AntiSMASH version 2 is a strong comprehensive tool [30] and includes the use of several of the other tools available, such as the CLUSEAN tools [33], NRPSpredic- tor1/2 [34,35] and a method by Minowa et al. [36]. Even though the occurrence of misidentifications is quite common, it is preferable to “over-identify” rather than missing potential gene clusters [30]. Therefore, comple- menting antiSMASH analysis with more specific tools aids in the gene cluster identification. In this study, we used three other tools: BAGEL3 for the identification of bacteriocins [37]; NapDos for the identification of keto-synthase (KS-domains) and condensation domains (C-domains) [38]; and NP.search for the identification of whole gene clusters that may be composed of several KS- and/or C-domains [39]. C- and KS-domains catalyze the chain formation of the subunits (peptides or acyl- CoAs), respectively and a high number of these domains reflects the richness of bonds possibly made by an organ- ism and the degree of diversity on non-ribosomal peptide synthesis. Even though scientists have started to explore several other habitats than the terrestrial, the marine environ- ment stands out as a hitherto under-explored niche for new bioactive molecules [6,16-19]. Previous studies have indicated that since the environmental conditions are very different from terrestrial habitats, novel compounds and chemical classes are present, and indeed some mar- ine natural products are characterized by the unique marine factors such as halogenation [20-22]. Background disease treatment [1-3]. Chemical synthesis has devel- oped to be faster and cheaper as compared to biological screenings of organisms and extracts, however, chemical synthetic libraries have not provided the expected novel drugs and a high percentage of new chemicals that are introduced into the markets by pharmaceutical compan- ies are actually derived from natural products [4]. Most of the natural products identified are produced by non- ribosomal peptide synthases (NRPSs) and/or polyketide synthases (PKSs) [5,6]. NRPSs and PKSs are multifunc- tional modular enzymes that assemble small molecules from monomers like pearls on a string. Both enzyme types The discovery and development of new molecules for medical treatment is in great need as the 21st century unfolds. Drug-resistant pathogenic microorganisms are becoming a significant threat to public health and the pharmaceutical discovery pipelines have not been deliv- ering the amount of new drugs required for efficient * Correspondence: henma@biosustain.dtu.dk 1Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Kogle Allè 6, DK-2970 Hørsholm, Denmark 2Department of Systems Biology, Technical University of Denmark, Matematiktorvet bldg 301, DK-2800 Kgs Lyngby, Denmark Machado et al. BMC Genomics (2015) 16:158 Page 2 of 12 Machado et al. BMC Genomics (2015) 16:158 Hitherto, the vast majority of bioactive compounds have been found using a classical bioassay-guided process, however, this bioprospecting of drugs is expensive and time-consuming, and re-discovery of known compounds is, despite dereplication steps, a major challenge. The process of drug discovery is currently undergoing changes as a result of the rapid developments in sequencing tech- nology and synthetic biology. The number of whole mi- crobial genomes and metagenomic data made publicly available is increasing exponentially and therefore, (meta) genome mining has become an extremely attractive tool for drug discovery [2,3,16,30,31]. It has led to the develop- ment of new bioinformatic tools used for screening and identification of the genetic background of the bioactiv- ities including gene clusters responsible for the production of the novel molecules. Many of these clusters are prob- ably silent under most laboratory culture conditions and require induction [32]. Several of the bioinformatic tools have been designed to search specifically for PKS and NRPS clusters, of which the structure is conserved. Sev- eral recent reviews provide a comparison between differ- ent tools, considering their modus operandi [30,31]. have core domains responsible for the recognition of the monomer, attachment to the enzyme, condensation and chain-termination. Results and discussion Marine bacterial genomes genome size The genomes were assembled using CLC Genomics Workbench 7 (CLC bio, Aarhus, Denmark) to obtain contig-based draft genomes of the strains. These draft genomes were then annotated using the Rapid Annotation using Subsystem Technology (RAST) [48,49]. The subse- quent analysis of the genomes was performed using CLC Main Workbench 7 (CLC bio, Aarhus, Denmark). During the last five years, we have demonstrated that marine Gram-negative bacteria produce an array of anti- biotic and anti-virulence compounds [19,28,29,40,41,43-45] and here, we ask the question if the classical biopros- pecting approach had fully revealed the potential of these bacteria. We present an in silico study of different marine bacterial genomes, which were analyzed using several of the prediction tools developed for the iden- tification of secondary metabolism pathways, namely antiSMASH, NapDos, Np.search, and BAGEL3 [37,46,47]. We combined the genome mining with phenotypic The genome size varied between 3.6 and 6.2 Mb in the 21 sequenced strains (Table 1). In the six Vibrionaceae, the genomes varied between 4 and 6.2 Mb, and the ge- nomes of the eight Pseudoalteromonas spp. ranged from 4.1 to 6.1 Mb. The genomes of the five strains from the Rhodobactereaceae family were slightly smaller; from 3.6 Table 1 Potential for production of bioactive secondary metabolites from 21 marine bacterial strains Strain Species Genome size (Mb) Antibacterial activity AntiSMASH (total hits) BAGEL3 NapDos NP.search KS-domains C-domains NRPS PKS Mix Trans PKS S2753 Photobacterium halotolerans 5.5 yes 12 0 3 19 1 0 1 0 S2052 Vibrio coralliilyticus 5.4 yes 7 2 7 13 2 0 2 0 S2043 Vibrio coralliilyticus. 5.4 yes 7 2 7 13 2 0 2 0 S2604 Vibrio nigripulchritudo 6.2 yes 9 0 6 17 1 0 0 0 S2394 Vibrio neptunius 5.2 yes 6 1 4 12 1 0 1 0 S2757 Vibrio sp. Background Marine natural products have been isolated and identified from several different sources such as algae, sponges or mol- luscs, however, several recent studies have attributed the production of many of these compounds to microorgan- isms associated with the eukaryotic producer previously identified [23], bringing marine microorganisms to the spotlight of natural product discovery. Following the success of terrestrial streptomycetes as producers of natural products, several researchers have focused their search on marine actinobacteria and the discovery of the first truly marine actinobacterium Salinispora has provided a number of very interesting bioactive compounds, including the anti-cancer compound salinosporamide [24,25]. Also, subsequent mining of the genome demonstrated an impressive number of potentially bioactive gene clusters [16]. The Gram-negative proteobac- teria have generally been thought to have less potential for the production of bioactives than actinobacteria, however, several bioactive compounds have been isolated from the marine genus Pseudoalteromonas and more recently also from strains of the Roseobacter clade and the Vibrionaceae family [19,26-29]. The strains investigated in this genome mining study were isolated during the Galathea 3 global expedition in 2006/7. Antagonistic activity towards the human patho- gen Staphylococcus aureus and the fish pathogen Vibrio anguillarum were the main selection criteria [19]. The Galathea 3 bacterial collection has been used in previous studies where identification of new bioactive compounds has been successful. For instance, Photobacterium halo- tolerans strain S2753 produces novel compound families, the solonamides and ngercheumicins, which interfere with virulence regulation in S. aureus [40-42]. Vibrio nigripul- chritudo strain S2604 produces a novel siderophore: Machado et al. BMC Genomics (2015) 16:158 Page 3 of 12 evaluation of molecules potentially involved in production or regulation of bioactive compounds; namely, quorum sensing signals, siderophores and chitinases. nigribactin [43]. However, also several known antibiotic compounds were re-discovered, for instance, S2753 produces holomycin [28], an antibiotic previously only isolated from terrestrial streptomycetes, and Vibrio coralliilyticus S2052 produces andrimid [28]. Also, in pigmented Pseudoalteromonas, we have re-identified a range of antibiotic compounds (indolmycin, pentabro- mopseudilin, prodigiosin) [44,45]. Results and discussion 4.0 no 2 0 5 0 0 0 0 0 S2040 Pseudoalteromonas piscicida 5.3 yes 14 1 8 58 7 0 1 0 S2724 Pseudoalteromonas piscicida 5.2 yes 10 1 7 30 2 0 2 0 S816 Pseudoalteromonas agarivorans 4.4 no 2 0 5 0 0 0 0 0 S3258 Pseudoalteromonas ruthenica 4.1 yes 3 0 5 0 0 0 0 0 S3137 Pseudoalteromonas ruthenica 4.1 yes 3 0 5 0 0 0 0 0 S4054 Pseudoaltermonas luteoviolacea 6.1 yes 20 1 14 48 3 0 4 1 S2471 Pseudoalteromonas rubra 5.8 yes 17 2 12 56 3 0 2 1 S2151 Halomonas sp. 5.2 no 5 0 7 0 0 0 0 0 S3726 Marinomonas sp. 5.4 yes 5 0 6 17 2 0 0 0 S2292 Spongiobacter sp. 4.7 yes 5 1 3 3 0 0 0 0 S4079 Loktanella sp. 3.6 no 5 1 3 3 0 0 0 0 S4493 Paracoccus sp. 4.0 yes 11 1 3 2 0 0 0 0 S1942 Ruegeria mobilis 4.8 yes 8 1 4 1 0 0 0 0 F1926 Ruegeria mobilis 4.6 yes 9 0 5 1 0 0 0 0 DSM17395 Phaeobacter inhibens 3.8 yes 9 + 1 0 4 1 1 0 0 0 tion of bioactive secondary metabolites from 21 marine bacterial strains ble 1 Potential for production of bioactive secondary metabolites from 21 marine bacterial strains Machado et al. BMC Genomics (2015) 16:158 Page 4 of 12 to 4.8 Mb. The in vitro bioactivity (antibacterial activity measured as zone size) [19] did not correlate to the gen- ome size (Table 1). general, few secondary metabolites have been identified in these strains [54-57]. Here, we show that using bio- informatics tools a few clusters could be identified, but still the bioactive potential harbored in the genome of these genera appears much lower than that observed in Gammaproteobacteria. It has been suggested that the potential for production of secondary metabolites would be related to genome size [11,50,51], with a larger genome allowing more genes to be allocated to secondary metabolism. This no- tion was to some extend developed by studies of the genus Streptomyces which is a prolific producer of sec- ondary metabolites and has relatively large genomes of approx. 8 Mb in size as compared to other bacteria. This understanding is changing, as the marine actinomycete Salinispora sp. has a genome size of approx. Results and discussion 5 Mb, of which approx. 10% is dedicated to secondary metabolism, whereas only approx. 8% of the genome of Streptomyces coelicolor has been reported as dedicated to secondary metabolism [11,16]. A number of strains that were not antagonistic in agar-based assays were included in the analysis and these contained only few gene clusters potentially coding for secondary metabolites (Table 1). This was the case for Vibrio sp. S2757 and P. agarivorans S816, for which anti- SMASH identified only two potential clusters (Table 1). Another interesting group of strains included those that received five hits in total in the antiSMASH analysis. This includes the bioactives Marinomonas sp. S3726 and Spongiobacter sp. S2292 and the non-bioactives Halomo- nas sp. S2151 and Loktanella sp. S4079. Although, all of them had a considerably lower number of hits in the anti- SMASH analysis than the pigmented pseudoalteromo- nads and the vibrios, the results of the other mining tools (NaPDoS) demonstrate that Marinomonas sp. S3726 has a great potential with 6 KS- and 17 C-domains identified (Table 1). Thus, the sole number obtained by one given analysis tool may not reflect the whole potential of the or- ganism, and complementary analysis should be performed to ensure discovery of the full bioactive potential. This should also be done to avoid further work on clusters that may not be true secondary metabolite clusters, it appears from the analysis that NapDos and NP.search tools seemed to identify only a subset of the NRPS/PKS clusters identified by antiSMASH. Identification of gene clusters potentially encoding secondary metabolites The genomes were mined using bioinformatic tools for the identification of clusters involved in secondary me- tabolism, namely antiSMASH, NapDos, Np.search, and BAGEL3 [37-39,47]. We found a high genetic potential for secondary metabolite production also in Gram-negative marine bacteria with genome sizes ranging from 4 to 6 Mb, with some strains reaching the considerable number of eight distinct PKS/NRPS clusters (Table 1 – NP.search). However, some strains with similar genome size harbored none or very few potential bioactive clusters and thus, there was no clear correlation between genome size and number of secondary metabolism gene clusters. Some strains, such as V. nigripulchritudo S2604 or Halomonas sp. S2151, with larger genomes had a low number of hits; and also contrarily, strains with smaller genomes had a greater number of hits e.g. P. piscicida strains S2040 and S2724 (Table 1). The potential for secondary metabolite production in the strains is clearly much larger than so far identified by bioassay-guided fractionation. For instance, the PK/NRP hybrid andrimid has been identified as the bioactive com- pound in V. coralliilyticus S2052 [28,29,58,59]. The gen- ome mining identified the gene cluster likely encoding for andrimid production genes (Figure 1(A)). Also, we found at least three more NRPS clusters using antiSMASH, Nap- Dos and NP.search (Table 1). Similarly, in P. halotolerans S2753, the dithiolopyrrolone holomycin was identified in extracts [28] and the corresponding gene cluster was found by the bioinformatic tools used (Figure 1(B)); again, four more NRPS/PKS clusters were found using anti- SMASH, although only one more was discovered when using NP.search (Table 1). As indicated, we and others have identified several bioactive compounds from pigmen- ted pseudoalteromonads and here we also identified the respective gene clusters for indolmycin [44], violacein [60] and pentabromopseudilin [5,61,62]. However, the pig- mented pseudoalteromonads contained a large number of potential bioactive clusters, including a very high number of C-domains as compared to the other studied strains (NapDos – Table 1). The case of Pseudoalteromonas ruthenica finding novel natural bioactive compounds, including antibiotics [63]. Yet, all the bioinformatic tools used to search for the biosynthetic capabilities and potential of P. ruthenica failed. This might be the case because the antagonistic activity is due to other biosynthetic path- ways as is for instance the case with the antibiotic tro- podithietic acid produced by some Roseobacter clade bacteria [64]. Also, it can be attributed to limitations in the prediction algorithms. The prediction algorithms of the bioinformatic tools are to some extend based on identification of known biosynthetic activities and one could speculate that truly novel biosynthetic pathways would not be identified. To identify the core genes of a biosynthetic pathway, most of the tools available use profile-HMMs or alignments of conserved domains in biosynthetic enzymes [30]. This is a problem in the iden- tification of non-standard pathways and antiSMASH has therefore implemented an algorithm to identify the dis- tribution of protein domains usually associated with sec- ondary metabolites [30], increasing the probability of identification of clusters responsible for secondary me- tabolites production. This not only increases the number of hits, but also the time needed for evaluation of the clusters, raising the question of the feasibility of using genome mining in groundbreaking discoveries. In in vitro assays, Pseudoalteromonas ruthenica is highly antagonistic against S. aureus and V. anguillarum caus- ing large clearing zones in agar-based screening assays [19]. However, we have not been able to identify the compound(s) responsible for this inhibition by bioassay- guided fractionation and anticipated that genome mining would reveal potential bioactive gene clusters. Anti- SMASH identified three gene clusters (one for sidero- phore and two for bacteriocin biosynthesis), but only the siderophore cluster was correctly identified, whereas the bacteriocin-related clusters were misidentified and encoded the flagella operon and a cluster encoding for hypothetical proteins, a muramoyltetrapeptide carboxy- peptidase and a 2,3,4,5-tetrahydropyridine-2,6-dicarboxy- late N-succinyltransferase, involved in the biosynthesis of peptidoglycan and lysine, respectively. A second analysis of the P. ruthenica strains with anti- SMASH based on PFAM domain probabilities increased the number of potential gene clusters from three to thir- teen. Mainly clusters encoding for acyl carrier proteins were identified, but we also identified some biosynthetic clusters such as lipopolysaccharide, capsular polysac- charide, legionaminic acid and fatty acid biosynthesis. From all the clusters, only one matched with the RAST annotation as behaving an open reading frame (ORF) encoding a non-ribosomal peptide synthase. The case of Pseudoalteromonas ruthenica Yet this ORF was only 663 bp, and when we blasted the pre- dicted aminoacid sequence against the NCBI protein database, it presented a high similarity with a methionyl- tRNA formyltransferase and not to an NRPS. In agree- ment with these were NapDos and NP.search, which did not identify any potential bioactive clusters (Table 1). This reduces the likelihood that the clusters identified by antiSMASH using PFAM domains are actually clusters responsible for the production of bioactives. Bioactivity potential - NRPS/PKS The presence of gene clusters likely encoding bioactive compounds is spread among the different families of Alpha- and Gammaproteobacteria. Although our collec- tion is limited in number, it appears that the Gamma- proteobacteria class is richer in NRPS and PKS clusters than the Alphaproteobacteria. The analysis using NapDos and NP.search, in general, identified the same number of potential bioactive gene clusters. A higher frequency of KS- and C-domains was identified in pigmented Pseudoal- teromonas strains (S2040; S2724; S4054; S2471) followed by Vibrionaceae, with the exception of S2757 (no hits), and Marinomonas sp. S3726 (high number of hits). Some species in the Rhodobactereaceae family (Ruegeria mobilis and Phaeobacter inhibens) are capable of inhibit- ing a wide range of other bacteria [52-54]; however, in Machado et al. BMC Genomics (2015) 16:158 Page 5 of 12 Figure 1 Previously known clusters identified in the studied marine bacteria, using genome mining. Andrimid gene cluster from V. coralliilyticus S2052 (A); Holomycin gene cluster from P. halotolerans S2753 (B). Figure 1 Previously known clusters identified in the studied marine bacteria, using genome mining. Andrimid gene cluster from V. coralliilyticus S2052 (A); Holomycin gene cluster from P. halotolerans S2753 (B). Acyl homoserine lactones Four of the 21 strains induced a clear response in the AHL (acyl homoserine lactone) biomonitors (Table 2). Three strains, Vibrio sp. S2757, Paracoccus sp. S4493 and P. luteoviolaceae 4054 induced both monitors whereas P. inhibens DSM17395 induced only A. tumefaciens. This is in agreement with previous studies where also Phaeo- bacter sp. strain S27-4 induced A. tumefaciens and chem- ical analysis identified 3-hydroxy-decanoyl-homoserine lactone [64]. Interestingly, antiSMASH detected homoser- ine lactone synthases in three of these four strains but not in Vibrio S2757. The response in the monitor strains could be caused by other compounds, such as diketopiperazines that have been demonstrated to induce the AHL monitors [65]. The same could be true for the extracts of the V. coralliilyticus strains S2052 and S2043, which resul- ted in a weak reaction in C. violaceum, and the genomes did not contain an AHL synthase gene. AntiSMASH Table 2 Iron system in the studied strains, comprising in silico and phenotypical results Strain Species Response in AHL monitor AntiSMASH Siderophore (CAS) AntiSMASH Fur Cv At HSL Siderophore NRPS S2753 Photobacterium halotolerans - - 0 + 1 4 1 S2052 Vibrio coralliilyticus (+) - 0 - 1 4* 1 S2043 Vibrio coralliilyticus (+) - 0 (+) 1 4* 1 S2604 Vibrio nigripulchritudo - - 0 - 0 4* 1 S2394 Vibrio neptunius - - 0 (+) 1 3* 1 S2757 Vibrio sp. + + 0 + 1 0 1 S2040 Pseudoalteromonas piscicida - - 0 + 0 11* 1 S2724 Pseudoalteromonas piscicida - - 0 + 0 5* 1 S816 Pseudoalteromonas agarivorans - - 0 + 1 0 1 S3258 Pseudoalteromonas ruthenica - - 0 + 1 0 1 S3137 Pseudoalteromonas ruthenica - - 0 - 1 0 1 S4054 Pseudoaltermonas luteoviolacea + + 1 (+) 0 11* 1 S2471 Pseudoalteromonas rubra - - 0 (+) 0 9* 1 S2151 Halomonas sp. - - 0 (+) 1 0 1 S3726 Marinomonas sp. - - 0 + 0 3* 1 S2292 Spongiobacter sp. - - 0 - 0 1 1 S4079 Loktanella sp. - - 1 (+) 0 1*** 1 S4493 Paracoccus sp. Bacteriocins Th b The number of clusters identified by antiSMASH as bac- teriocins varied between one and five in each strain, with an average of two clusters per strain. However, when the genomes were analyzed using the prediction tool BAGEL3 [37], the presence of bacteriocin-related genes was only confirmed in a few strains. The distribution of bacteriocin clusters did not follow a particular pattern with respect to genera or species. It seems evident that the specific prediction tools are more accurate in identifying their defined target; therefore, BAGEL 3 being most probably a better indicator of the number of bacteriocin-related genes than antiSMASH In genome mining, the identification of clusters likely involved in secondary metabolism, such as NRPS and PKS, have been used as a measure of the potential for Page 6 of 12 Machado et al. BMC Genomics (2015) 16:158 detected AHL synthase genes in three strains (Loktanella sp. and two Ruegeria mobilis) where no AHLs were de- tected by the monitors (Table 2). These genes could po- tentially encode novel AHLs not being in the detection range of the used biological monitors [66]. On the other hand, the bacteria may not have been cultured under con- ditions allowing the expression of the presumed AHL syn- thase genes or the AHL concentration produced was below the detection limit. We considered if the potential QS systems could be involved in production of secondary metabolites. In P. luteoviolaceae, the AHL synthase gene is adjacent to the gene cluster potentially involved in indolmycin production [67], but in the other five strains the HSL synthase genes detected by antiSMASH were not in proximity to identified natural product gene clus- ters. However, some were close to genes encoding acyl synthases, alcohol dehydrogenases or proteins contain- ing AMP-binding domains, which may potentially be in- volved in secondary metabolism. Due to draft genomes with multiple contigs, the association with natural prod- uct gene clusters could have been lost in the analyzed sequences. itself. This becomes more evident from the P. ruthenica case, where random genes were classified as bacteriocins (see above). Cv: Chromobacterium violaceum, At: Agrobacterium tumefaciens, HSL: homoserine lactone, CAS: chrome-azurol-S, + : strong bioactivity, (+) : weak bioactivity, −: no bioactivity detected under the tested conditions, NRPS: including single NRPS clusters and NRPS fusion clusters (e.g. NRPS-bacteriocin, NRPS-ectoine). Located on a plasmid; Cluster identified as a siderophore – ectoine cluster; *At least one NRPS is in proximity to siderophore-associated genes (tonB-dependent receptor etc.). Siderophores and iron regulation For twelve strains, these siderophore-associated genes were found close to the NRPS gene leading to the hy- pothesis that this NRPS gene could likely encode a siderophore-producing NRPS. This would demonstrate that all strains based on their genetic information would be capable of scavenging iron using siderophores. To detect this “hidden” activity for the five non-active strains, the strains might require optimization of culture conditions or certain biological cues from the environ- ment. Iron can also be scavenged by other molecules and non-siderophore iron sequestering systems may be oper- ational in the bacteria where siderophore genes were not detected. Indeed, several heme-related proteins were identified among the studied marine bacteria by an annotation-based search (data not shown). Iron is essential for almost all microorganisms being required for key biological processes [68] and is also one of the most important requirements for successful secondary metabolism. The iron levels in seawater are extremely low, and many marine bacteria are able to sequester iron using siderophores that can also serve as a tool in microbial com- petition. Hence, siderophores are included as secondary metabolites in the antiSMASH search. To complement the genetic search, we determined sidero- phore activity using the CAS assay [69]. Pronounced sidero- phore activity was detected in eight strains and a weak reaction was observed in eight further strains. Only five strains did not show any activity under the tested conditions (Table 2). The NRPS prediction tools, NapDos and NP.search, do not allow detailed prediction of the type of NRPS coding gene, however, antiSMASH is able to distinguish siderophore syn- thesis genes. The in silico analysis using antiSMASH identified putative siderophore gene clusters in five of the eight strains with a clear CAS reaction, and three of the eight with a weak reaction. In one strain, P. ruthenica, antiSMASH detected a siderophore synthesis gene, but the CAS assay was nega- tive. In contrast, the CAS reaction was positive for three strains (two P. piscicida and one Marinomonas) where a siderophore biosynthesis gene was not detected. Anti- SMASH predicts siderophore genes using the currently available sequence information on siderophore-producing NRPSs, which are mainly of terrestrial origin. Terrestrial siderophores differ structurally from marine siderophores that are usually associated with fatty acids [68]. We ana- lysed the identified NRPS gene clusters for siderophore- associated genes such as tonB-dependent receptor genes. Siderophores and iron regulation Even though iron is essential for growth, excess of iron can be toxic to bacteria and thus a tight regulation of up- take is crucial for microbial survival [68]. In Gram-negative bacteria, iron regulation is achieved by a repressor protein named Fur (Ferric-iron uptake regulator) which acts at the transcriptional level [70]. A Fur encoding gene could be identified in all the studied strains and the amino acid sequence predicted, with the sole exception of Paracoccus sp. S4493 (Table 2; Figure 2(A)). The verified exception of Paracoccus sp. S4493 might be due to sequencing limita- tions, or the fact that this organism has another regu- latory protein involved in iron sensing; in fact other uptake regulators for different metals could be identified (e.g. manganese, potassium, zinc, and nickel). Within the classes of Proteobacteria, the Fur proteins are relatively conserved at the amino acid level, presenting a higher variation at the C-terminus and the N-terminus. Siderophores and iron regulation MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QS PDNQH - I S AED V YK I L L D KGE E I GL A T V YR V L NQFDD AG I V TRHH F E MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QS PDNQH - I S AED V YK I L L D KGE E I GL A T V YR V L NQFDD AG I V TRHH F E MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QS PDNQH - I S AED V YK I L L D L GE E I GL A T V YR V L NQFDD AG I VSRHH F E MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QS PDNQH - I S AED V YK I L L D L GE E I GL A T V YR V L NQFDD AG I VSRHH F E MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QC PDNQH - I S AED V YK I L L D KGE E I GL A T V YR V L NQFDD AG I V TRHH F E MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QS P ENQH - I S AED V YK I L L D KGE E I GL A T V YR V L NQFDD AG I V TRHH F E MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QS PDNQH - I S AED V YK I L L DN S E E I GL A T V YR V L NQFDD AG I V TRHH F E MTDHN L E - - - - - L K K AGL K V T L PR I K I L E I L QS PDNQH - I S AED V YK I L L DN S E E I GL A T V YR V L NQFDD AG I V TRHH F E MSDNNQA - - - - - L KD AGL K V T L PR L K I L E V L QQP ECQH - I S AE E L YK K L I D L GE E I GL A T V YR V L NQFDD AG I V TRHH F E MSDNNQA - - - - - L KD AGL K V T L PR L K I L E V L QQP ECQH - I S AE E L YK K L I D L GE E I GL A T V YR V L NQFDD AG I V TRHH F E MSDNNQA - - - - - L KD AGL K V T L PR L K I L E V L QQP ECQH - I S AE E L YK K L I D L GE E I GL A T V YR V L NQFDD AG I V TRHH F E MSDNNQA - - - - - L KD AGL K V T L PR L K I L E V L QQP ECQH - I S AE E L YK K L I D L GE E I GL A T V YR V L NQFDD AG I V TRHH F E MPDNNQA - - - - - L K E AGL K V T L PR L K I L E V L QQPDC SH - I S AE E L YK K L I D L GE E I GL A T V YR V L NQFDD AG I V TRHH F E MSDNNQA - - - - - L KQAGL K V T L PR L K I L E V L QN P ECQH - I S AED L YK K L I D I GE E I GL A T V YR V L NQFDD AG I V TRHH F E MAD KNH E - - - - - L R K AGL K V T L PR VK I L Q I L E S A TDQHH L S AED V YK T L L E AGED VGL A T V YR V L TQF E S AGL V I RHN FD MT S ENQE - - - - - L K K AGL K V T L PR VK I L Q I L E S AGERH - M S ADD V YR T L L EQGED VGL A T V YR V L TQF EN AGL V VRHH F E MT P TH E EN AN KWL ME AGL R P TRQR V T L A A L L VGDGQHRH V T AE S L FD A AK DDGA A VS L A T V YN T L R A FCD AGV L QE I T VD MT P TH E EN AN KWL ME AGL R P TRQR V T L A A L L VGDGQHRH V T AE S L FD A AK DDGA A VS L A T V YN T L R A FCD AGV L QE I T VD MT P T P AD I A T NWL E T AGL R P TRQR V A L AE L L VGDGKHRH V T AE S L FD A AK D KGA A VS L A T V YN T L R A FCD AGV L QE I T VD MTD I V TQR S S DWL SGAGL R P TRQRMT L AS L L VGDGQDRH V T AE S L FD A AS S AGE K VS L A T V YN T L R A FCD AGL MR E I T VD GGK S V F E L SG S THHDH L VC L KCGK V VE F ED D V I E TRQE E I AN SNG I K L TN H S L Y L YGEC E D K E AC K K Y AE ENG - - - N GGK S V F E L SG S THHDH L VC L KCGK V VE F ED D V I E TRQE E I AN SNG I K L TN H S L Y L YGEC E D K E AC K K Y AE ENG - - - N GGK S V F E L SG S THHDH L VC L KCGK V VE F ED D L I ERRQL E I AE ENG I T L TN H S L Y L YGECQ D K A AC K A F SD SN - - - - S GGK S V F E L SG S THHDH L VC L KCGK V VE F ED DM I ERRQL E I AE ENG I T L TN H S L Y L YGECN D K A AC K E F SD AN - - - - S GGK S V F E L SG S THHDH L VC L KCGK V I E F ED E L I ERRQE E I AK ENG I K L TN H S L Y L YGECN D E E AC KN YGD E - - - - - - GGK S V F E L SG S THHDH L VC L KCGK V I E F ED E V I E TRQVE I A T SNG I K L TN H S L Y L YGEC V D E E ECRR F T E SD E - - - S GGK S V F E L AG S THHDH L VC L KCGR V I E F ED DM I ERRQVE I AE ENG I T L TN H S L Y L YGEC K D VE AC KN Y AE MSN - - - N GGK S V F E L AG S THHDH L VC L KCGR V I E F ED DM I ERRQVE I AE ENG I T L TN H S L Y L YGEC K D VE AC KN Y AE MSN - - - N GGK S V F E L S T QHHHDH L VC L DCGE V I E F AD D I I E ERQK E I ASQYN V I L TN H S L Y L YGKC A D - GSCRDN PD AH K - K K - GGK S V F E L S T QHHHDH L VC L DCGE V I E F AD D I I E ERQK E I ASQYN V I L TN H S L Y L YGKC A D - GSCRDN PD AH K - K K - GGK S V F E L S T QHHHDH L VC L DCGE V I E F AD D I I E ERQK E I AKQYN V I L TN H S L Y L YGKC A D - GSCRDN PD AH K L K K - GGK S V F E L S T QHHHDH L VC L DCGE V I E F SD D V I E ERQK E I AE K YN V I L TN H S L Y L YGKCG D - GSC KHN PD AH K P K S - GGK S V F E L S T QHHHDH L VC L DCGE V I E F SD E V I E ERQR E I A AK YN VE L TN H S L Y L YGKCG D - R SC KDD PN AH K P K K K GGK S V F E L S T QHHHDH L VC L DCGK V I E F SD D L I E ERQKQ I AE S YN I R L TN H S L Y L YGHC T A - GDCN KD E S L HN E K K - GGH A V F E L SQ E EHHDHMVC L E SGE I I E F FD E T I ERRQQE I AE EHGF E L VD H A L V L Y VR P K GS K A TRQEG I AK K - - - - AGT A V F E I AK GEHHDHMVCM D SGK V I E F YD P I I E KRQK E I A AEHGYE I ED HN L V L Y VR P K D - - - - - - - - - - - - - - - - GS K S Y FD - TN VHDH PH Y YWE GEGR VSD AP S E E L V I QS L PQ P P E - GME I AS VD V V I - - - - - - - - R L R K K AE L S - - - - - GS K S Y FD - TN VHDH PH Y YWE GEGR VSD AP S E E L V I QS L PQ P P E - GME I AS VD V V I - - - - - - - - R L R K K AE L S - - - - GS K S Y FD - TN THDH PH F YWE D EGR L SD AP S DQL V I K S L P A APQ - GVE I AS VD V V I - - - - - - - - R L R K - - - - - - - - - I GAK S Y FD - TN MTDH PH F YWE D T AH L TD AP A EQL E I AR VPH AP E - GAE I AS VD V V I - - - - - - - - R L RR K - - - - - - - - - A S1942 F1926 DSM17395 S4079 S2151 S3726 S2604 S2394 S2043 S2052 S2757 S2753 S4054 S3137 S3258 S816 S2292 S2724 S2040 S2471 0.1 Rhodobacteraceae Other Gammaproteobacteria Vibrionaceae Pseudoalteromonadaceae B Figure 2 Analysis of the Fur sequences of the marine bacterial strains used. Acyl homoserine lactones + + 4 - 0 2 0 S1942 Ruegeria mobilis - - 2 (+) 0 1* 1 F1926 Ruegeria mobilis - - 2 (+) 0 1* 1 DSM17395 Phaeobacter inhibens - + 2 + 1* 1 1 Cv: Chromobacterium violaceum, At: Agrobacterium tumefaciens, HSL: homoserine lactone, CAS: chrome-azurol-S, + : strong bioactivity, (+) : weak bioactivity, −: no bioactivity detected under the tested conditions, NRPS: including single NRPS clusters and NRPS fusion clusters (e.g. NRPS-bacteriocin, NRPS-ectoine). Located on a plasmid; Cluster identified as a siderophore – ectoine cluster; **At least one NRPS is in proximity to siderophore-associated genes (tonB-dependent receptor etc.). Table 2 Iron system in the studied strains, comprising in silico and phenotypical results Page 7 of 12 Machado et al. BMC Genomics (2015) 16:158 Siderophores and iron regulation Predicted Fur protein sequence alignment (A); Phylogenetic neighbor joining tree using Jukes-Cantor protein distance measurement method (B). A B S1942 F1926 DSM17395 S4079 S2151 S3726 S2604 S2394 S2043 S2052 S2757 S2753 S4054 S3137 S3258 S816 S2292 S2724 S2040 S2471 0.1 Rhodobacteraceae Other Gammaproteobacteria Vibrionaceae Pseudoalteromonadaceae B Rhodobacteraceae Other Gammaproteobacteria Vibrionaceae Vibrionaceae Figure 2 Analysis of the Fur sequences of the marine bacterial strains used. Predicted Fur protein sequence alignment (A); Phylogenetic neighbor joining tree using Jukes-Cantor protein distance measurement method (B). Machado et al. BMC Genomics (2015) 16:158 Page 8 of 12 Table 3 Chitinolytic systems in the studied strains, comprising in silico and phenotypic results Strain Species Chitinase activity Chitin Chitinase ChiS CBP S2753 Photobacterium halotolerans +++ 3 1 1 S2052 Vibrio coralliilyticus ++ 9 1 2 S2043 Vibrio coralliilyticus ++ 9 1 2 S2604 Vibrio nigripulchritudo + 8 2 0 S2394 Vibrio neptunius ++ 7 1 0 S2757 Vibrio sp. ++ 3 1 1 S2040 Pseudoalteromonas piscicida ++ 4 1 0 S2724 Pseudoalteromonas piscicida +++ 4 0 0 S816 Pseudoalteromonas agarivorans - 0 0 0 S3258 Pseudoalteromonas ruthenica ++ 3 0 1 S3137 Pseudoalteromonas ruthenica ++ 3 0 1 S4054 Pseudoaltermonas luteoviolacea + 10 0 0 S2471 Pseudoalteromonas rubra + 7 0 0 S2151 Halomonas sp. - 0 0 0 S3726 Marinomonas sp. - 0 0 0 S2292 Spongiobacter sp. - 0 0 1 S4079 Loktanella sp. - 0 0 0 S4493 Paracoccus sp. - 0 0 0 S1942 Ruegeria mobilis - 0 1 0 F1926 Ruegeria mobilis - 0 1 0 DSM17395 Phaeobacter inhibens - 0 0 0 ChiS: chitin catabolic cascade sensor histidine kinase, CBP: chitin binding proteins. - : no chitinase activity detected, + : low chitinase activity, ++ : medium chitinase activity, +++ : strong chitinase activity. Nevertheless, the conserved features such as DNA- binding-α-helix and Fe2+ and Zn2+ binding domains could be identified [71]. A neighbor joining tree using Jukes-Cantor protein distance measurement method (Figure 2(B)) demonstrates the conservation of closely related species, indicating that the fur gene is a phylogenetic trait instead of a random species variation or a product of recent horizontal gene transfer. In fact, the clusters based on protein sequences follow the phylogenetic distribution; the analyzed Alphaproteobacteria sequences form a separ- ate, distant group from the families of Gammaproteobac- teria, in which the Pseudoalteromonadaceae and the Vibrionaceae families form two distinct clusters from the other Gammaproteobacteria. Siderophores and iron regulation The only exception here was the Spongiobacter sp. S2292, which clustered together with the Pseudoalteromonas spp. This is interesting, since the 16S rRNA sequence (GenBank acc. no. FJ457273.1) would place Spongiobacter sp. S2292 closer to the Endozoicomonas genus and therefore within the order of Oceanopirillales, in which the species Halomonas and Marinomonas are also in- cluded [72]. This fact brings to question the phylogenetic placement of Spongiobacter as it remains an unclassified member of Gammaproteobacteria, and this association indi- cates a closer association with Pseudoalteromonas species than with the other Gammaproteobacteria. Table 3 Chitinolytic systems in the studied strains, comprising in silico and phenotypic results Interestingly, this clustering seems to be specific at species level, even for the Vibrio strains studied. Identify- ing Vibrio strains to species level typically requires multi- locus sequence analysis [73,74]. We recently showed that the fur gene is a good phylogenetic marker (Machado & Gram, submitted) to be added to the multilocus sequen- cing analysis performed nowadays in e.g. Vibrio species definition [73-75] and might also be possibly used in other genera for species differentiation. ChiS: chitin catabolic cascade sensor histidine kinase, CBP: chitin binding proteins. - : no chitinase activity detected, + : low chitinase activity, ++ : medium chitinase activity, +++ : strong chitinase activity. ChiS: chitin catabolic cascade sensor histidine kinase, CBP: chitin binding proteins. - : no chitinase activity detected, + : low chitinase activity, ++ : medium chitinase activity, +++ : strong chitinase activity. Bioinformatic analysis The draft genomes were annotated using RAST [49] and submitted to secondary metabolite gene cluster analysis using antiSMASH 2.0 [47], NapDos [38], NP.search [39], as well as to the bacteriocin-specific software BAGEL 3 [37]. Following RAST annotation, a homology search was conducted on the ferric-iron uptake regulator gene fur and an annotation-based search was performed for genes encoding, chitinases and the chitin catabolic cas- cade sensor gene chiS. Conclusions Here, we presented a straightforward, comprehensive gen- ome mining approach analyzing marine bacterial strains for secondary metabolism and associated features such as quorum-sensing, iron acquisition, chitin use as a carbon source and its regulation. The use of complementary tools for genome mining is of great value in narrowing down the potential gene clusters from a large pool obtained by broad prediction software such as antiSMASH. We dem- onstrated the great potential of marine bacteria for sec- ondary metabolite production, with special focus on Vibrio and pigmented Pseudoalteromonas species. Chitinases and regulation been also shown that natural substrates such as chitin in- fluence secondary metabolite production, such as the in- duction of their production [58]. We searched for the chiS gene, which was present in all of the six Vibrionaceae, one Pseudoalteromonas, and two Ruegeria (Table 3). The Alphaproteobacteria did not degrade chitin, though two Ruegeria mobilis strains harbor the chitin sensor genes. Interestingly, the ChiS regulator was only present in the Vibrio strains, suggesting that transcriptional shaping by chitin could be a trait associated with this family. Changes in secondary metabolism by chitin and the presence of the regulator ChiS requires further studies for confirmation. Production of acyl homoserine lactones Production of acyl homoserine lactones Production of acyl homoserine lactone (AHL) com- pounds was analysed using two AHL monitor systems Agrobacterium tumefaciens NT1(pZLR4) [84] and Chro- mobacterium violaceum CV026 [85] as described by Ravn et al. [86]. The strains were grown in 10 mL ½YTSS or sea salt medium (1.5% sea salt, 0.3% casamino acids, 0.4% glucose) in 50 mL Falcon tubes for 48 hours at 200 rpm and room temperature and extracted with 10 mL ethyl acetate acidified with 1% formic acid. The extract was dried under nitrogen, resuspended in 0.5 mL ethyl acetate containing 1% formic acid and stored at −20°C. The extracts were tested with the AHL-reporter strains in a plate well assay [87]. Chitinases and regulation Chitin is ─after cellulose ─the most abundant carbon source on Earth. Enzymes capable of degrading this organic compound are very useful in biotechnological industries. At the same time, chitin is also an important environmental clue influencing regulators of virulence and secondary metabolism [58,76-78]. We have previously shown that an andrimid producing V. coralliilyticus S2052 focuses its secondary metabolism exclusively on andrimid when grown on chitin as compared to growth on glu- cose and casamino acids [58]. This could be coupled with transcriptional changes and we therefore also mined the genomes for chitin catabolic cascade sensor histidine kinase (ChiS) and chitin binding proteins (CBP). genome mining revealed presence of three to nine chitinase encoding genes per strain in the chitinolytic bacteria. Chitin-related genes were present in Vibrio species, which is likely related to their ecology and close association with crustaceans [79,80]. The pigmented pseudoalteromo- nads are also often associated with eukaryotic surfaces [44] including organisms containing no chitin. However, several pseudoalteromonads had genes encoding for chitinases and showed prominent chitinolytic activity. The chitinolytic cascade has previously been studied in Vibrio species where its tight regulation was attributed to the hybrid chitin catabolic sensor/kinase (ChiS) to- gether with a periplasmic chitin oligosaccharide binding protein (CBP) [77]. This regulatory system has been shown to regulate expression of 50 genes, most of which involved in chitin catabolism [77]. Furthermore, it has By phenotypic assays, we identified several strongly chitinolytic strains and screened their genomes for chitinase encoding genes. All of the Vibrionaceae and pigmented Pseudoalteromonas sp., with the exception of P. agarivorans S816, were capable of degrading chitin (Table 3). The Machado et al. BMC Genomics (2015) 16:158 Page 9 of 12 Machado et al. BMC Genomics (2015) 16:158 Genomic Institute (Shenzhen, China). Libraries of 500 bp were used for 100 bp paired-end sequencing of genomes using the Illumina sequencing technology on a HiSeq2000 with a minimum coverage of 100. Genomic DNA se- quences were assembled in contigs using CLC Gen- omic Workbench (CLC Bio, Aarhus, Denmark). All the genomes had a coverage of 75x or higher. All of them were submitted to the National Center for Biotech- nology Information (NCBI) database under the accession numbers AUXW00000000, JMIB00000000, APME000000 00, AQCH00000000, CP002972, CP002973, CP002974, CP002975, JXXR00000000, JXXS00000000, JXXT000 00000, JXXU00000000, JXXV00000000, JXXW00000 000, JXXX00000000, JXXY00000000, JXXZ00000000, JXYA00000000, JXYB00000000, JXYC00000000, JXYD0 0000000, JXYE00000000, JXYF00000000, JXYG00000000. Verification of antibacterial activity Bacterial strains used in this study were isolated during the Danish Galathea 3 global research expedition (http:// www.galathea3.dk/uk) [19] and selected due to their antagonistic activity against a Gram-negative (Vibrio angu- illarum) and a Gram-positive (Staphylococcus aureus) pathogenic bacterium. Pure cultures of each strain were stored in cryoprotectant solution at −80°C from their isolation until the present study. Phaeobacter inhibens DSM17395 was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Germany). Some of the strains have previously been used in classical bioassay-guided bioprospecting and produce antibiotics or anti-virulence compounds [19,22,27-29,40-45,58]. Strains were routinely grown on Marine Agar (Difco 2216) and in Marine Broth (Difco 2216). The strains were re-tested for their antibacterial activity, as previously described [19]. Briefly, strains to be tested were grown in Marine Agar (Difco 2216) for 24 – 48 h and one colony was spotted in plates of artificial seawater agar with 3% Instant Ocean (IO; Aquarium Systems Inc., Sarrebourg, France) containing Vibrio anguillarum strain 90-11-287 serotype O1 [82] or Staphylococcus aureus strain 8325 [83] embedded. The plates were incubated and observed for clearing zones in the agar. Chitinase activity 10. Li B, Walsh CT. Identification of the gene cluster for the dithiolopyrrolone antibiotic holomycin in Streptomyces clavuligerus. Proc Natl Acad Sci U S A. 2010;107:19731–5. 10. Li B, Walsh CT. Identification of the gene cluster for the dithiolopyrrolone antibiotic holomycin in Streptomyces clavuligerus. Proc Natl Acad Sci U S A. 2010;107:19731–5. Chitinase activity was tested on chitin containing agar plates. Strains were grown on Marine Agar (Difco 2216) for 24 – 48 h and one colony was spotted on plates con- taining 20 g/L sea salts, 3 g/L casamino acids, 0.08% hy- drolyzed chitin, 20 g/L agar. The plates were incubated for 72 h and chitinase activity monitored at 24, 48 and 72 h. The natural turbidity of the media due to chitin al- lows the visual evaluation of chitin degradation, which leads to clearance of the media. Chitinase activity was graded qualitatively: low chitinase activity (<1.0 mm) zones were scored with one plus, medium chitinase activity zones (1.0 mm – 3.0 mm) with two pluses, and strong chitinase activity (>3.0 mm) with three plusses. 11. Udwary DW, Zeigler L, Asolkar RN, Singan V, Lapidus A, Fenical W, et al. Genome sequencing reveals complex secondary metabolome in the marine actinomycete Salinispora tropica. Proc Natl Acad Sci U S A. 2007;104:10376–81. 12. Yin H, Xiang S, Zheng J, Fan K, Yu T, Yang X, et al. Induction of holomycin production and complex metabolic changes by the argR mutation in Streptomyces clavuligerus NP1. Appl Environ Microbiol. 2012;78:3431–41. 12. Yin H, Xiang S, Zheng J, Fan K, Yu T, Yang X, et al. Induction of holomycin production and complex metabolic changes by the argR mutation in Streptomyces clavuligerus NP1. Appl Environ Microbiol. 2012;78:3431–41. 13. Li B, Walsh CT. Streptomyces clavuligerus HlmI is an intramolecular disulfide-forming dithiol oxidase in holomycin biosynthesis. Biochemistry. 2011;50:4615–22. 13. Li B, Walsh CT. Streptomyces clavuligerus HlmI is an intramolecular disulfide-forming dithiol oxidase in holomycin biosynthesis. Biochemistry. 2011;50:4615–22. 14. Bhattacharya D, Nagpure A, Gupta RK. Bacterial chitinases: properties and potential. Crit Rev Biotechnol. 2007;27:21–8. 14. Bhattacharya D, Nagpure A, Gupta RK. Bacterial chitinases: properties and potential. Crit Rev Biotechnol. 2007;27:21–8. 15. Ling LL, Schneider T, Peoples AJ, Spoering AL, Engels I, Conlon BP, et al. A new antibiotic kills pathogens without detectable resistance. Nature. 2015;517:455–9. 15. Ling LL, Schneider T, Peoples AJ, Spoering AL, Engels I, Conlon BP, et al. A new antibiotic kills pathogens without detectable resistance. Nature. Genomic DNA isolation and Sequencing Evi Based Complement Alternat Med. 2011;2011:384572. 18. Wietz M, Duncan K, Patin NV, Jensen PR. Antagonistic interactions mediated by marine bacteria: the role of small molecules. J Chem Eco 2013;39:879–91. 19. Gram L, Melchiorsen J, Bruhn JB. Antibacterial activity of marine cultu bacteria collected from a global sampling of ocean surface waters an surface swabs of marine organisms. Mar Biotechnol (NY). 2010;12:439– 20. Fenical W. Chemical studies of marine bacteria: developing a new res Chem Rev. 1993;93:1673–83. 21. Lane AL, Moore BS. A sea of biosynthesis: marine natural products me molecular age. Nat Prod Rep. 2011;28:411–28. 22. Wietz M, Mansson M, Vynne NG, Gram L. Small molecule antibiotics fr marine bacteria and strategies to prevent rediscovery of known compounds. In: Edited by Kim S. Marine microbiology : bioactive compounds and biotechnological applications. Wiley-VCH Verlag GmbH & Co. KGaA; 2013. p. 127-59. 23. Wilson MC, Mori T, Rückert C, Uria AR, Helf MJ, Takada K, et al. An environmental bacterial taxon with a large and distinct metabolic repertoire. Nature. 2014;506:58–62. 24. Beer LL, Moore BS. Biosynthetic convergence of salinosporamides A a in the marine actinomycete Salinispora tropica. Org Lett. 2007;9:845–8 25. Feling R, Buchanan G. Salinosporamide A: a highly cytotoxic proteaso inhibitor from a novel microbial source, a marine bacterium of the ne genus Salinospora. Angew Chemie. 2003;42:355–7. 26. Still PC, Johnson TA, Theodore CM, Loveridge ST, Crews P. Scrutinizing scaffolds of marine biosynthetics from different source organisms: Gram-negative cultured bacterial products enter center stage. J Nat P 2014;77:690–702. 27. Månsson M, Phipps RK, Gram L, Munro MHG, Larsen TO, Nielsen KF. Competing interests The authors declare that they have no competing interests. 20. Fenical W. Chemical studies of marine bacteria: developing a new resource. Chem Rev. 1993;93:1673–83. 20. Fenical W. Chemical studies of marine bacteria: developing a new resource. Chem Rev. 1993;93:1673–83. Abbreviations NRPS N b NRPS: Non-ribosomal peptide synthase; PKS: Polyketide synthase; KS-domains: Keto-synthase domains; C-domains: Condensation domains; RAST: Rapid Annotation using Subsystem Technology; PFAM: Protein families database; ORF: Open reading frame; profile-HMMs: Profile hidden Markov models; AHL: Acyl homoserine lactone; QS: Quorum sensing; HSL: Homoserine lactone; CAS: Chrome azurol S; Fur: Ferric-iron uptake regulator; ChiS: Chitin catabolic cascade sensor histidine kinase; CBP: Chitin binding protein. 17. Zhao X-Q. Genome-based studies of marine microorganisms to maximize the diversity of natural products discovery for medical treatments. Evid Based Complement Alternat Med. 2011;2011:384572. 17. Zhao X-Q. Genome-based studies of marine microorganisms to maximize the diversity of natural products discovery for medical treatments. Evid Based Complement Alternat Med. 2011;2011:384572. 18. Wietz M, Duncan K, Patin NV, Jensen PR. Antagonistic interactions mediated by marine bacteria: the role of small molecules. J Chem Ecol. 2013;39:879–91. 18. Wietz M, Duncan K, Patin NV, Jensen PR. Antagonistic interactions mediated by marine bacteria: the role of small molecules. J Chem Ecol. 2013;39:879–91. 19. Gram L, Melchiorsen J, Bruhn JB. Antibacterial activity of marine culturable bacteria collected from a global sampling of ocean surface waters and surface swabs of marine organisms. Mar Biotechnol (NY). 2010;12:439–51. 19. Gram L, Melchiorsen J, Bruhn JB. Antibacterial activity of marine culturable bacteria collected from a global sampling of ocean surface waters and surface swabs of marine organisms. Mar Biotechnol (NY). 2010;12:439–51. Authors’ contributions Authors’ contributions HM and LG designed the study and HM, ESC and JM carried out the experiments. All authors contributed to the writing of the manuscript. All authors read and approved the final manuscript. 21. Lane AL, Moore BS. A sea of biosynthesis: marine natural products meet the molecular age. Nat Prod Rep. 2011;28:411–28. 21. Lane AL, Moore BS. A sea of biosynthesis: marine natural products meet the molecular age. Nat Prod Rep. 2011;28:411–28. 22. Wietz M, Mansson M, Vynne NG, Gram L. Small molecule antibiotics from marine bacteria and strategies to prevent rediscovery of known compounds. In: Edited by Kim S. Marine microbiology : bioactive compounds and biotechnological applications. Wiley-VCH Verlag GmbH & Co. KGaA; 2013. p. 127-59. Siderophore activity 4. Newman DJ, Cragg GM. Natural products as sources of new drugs over the last 25 years. J Nat Prod. 2007;70:461–77. 4. Newman DJ, Cragg GM. Natural products as sources of new drugs over the last 25 years. J Nat Prod. 2007;70:461–77. Siderophore activity was tested using the liquid CAS assay [69]. The marine strains were grown in 10 mL sea salt medium or ½YTSS in 50 mL Falcon tubes at 25°C and 200 rpm for 24 and 48 hours at room temperature. 1 mL of culture was centrifuged for 5 min at 12,100 × g and the supernatant was mixed with CAS solution in a 1:1 ratio. Colour change from blue to orange indicating siderophore activity was observed after 5 min and 24 h. 5. Moore BS. Biosynthesis of marine natural products: microorganisms (Part A). Nat Prod Rep. 2005;22:580–93. 5. Moore BS. Biosynthesis of marine natural products: microorganisms (Part A). Nat Prod Rep. 2005;22:580–93. 6. Xiong Z-Q, Wang J-F, Hao Y-Y, Wang Y. Recent advances in the discovery and development of marine microbial natural products. Mar Drugs. 2013;11:700–17. 6. Xiong Z-Q, Wang J-F, Hao Y-Y, Wang Y. Recent advances in the discovery and development of marine microbial natural products. Mar Drugs. 2013;11:700–17. 8. Bérdy J. Thoughts and facts about antibiotics: where we are now and where we are heading. J Antibiot. 2012;65:441–41. 8. Bérdy J. Thoughts and facts about antibiotics: where we are now and where we are heading. J Antibiot. 2012;65:441–41. 9. Yu D, Xu F, Valiente J, Wang S, Zhan J. An indigoidine biosynthetic gene cluster from Streptomyces chromofuscus ATCC 49982 contains an unusual IndB homologue. J Ind Microbiol Biotechnol. 2013;40:159–68. 9. Yu D, Xu F, Valiente J, Wang S, Zhan J. An indigoidine biosynthetic gene cluster from Streptomyces chromofuscus ATCC 49982 contains an unusual IndB homologue. J Ind Microbiol Biotechnol. 2013;40:159–68. Chitinase activity 2015;517:455–9. 16. Ziemert N, Lechner A, Wietz M, Millán-Aguiñaga N, Chavarria KL, Jensen PR. Diversity and evolution of secondary metabolism in the marine actinomycete genus Salinispora. Proc Natl Acad Sci U S A. 2014;111:E1130–9. 16. Ziemert N, Lechner A, Wietz M, Millán-Aguiñaga N, Chavarria KL, Jensen PR. Diversity and evolution of secondary metabolism in the marine actinomycete genus Salinispora. Proc Natl Acad Sci U S A. 2014;111:E1130–9. Acknowledgments HM was supported by a PhD grant from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme FP7-People-2012-ITN, under grant agreement No. 317058, “BACTORY”. ECS was supported by the European Union’s Seventh Framework Programmes KBBE-2012-6-311975 MaCuMBA (Marine Microorganisms: Cultivation Methods for Improving their Biotechnological Applications) and KBBE-2012-6-312184 PharmaSea (Increasing Value and Flow in the Marine Biodiscovery Pipeline). The present work was carried out as part of the Galathea 3 expedition under the auspices of the Danish Expedition Foundation. This is Galathea 3 contribution no. p110. 23. Wilson MC, Mori T, Rückert C, Uria AR, Helf MJ, Takada K, et al. An environmental bacterial taxon with a large and distinct metabolic repertoire. Nature. 2014;506:58–62. 24. Beer LL, Moore BS. Biosynthetic convergence of salinosporamides A and B in the marine actinomycete Salinispora tropica. Org Lett. 2007;9:845–8. 25. Feling R, Buchanan G. Salinosporamide A: a highly cytotoxic proteasome inhibitor from a novel microbial source, a marine bacterium of the new genus Salinospora. Angew Chemie. 2003;42:355–7. 26. Still PC, Johnson TA, Theodore CM, Loveridge ST, Crews P. Scrutinizing the scaffolds of marine biosynthetics from different source organisms: Gram-negative cultured bacterial products enter center stage. J Nat Prod. 2014;77:690–702. Received: 16 January 2015 Accepted: 20 February 2015 Genomic DNA isolation and Sequencing High purity genomic DNA was extracted by succes- sive phenol:chloroform:isoamyl-alcohol purification steps followed by precipitation with isopropanol, treatment with RNase and a final purification and precipitation step [81]. Quantification was done in 1% agarose gel electrophoresis, NanoDrop Spectrometer (Saveen Werner, Sweden) and Qubit 2.0 Analyser (Invitrogen, United Kingdom). Se- quencing of the genomes was performed by Beijing Machado et al. BMC Genomics (2015) 16:158 Page 10 of 12 4. Newman DJ, Cragg GM. Natural products as sources of new drugs ov last 25 years. J Nat Prod. 2007;70:461–77. 5. Moore BS. Biosynthesis of marine natural products: microorganisms (P Nat Prod Rep. 2005;22:580–93. 6. Xiong Z-Q, Wang J-F, Hao Y-Y, Wang Y. Recent advances in the discov and development of marine microbial natural products. Mar Drugs. 2013;11:700–17. 7. Meier JL, Burkart MD. The chemical biology of modular biosynthetic enzymes. Chem Soc Rev. 2009;38:2012–45. 8. Bérdy J. Thoughts and facts about antibiotics: where we are now and where we are heading. J Antibiot. 2012;65:441–41. 9. Yu D, Xu F, Valiente J, Wang S, Zhan J. An indigoidine biosynthetic ge cluster from Streptomyces chromofuscus ATCC 49982 contains an unus IndB homologue. J Ind Microbiol Biotechnol. 2013;40:159–68. 10. Li B, Walsh CT. Identification of the gene cluster for the dithiolopyrrol antibiotic holomycin in Streptomyces clavuligerus. Proc Natl Acad Sci U 2010;107:19731–5. 11. Udwary DW, Zeigler L, Asolkar RN, Singan V, Lapidus A, Fenical W, et a Genome sequencing reveals complex secondary metabolome in the marine actinomycete Salinispora tropica. Proc Natl Acad Sci U S A. 2007;104:10376–81. 12. Yin H, Xiang S, Zheng J, Fan K, Yu T, Yang X, et al. Induction of holom production and complex metabolic changes by the argR mutation in Streptomyces clavuligerus NP1. Appl Environ Microbiol. 2012;78:3431–4 13. Li B, Walsh CT. Streptomyces clavuligerus HlmI is an intramolecular disulfide-forming dithiol oxidase in holomycin biosynthesis. Biochemis 2011;50:4615–22. 14. Bhattacharya D, Nagpure A, Gupta RK. Bacterial chitinases: properties a potential. Crit Rev Biotechnol. 2007;27:21–8. 15. Ling LL, Schneider T, Peoples AJ, Spoering AL, Engels I, Conlon BP, et new antibiotic kills pathogens without detectable resistance. Nature. 2015;517:455–9. 16. Ziemert N, Lechner A, Wietz M, Millán-Aguiñaga N, Chavarria KL, Jense Diversity and evolution of secondary metabolism in the marine actinomycete genus Salinispora. Proc Natl Acad Sci U S A. 2014;111:E1 17. Zhao X-Q. Genome-based studies of marine microorganisms to maxim the diversity of natural products discovery for medical treatments. References 27. Månsson M, Phipps RK, Gram L, Munro MHG, Larsen TO, Nielsen KF. Explorative Solid-Phase Extraction (E-SPE) for accelerated microbial natural product discovery, dereplication, and purification. J Nat Prod. 2010;73:1126–32. 28. Mansson M, Gram L, Larsen TO. Production of bioactive secondary metabolites by marine vibrionaceae. Mar Drugs. 2011;9:1440–68. 28. Mansson M, Gram L, Larsen TO. Production of bioactive secondary metabolites by marine vibrionaceae. Mar Drugs. 2011;9:1440–68. Page 11 of 12 Machado et al. BMC Genomics (2015) 16:158 deducing the ability of producing secondary metabolites. Proc Natl Acad Sci U S A. 2001;98:12215–20. 29. Wietz M, Mansson M, Gotfredsen CH, Larsen TO, Gram L. Antibacterial compounds from marine Vibrionaceae isolated on a global expedition. Mar Drugs. 2010;8:2946–60. deducing the ability of producing secondary metabolites. Proc Natl Acad Sci U S A. 2001;98:12215–20. 51. Bentley SD, Chater KF, Cerdeño-Tárraga A-M, Challis GL, Thomson NR, James KD, et al. Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature. 2002;417:141–7. 30. Weber T. In silico tools for the analysis of antibiotic biosynthetic pathways. Int J Med Microbiol. 2014;304:230–5. Streptomyces coelicolor A3(2). Nature. 2002;417:141–7. 52. Buchan A, González J, Moran M. Overview of the marine roseobacter lineage. Appl Environ. 2005;71:5665–77. 31. Fedorova ND, Moktali V, Medema MH. Bioinformatics approaches and software for detection of secondary metabolic gene clusters. In: Keller NP, Turner G, editors. Fungal secondary metabolism: methods and protocols. 944th ed. Totowa, NJ: Humana Press; 2012. p. 23–45. 53. Wagner-Döbler I, Biebl H. Environmental biology of the marine Roseobacter lineage. Annu Rev Microbiol. 2006;60:255–80. 32. Seyedsayamdost MR. High-throughput platform for the discovery of elicitors of silent bacterial gene clusters. Proc Natl Acad Sci U S A. 2014;111:7266–71. 54. Porsby CH, Nielsen KF, Gram L. Phaeobacter and Ruegeria species of the Roseobacter clade colonize separate niches in a Danish Turbot (Scophthalmus maximus)-rearing farm and antagonize Vibrio anguillarum under different growth conditions. Appl Environ Microbiol. 2008;74:7356–64. 33. Weber T, Rausch C, Lopez P, Hoof I, Gaykova V, Huson DH, et al. CLUSEAN: a computer-based framework for the automated analysis of bacterial secondary metabolite biosynthetic gene clusters. J Biotechnol. 2009;140:13–7. 55. Seyedsayamdost MR, Carr G, Kolter R, Clardy J. Roseobacticides: small molecule modulators of an algal-bacterial symbiosis. J Am Chem Soc. 2011;133:18343–9. 34. Rausch C, Weber T, Kohlbacher O, Wohlleben W, Huson DH. Specificity prediction of adenylation domains in Nonribosomal Peptide Synthetases (NRPS) using Transductive Support Vector Machines (TSVMs). Nucleic Acids Res. 2005;33:5799–808. 56. References Brinkhoff T, Bach G, Heidorn T, Liang L, Schlingloff A, Simon M. Antibiotic Production by a Roseobacter Clade-Affiliated Species from the German Wadden Sea and Its Antagonistic Effects on Indigenous Isolates. Appl Environ Microbiol. 2004;70:2560–5. 35. Röttig M, Medema MH, Blin K, Weber T, Rausch C, Kohlbacher O. NRPSpredictor2–a web server for predicting NRPS adenylation domain specificity. Nucleic Acids Res. 2011;39(Web Server issue):W362–7. 57. Geng H, Bruhn JB, Nielsen KF, Gram L, Belas R. Genetic dissection of tropodithietic acid biosynthesis by marine roseobacters. Appl Environ Microbiol. 2008;74:1535–45. y 36. Minowa Y, Araki M, Kanehisa M. Comprehensive analysis of distinctive polyketide and nonribosomal peptide structural motifs encoded in microbial genomes. J Mol Biol. 2007;368:1500–17. 58. Wietz M, Månsson M, Gram L. Chitin stimulates production of the antibiotic andrimid in a Vibrio coralliilyticus strain. Environ Microbiol Rep. 2011;3:559–64. 37. Van Heel AJ, de Jong A, Montalbán-López M, Kok J, Kuipers OP. BAGEL3: automated identification of genes encoding bacteriocins and (non-) bactericidal posttranslationally modified peptides. Nucleic Acids Res. 2013;41(Web Server issue):W448–53. 59. Jin M, Fischbach MA, Clardy J. A biosynthetic gene cluster for the acetyl-CoA carboxylase inhibitor andrimid. J Am Chem Soc. 2006;128:10660–1 60. Gauthier MJ. Morphological, physiological, and biochemical characteristics of some violet-pigmented bacteria isolated from seawater. Can J Microbiol. 1976;22:138–49. 38. Ziemert N, Podell S, Penn K, Badger JH, Allen E, Jensen PR. The natural product domain seeker NaPDoS: a phylogeny based bioinformatic tool to classify secondary metabolite gene diversity. PLoS One. 2012;7:e34064. 61. Laatsch H, Pudleiner H. Marine bakterien, I: synthese von pentabrompseudilin, einem phenylpyrrol aus Alteromonas luteoviolaceus. Liebigs Ann Chem. 1989;9:863–81. 39. Li MHT, Ung PMU, Zajkowski J, Garneau-Tsodikova S, Sherman DH. Automated genome mining for natural products. BMC Bioinformatics. 2009;10:185. 62. Moore BS. Biosynthesis of marine natural products: macroorganisms (Part B) Nat Prod Rep. 2006;23:615–29. 62. Moore BS. Biosynthesis of marine natural products: macroorganisms (Part B). Nat Prod Rep. 2006;23:615–29. 40. Mansson M, Nielsen A, Kjærulff L, Gotfredsen CH, Wietz M, Ingmer H, et al. Inhibition of virulence gene expression in Staphylococcus aureus by novel depsipeptides from a marine Photobacterium. Mar Drugs. 2011;9:2537–52. 63. Conway KR, Boddy CN. ClusterMine360: a database of microbial PKS/NRPS biosynthesis. Nucleic Acids Res. 2013;41(Database issue):D402–7. 41. Nielsen A, Månsson M, Bojer MS, Gram L, Larsen TO, Novick RP, et al. Solonamide B inhibits quorum sensing and reduces Staphylococcus aureus mediated killing of human neutrophils. PLoS One. 2014;9:e84992. 64. References Bruhn JB, Nielsen KF, Hjelm M, Hansen M, Bresciani J, Schulz S, et al. Ecology, inhibitory activity, and morphogenesis of a marine antagonistic bacterium belonging to the Roseobacter clade. Appl Environ Microbiol. 2005;71:7263–70. 42. Kjaerulff L, Nielsen A, Mansson M, Gram L, Larsen TO, Ingmer H, et al. Identification of four new agr quorum sensing-interfering cyclodepsipeptides from a marine Photobacterium. Mar Drugs. 2013;11:5051–62. 65. Holden MTG, Chhabra SR, De Nys R, Stead P, Bainton NJ, Hill PJ, et al. Quorum-sensing cross talk: isolation and chemical characterization of cyclic dipeptides from Pseudomonas aeruginosa and other gram-negative bacteria. Mol Microbiol. 1999;33:1254–66. 43. Nielsen A, Mansson M, Wietz M, Varming AN, Phipps RK, Larsen TO, et al. Nigribactin, a novel siderophore from Vibrio nigripulchritudo, modulates Staphylococcus aureus virulence gene expression. Mar Drugs. 2012;10:2584–95. 66. Rasmussen BB, Nielsen KF, Machado H, Melchiorsen J, Gram L, Sonnenschein EC. Global and phylogenetic distribution of quorum sensing signals, acyl homoserine lactones, in the family of vibrionaceae. Mar Drugs. 2014;12:5527–46. 44. Vynne NG, Månsson M, Nielsen KF, Gram L. Bioactivity, chemical profiling, and 16S rRNA-based phylogeny of Pseudoalteromonas strains collected on a global research cruise. Mar Biotechnol (NY). 2011;13:1062–73. 67. Vynne NG. Bioactivity and phylogeny of the marine bacterial genus Pseudoalteromonas. Kgs. Lyngby: Technical University of Denmark; 2011. 68. Sandy M, Butler A. Microbial iron acquisition: marine and terrestrial siderophores. Chem Rev. 2009;109:4580–95. 45. Vynne NG, Mansson M, Gram L. Gene sequence based clustering assists in dereplication of Pseudoalteromonas luteoviolacea strains with identical inhibitory activity and antibiotic production. Mar Drugs. 2012;10:1729–40. 69. Schwyn B, Neilands JB. Universal chemical assay for the detection and determination of siderophores. Anal Biochem. 1987;160:47–56. 46. Medema MH, Blin K, Cimermancic P, de Jager V, Zakrzewski P, Fischbach MA, et al. antiSMASH: rapid identification, annotation and analysis of secondary metabolite biosynthesis gene clusters in bacterial and fungal genome sequences. Nucleic Acids Res. 2011;39(Web Server issue):W339–46. 70. Hider RC, Kong X. Chemistry and biology of siderophores. Nat Prod Rep. 2010;27:637–57. 71. Rudolph G, Hennecke H, Fischer H-M. Beyond the Fur paradigm: iron-controlled gene expression in rhizobia. FEMS Microbiol Rev. 2006;30:631–48. 47. Blin K, Medema MH, Kazempour D, Fischbach MA, Breitling R, Takano E, et al. antiSMASH 2.0–a versatile platform for genome mining of secondary metabolite producers. Nucleic Acids Res. 2013;41 (Web Server issue):W204–12. 72. Williams KP, Gillespie JJ, Sobral BWS, Nordberg EK, Snyder EE, Shallom JM, et al. Phylogeny of gammaproteobacteria. J Bacteriol. 2010;192:2305–14. 48. Machado et al. BMC Genomics (2015) 16:158 75. Thompson FL, Gevers D, Thompson CC, Dawyndt P, Naser S, Hoste B, et al. Phylogeny and molecular identification of vibrios on the basis of multilocus sequence analysis. Appl Environ Microbiol. 2005;71:5107–15. 76. Meibom KL, Li XB, Nielsen AT, Wu C-Y, Roseman S, Schoolnik GK. The Vibrio cholerae chitin utilization program. Proc Natl Acad Sci U S A. 2004;101:2524–9. 76. Meibom KL, Li XB, Nielsen AT, Wu C-Y, Roseman S, Schoolnik 77. Li X, Roseman S. The chitinolytic cascade in Vibrios is regulated by chitin oligosaccharides and a two-component chitin catabolic sensor/kinase. Proc Natl Acad Sci U S A. 2004;101:627–31. 78. Frederiksen RF, Paspaliari DK, Larsen T, Storgaard BG, Larsen MH, Ingmer H, et al. Bacterial chitinases and chitin-binding proteins as virulence factors. Microbiology. 2013;159(Pt 5):833–47. 79. Souza CP, Almeida BC, Colwell RR, Rivera ING. The importance of chitin in the marine environment. Mar Biotechnol (NY). 2011;13:823–30. 80. Stauder M, Huq A, Pezzati E, Grim CJ, Ramoino P, Pane L, et al. Role of GbpA protein, an important virulence-related colonization factor, for Vibrio cholerae’s survival in the aquatic environment. Environ Microbiol Rep. 2012;4:439–45. 81. Sambrook J, Russel DW. Molecular cloning: a laboratory manual. Cold Spring Harbour, New York: Cold Spring Harbour Laboratory Press; 2001. 82. Skov MN, Pedersen K, Larsen JL. Comparison of pulsed-field gel electrophoresis, ribotyping, and plasmid profiling for typing of Vibrio anguillarum serovar O1. Appl Environ Microbiol. 1995;61:1540–5. 83. Novick R, Morse S. In vivo transmission of drug resistance factors between strains of Staphylococcus aureus. J Exp Med. 1967;125(1):45–59. 84. Cha C, Gao P, Chen YC, Shaw PD, Farrand SK. Production of acyl-homoserine lactone quorum-sensing signals by gram-negative plant-associated bacteria. Mol Plant Microbe Interact. 1998;11:1119–29. 85. Mcclean KH, Winson MK, Fish L, Taylor A, Chhabra SR, Camara M, et al. Quorum sensing and Chromobacterium violaceum: exploitation of violacein production and inhibition for the detection of N -acylhomoserine lactones. 1997. p. 3703–11. 86. Ravn L, Christensen AB, Molin S, Givskov M, Gram L. Methods for detecting acylated homoserine lactones produced by Gram-negative bacteria and their application in studies of AHL-production kinetics. J Microbiol Methods. 2001;44:239–51. 87. Gram L, Grossart H. Possible quorum sensing in marine snow bacteria: production of acylated homoserine lactones by Roseobacter strains isolated from marine snow. Appl. 2002;68:4111–6. References Overbeek R, Olson R, Pusch GD, Olsen GJ, Davis JJ, Disz T, et al. The SEED and the rapid annotation of microbial genomes using Subsystems Technology (RAST). Nucleic Acids Res. 2014;42(Database issue):D206–14. 73. Sawabe T, Ogura Y, Matsumura Y, Feng G, Amin AR, Mino S, et al. Updating the Vibrio clades defined by multilocus sequence phylogeny: proposal of eight new clades, and the description of Vibrio tritonius sp. nov. Front Microbiol. 2013;4(December):414. 49. Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, Edwards RA, et al. The RAST Server: rapid annotations using subsystems technology. BMC Genomics. 2008;9:75. 74. Pascual J, Macián MC, Arahal DR, Garay E, Pujalte MJ. Multilocus sequence analysis of the central clade of the genus Vibrio by using the 16S rRNA, recA, pyrH, rpoD, gyrB, rctB and toxR genes. Int J Syst Evol Microbiol. 2010;60(Pt 1):154–65. 50. Omura S, Ikeda H, Ishikawa J, Hanamoto A, Takahashi C, Shinose M, et al. Genome sequence of an industrial microorganism Streptomyces avermitilis: Page 12 of 12 Page 12 of 12 Machado et al. BMC Genomics (2015) 16:158 Machado et al. BMC Genomics (2015) 16:158 Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit • Convenient online submission
W4306385480.txt	https://www.frontiersin.org/articles/10.3389/fenvs.2022.1036464/pdf	en		Does e-commerce participation affect green agrotechnology adoption among reservoir resettlers? The case of China’s Three Gorges Reservoir area	Frontiers in environmental science	2,022	cc-by	9,703	Original Research 17 October 2022 10.3389/fenvs.2022.1036464 TYPE PUBLISHED DOI OPEN ACCESS EDITED BY Aaron Kinyu Hoshide, University of Maine, United States REVIEWED BY Muhammad Asad Ur Rehman Naseer, Bahauddin Zakariya University, Pakistan Atique ur Rehman, Bahauddin Zakariya University, Pakistan Muhammad Rizwan, Yangtze University, China CORRESPONDENCE Xu Zhao, zhaoxu@ctgu.edu.cn Does e-commerce participation affect green agrotechnology adoption among reservoir resettlers? The case of China’s Three Gorges Reservoir area Xu Zhao 1,2, Zhuo Cui 2 and Feifei Zhao 2 1 Research Center for Reservoir Resettled, China Three Gorges University, Yichang, China, 2College of Economics and Management, China Three Gorges University, Yichang, China SPECIALTY SECTION This article was submitted to Environmental Economics and Management, a section of the journal Frontiers in Environmental Science 04 September 2022 05 October 2022 PUBLISHED 17 October 2022 RECEIVED ACCEPTED CITATION Zhao X, Cui Z and Zhao F (2022), Does e-commerce participation affect green agrotechnology adoption among reservoir resettlers? The case of China’s Three Gorges Reservoir area. Front. Environ. Sci. 10:1036464. doi: 10.3389/fenvs.2022.1036464 COPYRIGHT © 2022 Zhao, Cui and Zhao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. This study explores how promoting e-commerce participation impacts the adoption of green agrotechnology by resettlers in China’s Three Gorges Reservoir area and helps rural revitalization and the realization of value from ecological produce. First, we combine induced innovation model theory with the risk perception factor of expected utility theory. A model of resettlers’ green agrotechnology adoption under different levels of e-commerce participation is constructed, and research hypotheses are proposed accordingly. Survey data gathered from resettled farmers in Zigui, the ﬁrst county of the studied area, are tested empirically with an ordered probit model. The results show ﬁrst, that e-commerce participation signiﬁcantly and positively affects the level of green agrotechnology adoption at the 1% level; and second, that expectations of the ecological value of agricultural products and the agrotechnology support provided by e-commerce are important driving factors. The promotion effect of different modes of e-commerce participation on agrotechnology adoption differ. The risk-averse behavior of resettlers can weaken the promotion effect of e-commerce participation on agrotechnology adoption. KEYWORDS e-commerce participation, green agrotechnology adoption, risk perception, reservoir resettlers, ordered-probit model 1 Introduction The increasing number of hydropower plants being built around the world to achieve clean energy has led to a large number of compulsory population movements, resulting in the emergence of resettled farmers as a group. In accordance with the World Bank’s resettled guidelines, rural resettlers in China are mostly resettled in the back-to-back mode of “returning land to land and agriculture to agriculture” for the sake of continuing their original livelihoods (Yan et al., 2018). However, due to the limited resource-carrying capacity of the Three Gorges Reservoir area, the arable land available there has become increasingly scarce. Most of the compensation land received by resettlers is fragmented and low in quality, causing not only limited output but also signiﬁcant increases in the cost Frontiers in Environmental Science 01 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 government subsidies, which can be used to guide technology application behavior in agricultural production. Reducing the costs involved in such production will support and guarantee the sustainable application of green agrotechnology (Dong et al., 2022). Second, research has explored the contribution of regional resource endowments, where the level of development, farming history, and soil quality are direct factors determining individual adoption intentions and behaviors (Zeng et al., 2019; Wu and Zhou 2021), while population aging and the stock of agricultural technicians have obvious indirect effects (Natkhov and Vasilenok 2021). A third stream, focusing on the role of rural households and individual characteristics, has found that the degree of green agrotechnology adoption is constrained by farmers’ relational capital (Wang et al., 2020), socioeconomic status (Bidogeza et al., 2009), and learning and training opportunities (Chatzimichael et al., 2014). In contrast, risk preferences play a dominant role in individual production decision behavior (Gao and Niu 2019), and risk perceptions, in turn, positively moderate the degree of association between risk preferences and agrotechnology adoption (Qiu et al., 2020). For example, perceived risk from agricultural production affects the use of irrigation technology (Koundouri et al., 2006), fertilizers (Adnan et al., 2019), etc. Since it is difﬁcult for individuals to fully understand the beneﬁts associated with the use of green agrotechnology, they tend to have doubts about the risks they need to take in technology adoption (Chavas and Nauges 2020). However, further studies have shown that farmers may use subjective risk judgements to weigh the pros and cons; that is, there is a negative relationship between the perceived risk of using green agrotechnology and the probability of adoption (Duong et al., 2019). The research outlined above is limited by the fact that the paths between variables have only been tested empirically, while speciﬁc inﬂuence mechanisms have not yet been theoretically deduced. Hence, such mechanisms are based only on the summary of phenomena and experiences and lack a scientiﬁc basis. Further exploration is required of such mechanisms and the paths whereby the interactive feedback model of e-commerce operation–agrotechnology application–income enhancement can be realized. The Chinese government has implemented appropriate support policies for Reservoir resettlers, and the livelihood level and social integration of this group have been signiﬁcantly improved. Ensuring the sustainable selfdevelopment of involuntary project resettlers has attracted research attention globally (Karimi and Taifur 2013). The current research was conducted as follows. First, we constructed a model of the beneﬁts of green agrotechnology adoption by resettled households under different e-commerce participation scenarios, based on utility optimization theory and induced agricultural technological innovation theory. Second, we introduced risk perception as a moderating variable and combined it with expected utility theory to of crop cultivation (Zhang 2021). Chinese rural re-settlers’ per capita income was only about 76% of that of ordinary farmers in 2014. Therefore, fewer than 30% of resettlers actually cultivate the land that the government transferred to them at a high cost, and many tend to abandon the land and go out to work instead (Ni and Shao 2013). Currently, in the context of natural capital stock constraint, the only way to guarantee that resettlers can increase their income from local farming is by applying and promoting green agrotechnology. Green agricultural technology aims to solve environmental problems and promote sustainable agricultural development. It will improve the soil by reducing pollution, leverage the ecological value of agricultural products, and raise crop yields (Zuo and Fu 2021). The waste of resources and environmental pollution caused by the unscientiﬁc nature of traditional agricultural technology hinders the green development of agriculture in resettlement areas. Achieving a green transformation of agricultural development is an important way to solve the current agricultural development challenges in resettlement areas. As resettlers’ livelihoods are highly vulnerable and they struggle to withstand the ex-post effects of risks, they mainly choose to avoid risk. Their conservative behavior in agricultural production hinders the adoption of green agrotechnology to a certain extent. Therefore, a long-term mechanism of green agrotechnology adoption is needed to solve the current problem of low level of green agrotechnology in resettlement areas and to achieve sustainable development of resettlement areas and individual income of resettlers. For involuntary resettlers with impaired livelihood capital and broken social networks of origin, rural e-commerce, a new rural industry with convenient participation channels and an almost zero threshold for resource consumption, has become the ﬁrst choice to improve the production and marketing of agricultural products (Yin and Choi 2022). For example, in the Three Gorges Reservoir area of China, the town of Xingshan is located in a mountainous area with a low land stock, but resettlers have achieved an average daily sale of 100,000 kg of citrus by using live streaming. The strong development of Internet technology has enriched agricultural e-commerce models, with traditional e-commerce, social e-commerce, and live e-commerce all expanding in scale (Chen et al., 2022). Heterogeneous participation in rural e-commerce brings innovation to the traditional agricultural products business model and concept while effectively enhancing the income of resettled households. At the same time, the online sales model requires quality standardization and ecological branding, which will certainly promote the application of green agrotechnology and bring high-quality change to resettled areas (Xiao et al., 2021). Current research on empowering rural agents to promote the adoption of green agrotechnology focuses on three aspects. The ﬁrst is the promotion of environmental regulations and Frontiers in Environmental Science 02 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 the post-resettled production and living standards of resettlers should not be lower than those before relocation. Therefore, whether resettled farmers can promote the application of green agrotechnology in agricultural production after participating in rural e-commerce is mainly determined by the related proﬁt, so it can be judged according to the proﬁt function corresponding to different participation situations of e-commerce, as shown in Eq. 1. improve the multi-factor inﬂuence mechanism model of resettled technology adoption. Third, we proposed research hypotheses according to the theoretical derivation and empirically tested the theoretical model using the ordered probit model with survey data gathered from resettled farmers in Zigui, Hubei province, China, the ﬁrst county in the Three Gorges Reservoir area. Next, we explored how the promotion of rural e-commerce participation affects the adoption of green agrotechnology by resettlers and the core elements within this process and investigated the perturbing inﬂuence of the particular risk preferences of involuntarily relocated populations in regard to promoting modern agricultural technology and achieving sustainable livelihood development. Last, we provide support for decision making in regard to compensating and supporting resettlers in water conservancy and hydropower projects. π i TR(Q) − TC(Q) where π i is the proﬁt of resettlers selling household agricultural products; i 0 means through an online e-commerce channel, and i 1 means along the traditional ofﬂine channel. Q is the agricultural production level of resettled households in the resettled area, TR(Q) is the total income derived from agricultural cultivation, and TC(Q) is the corresponding total expenditure. Here, it is assumed that the production level Q of resettled households can be expressed by Eq. 2. 2 Theories and hypothesis Q fIg, g In the theory of agricultural supply chain integration, in response to the impact of the agricultural business segment on the production segment, induced agricultural technological innovation theory was proposed at the start of the 20th century. This theory argues that the business model is closely related to agricultural technological innovation (Cowan et al., 2015) and that a proposed increase, decrease, or change of production and business factors will necessarily bring about technological innovation (Schultz 1987). Accordingly, it is clear that the promotion of Internet technologies and the development of rural e-commerce will inevitably promote changes in production by farmers, namely the use of green technologies in agriculture. On the one hand, the combination of e-commerce and the digital economy completes the construction of a digital system for the agricultural industry which is more conducive to the value-added and market demand transfer of green agricultural products and allows for more convenient mining and teaching of green agricultural technologies (Fu and Zhang 2022). On the other hand, the expansion of agricultural product sales by e-commerce and logistics platforms has led to a surge in the number of end customers with green consumption needs. The trend of standardization and branding of agricultural products has led to increasingly stringent requirements for the production process, which has also forced the improvement and standardization of the use of resettled agricultural technologies (Dong et al., 2021) as shown in Figure 1. In the process of compulsory relocation, Reservoir resettlers lose their livelihood capital, livelihood capacity, and social network. Thus, the focus of subsequent development has been on how to achieve maximum beneﬁts under resource constraints (Zhang 2021). It is a requirement of the principle of developmental resettled for hydropower projects in China that Frontiers in Environmental Science (1) (2) In the above equation, g is the degree of green agrotechnology adoption by resettled households, and I is other agricultural production factor inputs. The degree of households’ inputs is closely related to the level of agrotechnology they use, so the latter can be expressed in the functional form I(g). Assuming that the price of agricultural production factor I is P0 and the unit price of green agrotechnology input is P1 , the proﬁt of resettled agricultural output can be shown as in Eq. 3. π i TR(Q) − TC(Q) Pi Q − P0 Ig − P1 g (3) At this point, it is necessary to ﬁnd the optimal level of adoption of green agrotechnology for resettlers under the proﬁt maximization condition, which is the derivative of proﬁt πi to the degree of adoption of agrotechnology g as shown in Eq. 4. dπi zQ dI dQ dI + − P1 0 Pi · · − P0 · zI dg dg dg dg (4) In the above equation, Pi · dQ/dg can be regarded as the marginal beneﬁt of green agrotechnology adoption by resettlers, denoted as MRi , while P0 · dl/dg + P1 is the marginal cost of agrotechnology adoption, denoted as MCi . While the network direct sales model allows customers to sell their products directly, it also allows customers to obtain ecological and high-quality agricultural products more conveniently (Tian et al., 2022). Then it feeds this demand to resettled households quickly, which becomes a source of motivation for this group to apply green agricultural technology. For example, in the resettled area of Guojiaba Township, Zigui County in the Three Gorges Reservoir area, a water and fertilizer integrated navel orange planting base was built with the help of an e-commerce platform, and the track from the orchard to the road was electriﬁed and a full production cycle traceability system with an integrated QR code was designed. 03 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 FIGURE 1 Analytical framework based on induced agricultural technological innovation theory. it can be sold at 0.62 USD/kilogram (kg), an increase of more than 20%. So, when P1 > P0 , there is g1 > g0 , and the following research hypothesis can be proposed. Hypothesis 1. Participation in e-commerce has a catalytic effect on resettled growers’ green agrotechnology adoption behavior. Different e-commerce participation models will have different beneﬁts for resettled households as the participating subjects and circulation links vary. This paper classiﬁes resettled households’ e-commerce participation model as either platform e-commerce or social e-commerce, according to the survey data. The platform e-commerce model refers to resettled households selling through online trading platforms, such as Taobao and Jindong. The social e-commerce model refers to such households selling through a network of acquaintances to form a ﬁxed source of online customers, such as through WeChat or QQ. The platform e-commerce model may obtain higher product revenue as a large number of merchants are participating: While the platform products are highly competitive barriers to entry, the requirements for technical investment are also higher. The social e-commerce model has price advantages, but resettled households have a limited network group of acquaintances, and the fact that e-commerce has a limited effect on increasing income leads to less willingness to adopt technical innovation and lower rates of green agrotechnology adoption. Accordingly, the following research hypothesis can be proposed. FIGURE 2 Linkages between e-commerce participation and resettlers’ green agrotechnology adoption. The results calculated from Eq. 4 is shown in Figure 2. In the case of resettled households participating in e-commerce, when MC1 MR1 , the optimal green agrotechnology adoption level is g1 , and when resettled households do not participate in e-commerce, the optimal value of green agrotechnology adoption at this point can be obtained from MC0 MR0 as g0 . The current participation of resettlers in e-commerce will signiﬁcantly increase the sales price of agricultural products. For example, the traditional channel sale price of pomegranate in Yunnan reservoir area of the lower Jinsha River hydropower station in 2021 was 0.5 USD/ kilogram (kg). While through e-commerce do boutique retailing, Frontiers in Environmental Science Hypothesis 2. There are differences in the promotion of green agrotechnology adoption among resettled households according to various e-commerce participation models, with the platform e-commerce model outperforming the social e-commerce model. 04 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 1 E[U(W + ε)] EU(W) + U′(W)ε + U″(W)ε2 + R 2 1 ≈ U(W) + U′(W)E(ε) + U″(W)Eε2 2 However, a single proﬁt-seeking factor is not sufﬁcient in explaining the motivation of resettled households when they adopt green agrotechnology (Adnan et al., 2021). The increased vulnerability of livelihood and the relatively scarce livelihood capital led to a weakening in the ability of involuntary resettled groups to tackle risks to their livelihoods (Gong et al., 2020). However, the application of green agrotechnology may also give rise to additional risks, including technological, natural, and market risks. For example, compared with chemical pesticides, organic pesticides are difﬁcult to operate and highly targeted. Resettlers faced with planting new crops in resettled areas may be vulnerable to yield reduction after application of a new technology. Especially if they are not familiar with pesticide selection, application dosage, and application time (Fang et al., 2021). In addition, resettlers whose original social networks are a great distance away from the resettled area and who have poor information channels may face failure to achieve high quality and high prices for their agricultural products, and with high technology costs (Lu et al., 2018). Therefore, resettlers have a lot of uncertainty when adopting green agrotechnology, and thus perceive the existence of risks, and in view of their relatively weak risk tolerance, they are mostly risk-averse subjects. In other words, the greater levels of risk perception in the resettlers’ technology adoption behavior will directly affect the role played by e-commerce in the promotion of such behavior. From the expectation-utility theory, the utility function of resettlers’ participation in rural e-commerce is shown in Eq. 5. U(W − e) E[U(W + ε)] Since resettlers are mostly risk-averse after experiencing loss of livelihood, R is basically negligible and has E(ε) 0, therefore at this point, E(ε2 ) Var(ε). According to Eqs 5, 7, 8 are equivalent by association, which leads to Eq. 9. 1 U(W) + U″(W)Var(ε) U(W) − U′(W)e 2 e− 1 U″(W) 1 Var(ε) k(e)Var(ε) 2 U′(W) 2 (10) It can be calculated that U″(W)/U′(W) k(e) in the above equation, while Var(ε) represents the external factors affecting resettlers’ returns. This is the variance of random returns, which can be considered as being the perceived income risk held by resettled households. In the case that resettled households tend towards risk-aversion, k(e) is relatively stable. If the degree of resettled risk perception Var(ε) is higher, the value of e increases, indicating that resettlers are more resistant to green farming techniques. According to the above analysis, the following research hypothesis can be proposed: (5) Hypothesis 3. Individual risk perceptions will constrain the degree to which e-commerce participation promotes the adoption of green agrotechnology by resettled households. 3 Data, models and variables 3.1 Location selection and data sources The Three Gorges Reservoir has ﬂooded 260 km2 of arable land, and there are 354,000 rural resettlers settled in the area. This has resulted in an extremely limited environmental capacity to produce food. Zigui County in Hubei, at the head of the Three Gorges Project Reservoir, is both a resettled area and classed as a “national poverty-stricken county,” with 25% of its total population being Reservoir resettlers. As the climate and soil environment are very suitable for the growth of Navel oranges, Zigui County has become famous as China’s “Navel Orange Township.” Therefore, in recent years, through the Three Gorges resettled support funds and other promotional projects, Zigui County has been encouraged to adopt green agriculture. At present, the Zigui navel orange production area covers 23,200 ha and has an annual output of 605,000 metric tons. Across the area’s 12 towns and 116 villages, about 198,000 people 1 EU(W) + U′(W)ε + U″(W)ε2 + R U(W) + U′(W)e + R 2 (6) R in the above equation is the higher order remainder term of Eq. 6, and the process of transforming the resettled participation in the electric utility U in Eq. 5 by the equation is shown in Eq. 7. (7) Similarly, an equation transformation of the resettled expected utility function E in Eq. 5 is shown in Eq. 8. Frontiers in Environmental Science (9) An equation transformation of Eq. 9 can locate the willingness to resist green agrotechnology e of e-commerce resettled households, as shown in Eq. 10. Where U and E are the utility and expected utility functions of resettled households, respectively, W is the production and marketing input, ε is the stochastic return, and e is the risk premium, which indicates the degree of resistance to green agrotechnology among resettled e-commerce participants. In order to analyze whether risk perception affects the propulsive effect of e-commerce on agrotechnology adoption behavior, the relationship between resettled risk perception and risk preference e must ﬁrst be clariﬁed. It is assumed that the beneﬁt utility U(W) of resettled e-merchants is second-order derivable. Then a Taylor series expansion is carried out for both sides of Eq. 5 based on the point W. The results are shown in Eq. 6. U(W − e) U(W) + U′(W)e + R ≈ U(W) + U′(W)e (8) 05 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 TABLE 1 Number of resettled households’ adoption of various green agricultural technologies. Sample townships Resettlers E-commerce participation Physical control Pollutionfree pesticides Soil testing fertilizer Laminated water control Grafting Fertilizer integration Maoping 159 126 54 48 36 27 51 9 Guojiaba 240 153 126 129 126 24 192 21 18 Guizhou 186 81 105 114 93 24 156 Shuitianba 75 39 51 30 36 3 57 18 Tatal 660 399 336 321 291 66 456 66 After samples with no response or doubtful key information were excluded, 660 valid questionnaires were obtained. The number of resettled households’ adoption of various green agricultural technologies is shown in Table 1. According to the list of technical systems in the “Technical Guidelines for Green Agricultural Development (2018–2030),” resettled households mainly apply six types of green agrotechnologies: physical control technology, pollution-free pesticide technology, soil formula fertilizer technology, ﬁlm and water control technology, water and fertilizer integration technology, and grafting technology. Of these, 120 households (18.2%) have adopted two kinds of technologies, and 249 households (37.7%) have adopted four or more kinds of technologies, which shows that green agrotechnologies are in the emerging stage in the resettlement area. are engaged in related industries, and this is the only national citrus production area that produces fresh fruit throughout the year. In order to solve the difﬁculties caused by the lack of land to compensate resettlers in the county, as well as a lack of social resources and the low premium capacity of agricultural products, Zigui County took the lead in innovating an e-commerce development model in 2014 and was selected as one of the second batch of national “e-commerce-demonstrating rural counties” in 2019. In recent years, the proportion of online sales of Navel oranges has accounted for about 55% of total sales, and the per capita net income of orange farmers increased from 1,967 USD in 2014 to 3,372 USD in 2018, becoming an important way to enhance the income of resettlers. In addition, e-commerce promotes the adoption of solar pest control lights, water and fertilizer integration, Internet of Things (IoT, through which things are connected through the Internet) management, and other green agrotechnology, using networks to strengthen the ecological brand marketed as “a river of clear water, green mountains on both sides of the river, four seasons of fresh oranges.” The data used in this study come from the Navel orange electric business and green agrotechnical survey conducted in the resettlement area of Zigui County, Hubei Province during December 2020. The sampling points were selected from a total of 34 resettlement villages and groups in three towns and one township, namely Maoping Town, Guojiaba Town, Guizhou Town, and Shuitianba Township. At present, each of the four has an electric e-commerce logistics center, an e-commerce service center, and other infrastructure and service facilities and has several green agrotechnology demonstration orchards or planting bases, such as the Bajiaolou Green Technology Tour Park, the Flying Green Plant Protection base in Guogutai Village, Guojiaba Township, and the Alibaba Group’s Future Farm in Choumushu Village, Shuitianba Township. All these were sampled. The speciﬁc sampling method was to randomly select eight to 10 sample villages in each township, then randomly conduct household surveys. A total of 688 resettlers were interviewed. Frontiers in Environmental Science 3.2 Model construction Using Li et al. (2020) classiﬁcation of degree of agrotechnology adoption, we divided resettlers’ agrotechnology adoption into ﬁve categories, from low to high, and assigned them the following values: lower adoption = 1, low adoption = 2, moderate adoption = 3, high adoption = 4, and higher adoption = 5. Since, as an explanatory variable, the degree of resettlers’ agrotechnology adoption g is a multi-valued ordered variable, we used the ordered probit model to explore the inﬂuencing factors involved, and the underlying regression model constructed is shown in Eq. 11. gpi α1 + β1 EPi + β2 GPEi + β3 TTSi + λ1 CVi + σ i (11) In Eq. 11, EPi , GPEi , and TTSi represent e-commerce participation, agricultural price expectation, and agrotechnology training of the ith resettled household, respectively, while CVi represents a series of control variables and σ i is a random disturbance term. gpi is a latent variable of the degree of green agrotechnology adoption of resettled household i. Let C1<C2<C3<C4<C5 be the threshold, then gi values can be discretized by gpi as shown in Eq. 12. 06 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 1 ⎪ ⎧ ⎪ ⎪ ⎪ ⎪ ⎨2 gi ⎪ 3 ⎪ ⎪ 4 ⎪ ⎪ ⎩ 5 gpi ≤ C1 C1 ≤ gpi ≤ C2 C2 ≤ gpi ≤ C3 C3 ≤ gpi ≤ C4 C4 ≤ gpi ≤ C5 ej −m i1 pij ln pij / ln m. 3) Derive the entropy weight wj (1 − ej )/nj1 (1 − ej ) of the jth indicator. 4) Obtain the weight coefﬁcients by calculating the weight of each secondary indicator of the evaluation dimensions, as shown in Table 2. (12) 3.3.2 Independent variables If the random disturbance term σ i obeys the standard normal distribution, X is a vector of actual observations of sample households for all independent variables, and ϕ denotes the cumulative distribution function. The impact mechanism of each adoption degree is shown in Eq. 13. Pgi ⎪ ⎧ ⎪ ⎪ ⎪ ⎪ ⎨ Pgi Pgi ⎪ ⎪ ⎪ Pgi ⎪ ⎪ ⎩ Pgi 1X Pgpi ≤ C1 Φ1 p 2X PC1 ≤ gip ≤ C2 Φ2 3X PC2 ≤ gi ≤ C3 Φ3 4X PC3 ≤ gpi ≤ C4 Φ4 5X PC4 ≤ gpi ≤ C5 Φ5 According to the model deduction of the previous theoretical analysis, it can be seen that the participation behavior of e-commerce. The expected sales price of agricultural products are the keys to promoting the adoption of green agrotechnology among resettlers. The improvement of agrotechnology application capacity through training is also an important factor (Liu et al., 2022). (13) 3.3.3 Moderating and controlling variables According to the aforementioned theoretical model, it is clear that the perceived risk factors of resettled households have an impact on e-commerce’s promotion of the use of agrotechnology. These involve various aspects and varying degrees of perception, such as natural conditions, market environment, and technical capacity. According to previous studies, there are two main types of control variables. One is demographic characteristics, including gender, age, and the level of education of respondents. The other is household endowment, including the maximum years of education of members, the proportion of household labor force, annual household income, navel orange planting area, and support from local cooperative organizations. The deﬁnition and descriptive statistics of each variable are shown in Table 3, in which the mean value of adoption of green agrotechnology is 1.763. The overall application degree still needs to be improved; however, participation in e-commerce is more than half, indicating that resettlers generally have enthusiasm to engage in e-commerce. At the same time, the current support for training in the use of agricultural technology is still insufﬁcient with the mean value is 0.414. The risk perception of the application of agricultural technology is around the mean value, which needs to be controlled and further development of e-commerce is required to promote the popularity of green agricultural technology in resettled areas. After an ordered-probit model is constructed in Eq. 13, the regression coefﬁcients can be estimated using the maximum likelihood estimation (MLE) method. In addition, we conducted additional analyses independent of the probit model. To analyze the extent to which e-commerce participation drives the adoption of agricultural technologies, we analyze the marginal effects of each independent variable, as shown in Eq. 14. zPgi nXzxj −Φn βj (14) Where n = 1,2,......5 represents the ﬁve degrees of resettled green agrotechnology adoption, xij is the jth independent variable of sample i, and βij is the coefﬁcient to be estimated for xij . The marginal effects were analyzed according to the sign and coefﬁcients of the results. 3.3 Variable selection 3.3.1 The dependent variables of this paper is resettled households’ adoption of green technologies in navel orange cultivation The speciﬁc measurement has six categories of green agricultural technologies adopted by resettled households, and the value is assigned as 1 if adopted, and 0 if not. Due to variations in natural capital among resettled households, it is not suitable to use equal weighting among the categories because of limitations placed on the use of different green agricultural technologies. Therefore, the four dimensions of economic beneﬁts, resource saving degree, ecological beneﬁts, and operational feasibility are considered comprehensively, and the entropy value method is applied to determine their weighting coefﬁcients. The steps are as follows: 1)Calculate the indicator weight pij Xij /m i1 Xij for item i under the jth indicator of resettled household Xij, where m is the number of evaluation dimensions. 2) Measure the entropy value of indicator j Frontiers in Environmental Science 4 Empirical results 4.1 Impact of e-commerce participation behavior on the degree of green agrotechnology adoption Before testing the role of e-commerce participation behavior in the promotion of the application of resettled farming techniques, a multiple cointegration test between the relevant independent variables was required. The resulting variance inﬂation factor (VIF) was far below 10, without cointegration 07 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 TABLE 2 Evaluation indicators and weights of the degree of adoption of various types of green agrotechnology by resettled households. Target layer Level 1 indicators Secondary indicators Secondary indicator weights Weighting of primary indicators Degree of adoption of green agrotechnology Physical control technology Economic beneﬁts 0.019 0.075 Resource conservation degree 0.019 Eco-friendly effect 0.018 Operability 0.019 Pollution-free pesticide technology Soil testing and fertilizer technology Lamination and water control technology Water and fertilizer integration technology Grafting and splicing technology Economic beneﬁts 0.018 Resource conservation degree 0.018 Eco-friendly effect 0.017 Operability 0.017 Economic beneﬁts 0.014 Resource conservation degree 0.014 Eco-friendly effect 0.014 Operability 0.014 Economic beneﬁts 0.096 Resource conservation degree 0.095 Eco-friendly effect 0.096 Operability 0.096 Economic beneﬁts 0.095 Resource conservation degree 0.098 Eco-friendly effect 0.096 Operability 0.097 Economic beneﬁts 0.007 Resource conservation degree 0.007 Eco-friendly effect 0.007 Operability 0.009 0.056 0.383 0.386 0.030 that support technology adoption support regarding such as price expectation of agricultural products and technical training support. According to the model regression results above, participation in e-commerce has a signiﬁcant positive effect on resettled households’ adoption of green agrotechnology. Speciﬁcally, various government subsidies and agricultural policies increased their green agricultural product price expectations after participating in e-commerce. For example, Zigui County has successively issued documents such as “Implementation Opinions on Accelerating E-commerce Development” and “Implementation Plan of E-commerce in Rural Areas Project,” which have greatly enhanced the information used by resettlers in their agrotechnology inputs. At the same time, resettlers have been given more technical training opportunities. For example, platforms such as Suning University and Jingdong Business School dispatched lecturers to problems. We then examined whether the original data satisﬁed the parallel regression hypothesis and found that the chi-square value was not signiﬁcant. This meant the hypothesis was valid for analysis using the ordered-probit model. Finally, regression models were constructed when no-control variables (Model I), household head characteristics (Model II), and household head characteristics and household endowment (Model III) were added. The regression models are shown in Table 4. As shown in Table 4, e-commerce participation signiﬁcantly and positively inﬂuenced resettled navel orange growers’ green agrotechnology adoption behavior at the 1% level, i.e., Hypothesis 1 holds. To analyze the inﬂuence mechanism, the relationship between e-commerce participation behavior, agricultural price expectation, and technical training support was further investigated here, as shown in Table 5. The regression results show that there is a signiﬁcant positive correlation between participation in e-commerce and factors Frontiers in Environmental Science 0.070 08 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 TABLE 3 Deﬁnition of variables and descriptive statistics. Variable types Variable name Variable deﬁnition and assignment Average Standard deviation Dependent variable Level of adoption of green agrotechnology Comprehensive calculation of the application of each type of agrotechnology (entropy method) 1.763 1.113 Independent variables E-commerce participation Participation = 1; No participation = 0 0.605 0.490 Green product price expectations Very low = 1; relatively low = 2; average = 3. relatively high = 4; very high = 5 3.436 0.648 Technical training support Accepted = 1; Not accepted = 0 0.414 0.494 Risk perception Perceived risk in green agrotechnology application: very small = 1; relatively small = 2; average = 3; relatively large = 4; very large = 5 2.782 1.076 Gender Male = 1, Female = 0 0.664 0.474 Age Actual age (years) 48.514 11.316 Education level Years of education (years) 9.777 2.925 Maximum number of years of education for members Maximum number of years of education for family members (years) 12.891 3.059 Moderating variables Control variables Household labor force ratio Ratio of labor force population to total population 0.725 0.235 Annual household income Total household income in 2020 (USD) 1.892 2.272 Orange planting area Household-owned orange cultivation area in 2020 (hectare) 0.7367 4.5678 Local partner organizations support Yes = 1; No = 0 0.164 0.371 TABLE 4 Regression results of the impact of green farming technology adoption among resettled orange growers. Variable types Variable name Model I Model II Model III Coefﬁcient (robust standard error) Coefﬁcient (robust standard error) Coefﬁcient (robust standard error) Independent variables E-commerce participation 0.821* (0.191) 0.793* (0.210) 0.859* (0.211) Green product price expectations 0.647* (0.135) 0.648* (0.135) 0.664* (0.136) Technical training support 0.516* (0.168) 0.533* (0.169) 0.481* (0.175) Control variables Risk perception −0.217 (0.176) −0.151 (0.182) Gender 0.003 (0.009) 0.002 (0.010) Age 0.026 (0.035) 0.029 (0.037) Education level 0.002 (0.037) Maximum number of years of education for members −0.164 (0.369) Household labor force ratio −0.021 (0.009) Annual household income 0.001 (0.001) −0.081 (0.228) Orange planting area Pseudo-R2 0.1467 0.1505 0.1635 Observed values 660 660 660 Note: , , and are statistically signiﬁcant at the 10%, 5%, and 1% levels, respectively. The results in Table 6 show that the percentage of resettled households with low technology adoption will decrease by 25% after they participate in e-commerce, while resettled households with moderate, higher, and high adoption will increase by 6.3%, 3.7%, and 7.4%, respectively. It is conﬁrmed that the awareness and adoption of green farming techniques among resettlers are 12 townships in Zigui County to teach e-commerce production and operation. Based on the Hypothesis 1 test, the degree of resettlers’ agro technology adoption under the role of each factor was further analyzed to explore whether the marginal effects of different degree types affected were sensitive, as shown in Table 6. Frontiers in Environmental Science 09 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 TABLE 5 Correlation between e-commerce participation and green agrotechnology adoption support elements for resettled households. Dependent variable Independent variables E-commerce participation Green product price expectations Technical training support 0.733* (0.180) 0.542*(0.209) Control variables Controlled Controlled Pseudo-R2 0.575 0.0988 TABLE 6 Marginal effects of each inﬂuencing factor on the degree of resettlers’ adoption of agricultural technology. Very low adoption type Low adoption type Moderate adoption type High adopted type Very high adoption type 0.074* (0.024) E-commerce participation −0.252* (0.052) 0.076* (0.017) 0.063* (0.018) 0.037 (0.015) Green product price expectations −0.198* (0.037) 0.060* (0.015) 0.050* (0.013) 0.029* (0.011) 0.059* (0.017) Technical training support −0.158* (0.049) 0.048* (0.017) 0.040* (0.015) 0.023 (0.010) 0.047* (0.018) TABLE 7 Differences in the impact of different e-commerce participation models on resettlers’ green agrotechnology adoption. Group Processing effects Processing group Control group Difference Standard error t-test value Overall sample ATT 2.092 1.565 0.527 0.210 2.50 Platform e-commerce model ATT 2.551 1.265 1.286 0.319 4.03 Social e-commerce model ATT 1.815 1.284 0.531 0.258 2.06 participating in e-commerce was 0.527, while the degrees of increase for the platform e-commerce and social e-commerce models were 1.286 and 0.531, respectively. The results indicate that the degree of increase for the platform e-commerce model was greater, suggesting that it could bring about a greater increase in agrotechnology adoption than the social e-commerce model. enhanced under the e-commerce model. Resettled households with a low level of green agrotechnology application also decreased by 19.8% and 15.8%, respectively, as the price of agricultural products improved and agrotechnology training grew in popularity, but it was not as signiﬁcant as the effect of e-commerce participation. 4.2 The impact of e-commerce participation model on the degree of adoption of green agricultural technology 4.3 Results of the moderating effect of risk perception To analyze the heterogeneity in the impact of different e-commerce participation models, this paper further empirically tested the impact of e-commerce participation models on resettled households’ green farming skills. Statistics found that, among the sampled households participating in e-commerce, 153 (38.3%) used the platform e-commerce model sample and 246 (62.7%) used the social e-commerce model sample. The empirical results are shown in Table 7. According to the level of technology adoption classiﬁed by the previous section (1–5), from the overall sample, the degree of increase in agrotechnology adoption by farmers To verify the moderating role of risk perception in resettled agrotechnology adoption, Model IV was constructed by introducing the interaction term between e-commerce participation and risk perception. The results are shown in Table 8, which indicate that while e-commerce participation signiﬁcantly increases the adoption of green agrotechnology, risk perception has a signiﬁcant negative correlation to it. Indeed, the negative coefﬁcient of the interaction term conﬁrms that risk perception weakens the positive relationship between e-commerce participation and agrotechnology adoption, so Hypothesis 3 is valid. Frontiers in Environmental Science 10 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 TABLE 8 Analysis of moderating effects of risk perception. Variable types Independent variables Variable name Model IV Model V Model VI Coefﬁcient (robust standard error) Coefﬁcient (robust standard error) Coefﬁcient (robust standard error) E-commerce participation 1.971* (0.562) 3.748* (0.642) 2.220* (0.425) Risk perception −0.174 (0.151) Green product price expectations 1.016* (0.246) 0.558* (0.154) Technical training support 0.744** (0.308) 0.352* (0.199) −0.518*** (0.129) Interaction items E-commerce participation * Risk perception −0.361 (0.184) −0.799* (0.195) Control variables Controlled Controlled Controlled Pseudo-R2 0.1538 0.1991 0.1987 Observed values 660 660 528 5 Discussion For example, when the rainfall in Zigui and other places reached historical extremes in the summer of 2021, to ensure the taste of fresh navel oranges was maintained, the local migration management recommended the adoption of land mulching technology in the selenium-rich planting bases of Seven Princesses, Shi Wai Tian Yuan and other major e-commerce companies. However, if the residual ﬁlm cannot be effectively recovered or is uncovered too late, the ﬁlm will be left in the soil by weathering and decaying, and its long-term accumulation will result in serious damage and pollution to the soil. This means that the households with high soil quality requirements are cautious about employing this technology. Hypothesis 3 also illustrates that both the objective risk caused by external shocks and the subjective risk caused by incomplete information after relocation may have a signiﬁcant impact on agricultural production decisions. Resettled households with weak resilience to potential risks to their livelihood, especially, are often forced to make careful trade-offs between low risk and high proﬁt. Based on relevant theories, this study reveals the impact of resettled households’ e-commerce participation on their green technological innovations, determines the impact mechanisms, and conducts an empirical test. Our results summarize the results and experiences of modernizing involuntary among Three Gorges Reservoir area in China, which may provide a modellable case study for future rural revitalization in China. This may solve the sustainable livelihood problems of involuntary resettlers in other countries. Due to the limited environmental capacity in the resettlement area, participation in social e-commerce among Reservoir resettlers is high, and it has gradually replaced the traditional agricultural cooperatives as the ﬁrst choice to expand the distribution channels of agricultural products and generate family income. At the same time, although the adoption of green agrotechnology among resettlers is gradually expanding, adoption levels remain low to medium. Previous studies have mostly focused on the income-generating effects of e-commerce and the adoption factors and environmental effects of green agrotechnology (Takahashi and Muraoka 2019; Peng et al., 2021; Li et al., 2022), with less research on the interrelationship between the two, but the huge impact of e-commerce on agricultural production cannot be ignored. Our results show that e-commerce participation signiﬁcantly increases the level of resettlers’ green agrotechnology adoption, shrinks the proportion of households with low adoption, and increases the proportion of resettled households with high adoption. This is similar to the ﬁndings of Li et al. (2021). Speciﬁcally, agricultural price expectations and related agrotechnology training support from e-commerce were the most important agrotechnology adoption drivers, and the former was more sensitive, indicating that resettlers urgently need to rely on green agricultural development to cope with an industrial hollowing-out in the 4.4 Robustness test To test the robustness of our empirical results, one method was to construct another ordered-logit model (Model V) for the ordered variable of degree of resettlers’ green farming technology adoption to conduct a regression analysis. The other was to extract 80% of resettlers’ sample households and test them again through the ordered-probit model (Model VI). The results of both are shown in Table 8. As displayed in Table 8, there is essentially no difference in the sign (positive and negative values) of the regression coefﬁcients for Models V and VI, or the signiﬁcance level of the coefﬁcients. This indicates the strong robustness of the hypotheses derived from the theoretical model, as well as the mechanisms of inﬂuence of the selected independent and control variables on the dependent variable. Frontiers in Environmental Science 11 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 can continue to be improved. Subsequent studies can improve on the above aspects. Three Gorges Reservoir area due to natural capital loss and compensation resource constraints. However, green value-added agricultural products and e-commerce-assisted agrotechnical training mainly act on resettled households with low agrotechnical skills and, without the participation of e-commerce, will have limited impact effects. Heterogeneous e-commerce participation models impact agrotechnology adoption differently. Many studies have shown that as e-commerce continues to develop, different e-commerce models, such as social e-commerce and platform e-commerce, are available (Luo 2022). Because the platform e-commerce model requires high household endowment, which is difﬁcult for resettled households to afford, most such households choose the social e-commerce model. Chinese people value social relationships and are very good at using them (Zhang and Yang 2022). Resettled households share product information through social networks and word of mouth and can expand their product sales network more simply. Resettlers’ awareness and adoption of green farming techniques have increased under various e-commerce models. Of these, platform e-commerce participation brings the greatest degree of enhancement; the traditional e-commerce participation model has more intense market competition and consumers’ product requirements are higher (Heuer et al., 2015), which pushes resettled households to continuously improve the level of green agrotechnology adoption. In contrast, due to the limited network of resettled households’ acquaintances, the demand for green agrotechnology adoption driven by social e-commerce is lower, as is the degree of improvement it brings. Related studies have shown that uncertainty in the application and use of technologies when new technologies are introduced will also cause uncertainty in income (Chavas and Shi 2015; Hörner and Wolln 2022). Involuntary resettlers have a tendency to avoid livelihood risks and fear their own lack of resilience, thus, if they perceive risks in technology application, they are likely to weaken the green agrotechnology promotion effect brought about by e-commerce participation. Theoretical improvements and innovations are often not convincing to farmers (Bozzola and Finger 2020), especially when the technical barriers to green agrotechnology application are high or the effectiveness of use is uncertain, which will lead to higher risk perceptions in technology adoption and thus make e-commerce resettlers cautious about green agrotechnology. Some shortcomings must be acknowledged. First, the agrotechnology selected in this paper is citrus cultivation technology, and the signiﬁcance of other agricultural products still needs to be veriﬁed. Second, we tested limited factors inﬂuencing green agrotechnology. Government regulation and incentives, the role of markets, and individual capabilities still need to be further veriﬁed. Third, the rhosquared value is not at a high degree, so the model in this paper Frontiers in Environmental Science 6 Conclusion and policy recommendations Based on previous technology adoption theories, this paper constructs a theoretical model of green agrotechnology adoption by resettled households in the development of e-commerce in hydropower project reservoirs, taking into account the characteristics of Chinese involuntary project resettlers, the objective of maximizing expected returns, and the perceived risk to technology application posed by resettlers’ livelihood risk resistance. Drawing on survey data from resettled households in Zigui, the ﬁrst county in the Three Gorges Reservoir area, the hypotheses derived from the theoretical model were tested empirically by using the ordered probit model. The conclusions are as follows: Green agrotechnology has become more popular among resettlers, but the overall adoption level is still low; participation in e-commerce has a signiﬁcant positive impact on the adoption of green agrotechnology at the 1% level. The ecological value expectation of agricultural products and the agrotechnology support provided by e-commerce are the most important driving factors, but their current effects are mostly limited to resettled households with a low level of agrotechnology. Compared with the social e-commerce participation model, platform e-commerce brings more signiﬁcant improvements in technology adoption. The risk perception in resettled households’ agrotechnology application weakens the promotion effect of e-commerce participation on agrotechnology adoption, while the risk perception of e-commerce participation weakens the promotion effect of the latter on the adoption of green agrotechnology. Based on these ﬁndings, the following policy recommendations are proposed: 1) supporting fund should be used to improve network and logistics infrastructure. Through the combination of e-commerce policy and late-stage supporting system for resettlers, we can increase the participation rate of e-commerce and promote the application of green agricultural technology. 2) On the basis of the regional brand of the products in the resettled area, we embed the spirit of resettlers and create a special brand. On the one hand, it can realize the traceability of the whole growth cycle of agricultural products, and on the other hand, it can do the whole cycle of product development and expand a variety of products. This can enhance the agronomic value expectation of resettlers by adding value to agricultural products. 3) The government-led e-commerce associations in resettled areas should promote each other with the informal business organizations of resettlers formed by e-commerce platforms. It is necessary to provide green agrotechnology training more precisely. However, it is necessary to promote 12 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 Author contributions close interaction between resettlers and local residents through joint participation in e-commerce. Through small-scale technology demonstration and ﬁeld guidance, the dissemination and exchange of tacit knowledge in the application of agricultural technology in resettler communities should be promoted. 4) To address the riskaverse tendency of reservoir resettlers in the application of green agricultural technology, the government of resettled areas should strengthen the publicity and popularize the knowledge of technical risks to avoid excessive precautionary behavior by resettlers. Meanwhile, the government should cooperate with insurance agencies to encourage farmers to purchase agricultural insurance policies. Such insurance policies are speciﬁcally used to share the technical costs incurred by resettler groups during the medium- and long-term growth periods of agricultural products and are incorporated into latestage systems of risk sharing. This can reduce resettlers’ hesitancy to adopt green agrotechnologies. All authors contributed to the study conception and design. The ﬁrst draft of the manuscript was written by ZC, XZ, and FZ. All authors commented on previous version of the manuscript. All authors read and approved the ﬁnal manuscript. Funding This research was ﬁnancially supported by National Natural Science Foundation of China: Livelihood Elasticity Measurement and Simulation Research on Later Supporting Strategies of Involuntary Resettlers from the Perspective of Habitat Mutation (Project Number: 72271142). The Major Program of National Social Foundation of China: Study on the governance and late-stage support of inter-county relocation communities (Project Number: 21&ZD183). Data availability statement Conﬂict of interest The raw data supporting the conclusion of this article will be made available by the authors, without undue reservation. The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Ethics statement Publisher’s note Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent from the participants was not required to participate in this study in accordance with the national legislation and the institutional requirements. All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. References Chavas, J. P., and Shi, G. M. (2015). An economic analysis of risk, management and agricultural technology. J. Agric. Resour. Econ. 40 (1), 63–79. Adnan, N., Nordin, M. D. H., and Anwar, A. (2021). Transition pathways for Malaysian paddy farmers to sustainable agricultural practices: An integrated exhibiting tactics to adopt green fertilizer. Land Use Policy 90 (6), 104255–108129. doi:10.1016/j.landusepol.2019.104255 Chen, X., Wu, Y. L., and Zhong, R. J. (2022). Optimal marketing channel and strategy in social commerce. RAIRO-Oper. Res. 56 (3), 1203–1221. doi:10.1051/ro/ 2022053 Adnan, N., Nordin, S. M., and Rasli, A. M. (2019). A possible resolution of Malaysian sunset industry by green fertilizer technology: Factors affecting the adoption among paddy farmers. Environ. Sci. Pollut. Res. 26 (2), 27198–27224. doi:10.1007/s11356-019-05650-9 Cowan, B. W., Lee, D., and Shumway, C. R. (2015). The induced innovation hypothesis and U.S. public agricultural research. Am. J. Agric. Econ. 97 (3), 727–742. doi:10.1093/ajae/aau090 Bidogeza, J. C., Berentsen, P. B. M., De Graaff, J. A., and Oude Lansink, G. J. M. (2009). A typology of farm households for the Umutara Province in Rwanda. Food Secur. 1 (1), 321–335. doi:10.1007/s12571-009-0029-8 Dong, H., Wang, H., and Han, J. (2022). Understanding ecological agricultural technology adoption in China using an integrated technology acceptance model—theory of planned behavior model. Front. Environ. Sci. 10, 927668. doi:10.3389/fenvs.2022.927668 Bozzola, M., and Finger, R. (2020). Stability of risk attitude, agricultural policies and production shocks: Evidence from Italy. Eur. Rev. Agric. Econ. 48 (3), 477–501. doi:10.1093/erae/jbaa021 Dong, S., Yang, L., Shao, X., Zhong, Y., and Qiao, P. (2021). How can channel information strategy promote sales by combining ICT and blockchain? Evidence from the agricultural sector. J. Clean. Prod. 299 (22), 126857. doi:10.1016/j.jclepro. 2021.126857 Chatzimichael, K., Genius, M., and Tzouvelekas, V. (2014). Informational cascades and technology adoption: Evidence from Greek and German organic growers. Food Policy 49 (1), 186–195. doi:10.1016/j.foodpol.2014.08.001 Duong, T. T., Tom, B., Jo, L., and Kerstin, Z. (2019). A global review of farmers’ perceptions of agricultural risks and risk management Strategies. Agriculture 9 (1), 10. doi:10.3390/agriculture9010010 Chavas, J. P., and Nauges, C. (2020). Uncertainty, learning and technology adoption in agriculture. Appl. Econ. Perspect. Policy 42 (1), 42–53. doi:10.1002/aepp.13003 Frontiers in Environmental Science 13 frontiersin.org Zhao et al. 10.3389/fenvs.2022.1036464 Fang, L., Hu, R., Mao, H., and Chen, S. (2021). How crop insurance inﬂuences agricultural green total factor productivity: Evidence from Chinese farmers. J. Clean. Prod. 321 (7), 128977. doi:10.1016/j.jclepro.2021.128977 Ni, J., and Shao, J. (2013). The drivers of land use change in the migration area, Three Gorges Project, China: Advances and prospects. J. Earth Sci. 24 (1), 136–144. doi:10.1007/s12583-013-0306-5 Fu, W. Q., and Zhang, R. W. (2022). Can digitalization levels affect agricultural total factor productivity? Evidence from China. Front. Sustain. Food Syst. 6, 860780. doi:10.3389/fsufs.2022.860780 Peng, C., Biao, M. A., and Zhang, C. (2021). Poverty alleviation through e-commerce: Village involvement and demonstration policies in rural China. J. Integr. Agric. 20 (4), 998–1011. doi:10.1016/S2095-3119(20)63422-0 Gao, Y., and Niu, Z. H. (2019). Risk aversion, information acquisition ability and farmers’ adoption behavior of green control techniques. Chin. Rural. Econ. (8), 109–127. Qiu, H. G., Su, L. F., Zhang, Y. T., and Tang, J. J. (2020). Risk preference, risk perception and farmers’ adoption of conservation tillage. Chin. Rural. Econ. (7), 59–79. Gong, Y. C., Zhang, R. L., Yao, K. W., Liu, B. W., and Wang, F. L. (2020). A livelihood resilience measurement framework for dam-induced displacement and resettlement. Water 12 (11), 3191. doi:10.3390/w12113191 Schultz, T. W. (1987). Transforming traditional agriculture. Beijing: The Commercial Press. Takahashi, K., Muraoka, R., and Otsuka, K. (2019). Technology adoption, impact, and extension in developing countries’ agriculture: A review of the recent literature. Agric. Econ. 51 (1), 31–45. doi:10.1111/agec.12539 Heuer, D., Brettel, M., and Kemper, J. (2015). Brand competition in fashion e-commerce. Electron. Commer. Res. Appl. 14 (6), 514–522. doi:10.1016/j.elerap. 2015.07.007 Tian, T., Zhang, Y., and Mei, Y. (2022). Intelligent analysis of precision marketing of green agricultural products based on big data and GIS. Earth Sci. Inf. 15, 1395–1406. doi:10.1007/s12145-022-00789-5 Hörner, D., and Wolln, M. (2022). Does integrated soil fertility management increase returns to land and labor? Agric. Econ. 53 (3), 337–355. doi:10.1111/agec. 12699 Karimi, S., and Taifur, W. D. (2013). Resettlement and development: A survey of two of Indonesia’s koto panjang resettlement villages. Int. J. Water Resour. Dev. 29 (1), 35–49. doi:10.1080/07900627.2012.739539 Wang, W. K., Wang, J., Liu, K. B., and Wu, Y. J. (2020). Overcoming barriers to agriculture green technology diffusion through stakeholders in China: A social network analysis. Int. J. Environ. Res. Public Health 17 (19), 6976–76. doi:10.3390/ ijerph17196976 Koundouri, P., Nauges, C., and Tzouvelekas, V. (2006). Technology adoption under production uncertainty: Theory and application to irrigation technology. Am. J. Agric. Econ. 88 (3), 657–670. doi:10.1111/j.1467-8276.2006.00886.x Wu, Q., and Zhou, J. (2021). Need for cognitive closure, information acquisition and adoption of green prevention and control technology. Ecol. Chem. Eng. S 28 (1), 129–143. doi:10.2478/eces-2021-0011 Li, B., Xu, C., Zhu, Z., and Kong, F. (2021). Does e-commerce drive rural households engaged in non-timber forest product operations to adopt green production behaviors? J. Clean. Prod. 320, 128855. doi:10.1016/j.jclepro.2021. 128855 Xiao, J., Wang, W., and Tsai, S. (2021). Coupling of agricultural product marketing and agricultural economic development based on big data analysis and “Internet+”. Mob. Inf. Syst., 2021, 1–10. doi:10.1155/2021/3702064 Yan, D. C., Wang, M., Wang, H. B., and Shi, G. Q. (2018). Policy and implementation of land-based resettlement in China (1949–2014). Int. J. Water Resour. Dev. 34 (3), 453–471. doi:10.1080/07900627.2017.1417824 Li, H., Tang, M., Cao, A., and Guo, L. (2022). Assessing the relationship between air pollution, agricultural insurance, and agricultural green total factor productivity: Evidence from China. Environ. Sci. Pollut. Res. Int. doi:10.1007/s11356-022-21287-7 Yin, Z. H., and Choi, C. H. (2022). Does e-commerce narrow the urban–rural income gap? Evidence from Chinese provinces. Internet Res. 32 (4), 1427–1452. doi:10.1108/INTR-04-2021-0227 Li, X. J., Chen, Z., Liu, F., and Xia, X. L. (2020). Does participating in E-commerce promote the adoption of green production technologies by kiwifruit growers? A counterfactual estimation based on propensity score matching method. Chin. Rural. Econ. (3), 118–135. Zeng, Y. M., Zhang, J. B., He, K., and Cheng, L. L. (2019). Who cares what parents think or do? Observational learning and experience-based learning through communication in rice farmers’ willingness to adopt sustainable agricultural technologies in Hubei province, China. Environ. Sci. Pollut. Res. 26 (12), 12522–12536. doi:10.1007/s11356-019-04609-0 Liu, Y. Y., Shi, R. L., Peng, Y. T., Wang, W., and Fu, X. H. (2022). Impacts of technology training provided by agricultural cooperatives on farmers’ adoption of biopesticides in China. Agriculture 12 (3), 316. doi:10.3390/agriculture12030316 Zhang, Q., and Yang, J. (2022). Model construction of hierarchical polarization characteristics combined with social E-commerce consumer behavior. Secur. Commun. Netw. 2022, 9666677. doi:10.1155/2022/9666677 Lu, Y. H., Chen, S. J., and Li, X. M. (2018). Relationship poverty: The fracture and reconstruction of immigrant social relations network: Take S village of resettled in danjiangkou reservoir for example. China Agric. Univ. J. Soc. Sci. Ed. 35 (2), 114–122. doi:10.13240/j.cnki.caujsse.2018.02.025 Zhang, Y. (2021). Understanding China’s dam-induced resettlement under the institutionalised governance process of policy coevolution. J. Environ. Manag. 283 (7), 111968. doi:10.1016/j.jenvman.2021.111968 Luo, N. (2022). Innovation of E-commerce development model under the background of artiﬁcial intelligence and wireless communication. Wirel. Commun. Mob. Comput., 2022, 8572911–8572917. doi:10.1155/2022/8572911 Zuo, Z. Y., and Fu, Z. H. (2021). The environmental and economic effects of green agricultural subsidy policies: A regression of discontinuity design based on agricultural non-point source pollution control programs with the World Bank’s loan in guangdong province. Chin. Rural. Econ. (2), 106–121. Natkhov, T., and Vasilenok, N. (2021). Skilled immigrants and technology adoption: Evidence from the German settlements in the Russian empire. Explor. Econ. Hist. 81, 101399. doi:10.1016/j.eeh.2021.101399 Frontiers in Environmental Science 14 frontiersin.org
https://openalex.org/W2604759123	https://sam.ensam.eu/bitstream/10985/13117/3/LE2I_VR_PLOUZEAU_2017.pdf	English	null	Mechanism of integrating force and vibrotactile cues for 3D user interaction within virtual environments	null	2,017	cc-by	3,006	Science Arts & Métiers (SAM) is an open access repository that collects the work of Arts et Métiers Institute of Technology researchers and makes it freely available over the web where possible. This is an author-deposited version published in: https://sam.ensam.eu Handle ID: .http://hdl.handle.net/10985/13117 To cite this version : Aïda ERFANIAN, Stanley TARNG, Yaoping HU, Jeremy PLOUZEAU, Frédéric MERIENNE - Mechanism of Integrating Force and Vibrotactile Cues for 3D User Interaction within Virtual Environments - In: IEEE Virtual Reality, Etats-Unis, 2017-03-18 - IEEE Virtual Reality - 2017 Science Arts & Métiers (SAM) is an open access repository that collects the work of Arts et Métiers Institute of Technology researchers and makes it freely available over the web where possible. This is an author-deposited version published in: https://sam.ensam.eu Handle ID: .http://hdl.handle.net/10985/13117 To cite this version : Aïda ERFANIAN, Stanley TARNG, Yaoping HU, Jeremy PLOUZEAU, Frédéric MERIENNE - Mechanism of Integrating Force and Vibrotactile Cues for 3D User Interaction within Virtual Environments - In: IEEE Virtual Reality, Etats-Unis, 2017-03-18 - IEEE Virtual Reality - 2017 Science Arts & Métiers (SAM) is an open access repository that collects the work of Arts et Métiers Institute of Technology researchers and makes it freely available over the web where possible. This is an author-deposited version published in: https://sam.ensam.eu Handle ID: .http://hdl.handle.net/10985/13117 1 INTRODUCTION Virtual environments (VEs) need to provide human users with multi-sensory feedback [3]. As user tasks within VEs become progressively demanding in precision and accuracy, it is necessary to integrate multi-sensory feedback for meeting the demand. Each human sensory modality responds to various cues, which stimulate certain sub-categories of the modality. For the modality of haptics, sub-categorical cues include kinesthetic force and pressure, as well cutaneous textures of surfaces (e.g., vibration, etc.) [6, 12]. Such cues have been applied to enhance user interaction within VEs [11]. Force and vibrotactile cues are often used to aid the users in manipulating objects, along with visual information of the objects. y Ten male participants (26.78 ± 5.77 years old, and naïve to the purpose of the study) took part in the study. A pre-screening verified each participant to be right-handed, with normal (or correct-to-normal) vision including stereo acuity and color recognition. During the pre-screening, the recorded physiological data of the E4 wristband provided a baseline for each participant [2]. The study had an ethics approval by the University of Calgary. The task of the participant was to inspect the defects on the line, by using the stylus of the device to fly the drone and slide the clamp along the line. A continuous force, ܨԦ௖, tangentially along the line facilitated the participant to govern the flying drone. With a constant magnitude of 0.5 N, the force updated dynamically its direction with respect to the location of the clamp on the line. Any defect was signaled via the individual or combined force and vibrotactile cues, which formed the following 5 haptic profiles: Ten male participants (26.78 ± 5.77 years old, and naïve to the purpose of the study) took part in the study. A pre-screening verified each participant to be right-handed, with normal (or correct-to-normal) vision including stereo acuity and color recognition. During the pre-screening, the recorded physiological data of the E4 wristband provided a baseline for each participant [2]. The study had an ethics approval by the University of Calgary. When manipulating physical objects, the users expect feedbacks in contact with the objects. That is, force and vibrotactile cues are collocated to cause the users’ cognitive responses for their action [1]. Due to design restriction and implementation of haptic devices, force and vibrotactile cues could be collocated or dislocated for 3D user interaction within VEs [13, 15]. Jérémy Plouzeau† Frédéric Merienne† * Dept. of Electrical and Computer Engineering Schulich School of Engineering University of Calgary, CANADA † LE2I UMR6306 Arts et Métiers CNRS University, FRANCE ABSTRACT facilitating 3D interaction. Existing reports on multi-sensory integration have focused on cues of the visual and haptic modalities [5, 14], disparity and texture cues of the visual modality [8], and force and position cues of the haptic modality [4]. These integrations had collocated cues and followed maximum likelihood estimation (MLE) [5]. However, little effort is devoted to the mechanism of integrating force and vibrotactile cues. We thus undertook a human study to investigate the suitability of MLE for integrating force and vibrotactile cues of the haptic modality. facilitating 3D interaction. Existing reports on multi-sensory integration have focused on cues of the visual and haptic modalities [5, 14], disparity and texture cues of the visual modality [8], and force and position cues of the haptic modality [4]. These integrations had collocated cues and followed maximum likelihood estimation (MLE) [5]. However, little effort is devoted to the mechanism of integrating force and vibrotactile cues. We thus undertook a human study to investigate the suitability of MLE for integrating force and vibrotactile cues of the haptic modality. Proper integration of sensory cues facilitates 3D user interaction within virtual environments (VEs). Studies showed that the integration of visual and haptic cues follows maximum likelihood estimation (MLE). Little effort focuses however on the mechanism of integrating force and vibrotactile cues. We thus investigated MLE’s suitability for integrating these cues. Within a VE, human users undertook 3D interaction of navigating a flying drone along a high-voltage transmission line for inspection. The users received individual force or vibrotactile cues, and their combinations in collocated and dislocated settings. The users’ task performance including completion time and accuracy was assessed under each individual cue and setting. The presence of the vibrotactile cue promoted a better performance than the force cue alone. This agreed with the applicability of tactile cues for sensing 3D surfaces, herein setting a baseline for using MLE. The task performance under the collocated setting indicated a degree of combining the individual cues. In contrast, the performance under the dislocated setting was alike under the individual vibrotactile cue. These observations imply a possible role of MLE in integrating force and vibrotactile cues for 3D user interaction within VEs. * email: {aerfania, stanley.tarng, huy}@ucalgary.ca † email: {jeremy.plouzeau, frederic.merienne}@ensam.eu 2 EMPIRICAL STUDY Using the Unity game engine, we developed a VE for a 3D interactive task. The VE employed one motor of a VibroTac bracelet (SENSODRIVE Gmbh, Germany) and a PHANToM Omni device (Geomagic Inc., USA) to provide vibrotactile and force cues, respectively. While a participant (the user) used his/her right hand to hold the device’s stylus, the bracelet could be collocated with the device’s stylus on the right hand or dislocated on his/her right forearm. For each participant, a pair of ear plugs blocked out the noise generated by the bracelet. An E4 wristband (Empatica Inc., Italy) on the left wrist of the participant monitored his/her physiological signals. On a wall-sized screen, the VE displayed the stereoscopic scene of a high-voltage power transmission line located in an inaccessible mountainous region. The line was curved downwards between two supporting towers due to the gravity. The participant viewed the scene through a pair of passive stereo goggles and used the device’s stylus to fly a drone along the curved line. A camera on the front of the drone allowed the participant to view the scene from the drone’s perspective. A rigid arm attached to the bottom of the drone had a clamp at its distal end. The clamp covered the line for sensing defects. The defects were minuscular for visual differentiation and randomly distributed on the line. The participant declared a defect by pressing down both buttons on the device’s stylus. Index Terms: H.5.1 [Multimedia Information Systems]: Artificial, augmented and virtual realities; H.5.2 [User Interfaces]: Haptic I/O. To cite this version : To cite this version : Aïda ERFANIAN, Stanley TARNG, Yaoping HU, Jeremy PLOUZEAU, Frédéric MERIENNE - Mechanism of Integrating Force and Vibrotactile Cues for 3D User Interaction within Virtual Environments - In: IEEE Virtual Reality, Etats-Unis, 2017-03-18 - IEEE Virtual Reality - 2017 Aïda ERFANIAN, Stanley TARNG, Yaoping HU, Jeremy PLOUZEAU, Frédéric MERIENNE Mechanism of Integrating Force and Vibrotactile Cues for 3D User Interaction within Virtual Environments - In: IEEE Virtual Reality, Etats-Unis, 2017-03-18 - IEEE Virtual Reality - 2017 Any correspondence concerning this service should be sent to the repository Administrator : scienceouverte@ensam.eu Any correspondence concerning this service should be sent to the repository Administrator : scienceouverte@ensam.eu Any correspondence concerning this service should be sent to the repository Administrator : scienceouverte@ensam.eu Any correspondence concerning this service should be sent to the repository Administrator : scienceouverte@ensam.eu REFERENCES [1] J. Brown and R. Gillespie. The effect of force/motion coupling on motor and cognitive performance. Proc. IEEE WHC, pages 197-202, 2011. [2] J. J. Braithwaite, D. G. Watson, R. Jones, and M. Rowe. A guide for analysing electrodermal activity & skin conductance responses for psychological experiments. Psychophysiology, 49: 1017-1034, 2013. [3] Q. Dinh, N. Walker, C. Song, A. Kobayashi, and L. F. Hodges. Evaluating the importance of multi-sensory input on memory and the sense of presence in virtual environments. Proc. IEEE Virtual Reality, pages 222–228, 1999. [4] K. Drewing, M. O. Ernst. Integration of force and position cues for shape perception through active touch. Brain Res., 1078:92-100, 2006. [5] M. O. Ernst and M. S. Banks. Humans integrate visual and haptic information in a statistically optimal fashion. Nature, 415(6870):429– 433, 2002. [6] S. Gallo, L. Santos-Carreras, G. Rognini, M. Hara, A. Yamamoto, and T. Higuchi. Towards multimodal haptics for teleoperation: Design of a tactile thermal display. Proc. IEEE ADC, pages 1-5, 2012. [7] S. G. Hart. Nasa-Task Load Index (NASA-TLX); 20 Years Later. Proc. Annual Meeting of Human Factors Ergonomics Society, 50(9):904–908, 2006. The integration of both cues was examined on Acc by using MLE [14], because TCT was non-differentiable. Each testing block yielded an experimentally estimated Gaussian distribution, according to the MLE’s rules. Each distribution was characterized by its magnitude (M), mean (ߤሻ, and standard deviation (ߪ). [8] J. M. Hillis, M. O. Ernst, M. S. Banks, and M. S. Landy. Combining sensory information: mandatory fusion within, but not between, senses. Science, 298(5598):1627–1630, 2002. [9] R. S. Kennedy, N. E. Lane, K. S. Berbaum, and M. G. Lilienthal. Simulator sickness questionnaire: an enhanced method for quantifying simulator sickness. Int. J. Aviation Psych., 3(3):203-220, 1993. Under the collocated setting, the Gaussian distribution of the FV_co block overlapped evenly with those of both V_co and F_only blocks. The ߤ of the FV_co block was very similar to that of a predicted distribution, which combined theoretically both cues by the rules of MLE. This is much in the same way as multi-sensory integration of visual and haptic cues [5]. However, the combination of both force and vibrotactile cues is not optimal, due to the different M and ߪ between the FV_co block and its prediction. [10] R. H. LaMotte. Softness discrimination with a tool. J. Neurophysiology., 83(4):1777–1786, 2000. [11] S.J. Lederman, A. Martin, C. Tong, and R.L. Klatzky. 1 INTRODUCTION Proper integration of both cues is therefore paramount for creating intuitive VEs and The task of the participant was to inspect the defects on the line, by using the stylus of the device to fly the drone and slide the clamp along the line. A continuous force, ܨԦ௖, tangentially along the line facilitated the participant to govern the flying drone. With a constant magnitude of 0.5 N, the force updated dynamically its direction with respect to the location of the clamp on the line. Any defect was signaled via the individual or combined force and vibrotactile cues, which formed the following 5 haptic profiles: F_only: An individual force cue had a constant magnitude of 0.6 N to last 1.0 s. The cue was 20% stronger than ܨԦ௖ (> 15% – a just noticeable difference [10]) but with the same direction. F_only: An individual force cue had a constant magnitude of 0.6 N to last 1.0 s. The cue was 20% stronger than ܨԦ௖ (> 15% – a just noticeable difference [10]) but with the same direction. Table 1: Means and standard deviations of the objective and subjective data; and ANOVA results among all testing blocks. V_co: An individual vibrotactile cue was a vibration at 200 Hz. This frequency was within the sensing range of the humans [15]. The cue lasted 1.0 s on the right hand. distribution of the FV_dis block has a much similar ߪ as that of the prediction. However, the distribution of the FV_dis block resembles that under the individual vibrotactile cue. subjective data; and ANOVA results among all testing blocks. Data (Objective/ Subjective) Testing Blocks (ࣆ ± ࣌) ANOVA F_ only V_ co V_ dis FV_ co FV_ dis F(4, 49) p <0.05 TCT (s) 4.1±0.6 4.2±0.9 4.2±0.8 4.9±2.1 4.0±0.8 1.39 — Acc (%) 73 ± 10 89 ± 7 86 ± 9 82 ± 12 91 ± 7 4.53  Usf (%) 76 ± 15 73 ± 8 75 ± 9 74 ± 11 77 ± 9 0.58 — Eff (%) 62 ± 22 73 ± 11 75 ± 13 73 ± 11 74 ± 11 3.77  Pls (%) 56 ± 21 71 ± 12 72 ± 12 72 ± 13 75 ± 12 3.95  Wld 116 ±30 109 ±31 104 ±34 110 ±46 107 ±32 0.52 — humans [15]. The cue lasted 1.0 s on the right hand. 3 RESULTS AND DISCUSSION None of the participants had cybersickness, according to their SSQ responses and the consistency between their physiological baseline and recoded data. There were no outliers among the participants, because each of them detected much more than 7 defects. All objective and subjective data were analyzed using one-way repeated measures analysis of variance (ANOVA). Normality tests were conducted to verify the eligibility of these data for ANOVA analyses. As summarized in Table 1, ANOVA analyses on TCT, Usf and Wld revealed no significant difference among all testing blocks. In contrast, a significant differentiability was observed for each of Acc, Pls and Eff. Post-hoc Bonferroni tests indicated that, for Acc, 3 pairs of the blocks (F_only vs. V_co; F_only vs. V_dis; and F_only vs. FV_dis) were differentiable. For Pls and Eff, there was a significant difference between the F_only block and every other block. That is, the vibrotactile cue promoted a better Acc, Eff, and Pls than the force cue alone. This agreed with the observations on tactile cues for sensing 3D surfaces [11], establishing a baseline for examining the integration of force and vibrotactile cues. 4 CONCLUSION We investigated the suitability of MLE for integrating force and vibrotactile cues. Our observations imply that MLE could play a role in integrating force and vibrotactile cues for 3D user interaction within VEs. Future work will investigate unique particularities arose from the integration of both cues. 1 INTRODUCTION g V_dis: An individual vibrotactile cue had the same vibration as V_co, but was located on the right forearm. FV_co: A collocated setting of the concurrent F_only and V_co profiles on the right hand. FV_dis: A dislocated setting of the concurrent F_only and FV_dis: A dislocated setting of the concurrent F_only and V_dis profiles on the right hand and forearm, respectively. Each of these haptic profiles corresponded to one testing block. _ g _ y V_dis profiles on the right hand and forearm, respectively. Each of these haptic profiles corresponded to one testing block. V_dis profiles on the right hand and forearm, respectively. Each of these haptic profiles corresponded to one testing block. There was one practice block before each testing block. During each practice block, the participant learnt the task by utilizing an individual or combined cues. The locations of 15 defects differed among all practice and testing blocks. After each block, the participant answered two questionnaires: one on cybersickness (SSQ) [9]; and another on the subjective data of each participant’s perceived usefulness (Usf), effectiveness (Eff), pleasure (Pls), and workload (Wld). The perceived workload was devised by using the NASA task load index [7]. As objective data, our VE application logged his/her task performance including task completion time (TCT) and accuracy in identifying defects (Acc). The participant spent averagely 1.5 hours in the study. The testing blocks were counterbalanced for all participants in a within-subject design. distribution of the FV_dis block has a much similar ߪ as that of the prediction. However, the distribution of the FV_dis block resembles that under the individual vibrotactile cue. distribution of the FV_dis block has a much similar ߪ as that of the prediction. However, the distribution of the FV_dis block resembles that under the individual vibrotactile cue. distribution of the FV_dis block has a much similar ߪ as that of the prediction. However, the distribution of the FV_dis block resembles that under the individual vibrotactile cue. REFERENCES Relative performance using haptic and/or touch-produced auditory cues in a remote absolute texture identification task. Proc. HI VE & Teleoperation, pages 151-158, 2003. [12] K. E. MacLean. Designing with haptic feedback. Proc. IEEE ICRA, San Francisco, CA, pages 783-788, 2000. Under the dislocated setting, the Gaussian distribution of the FV_dis block overlapped that of the V_dis block much more than the F_only block. The ߤ of the FV_dis block was within 1ߪ from the ߤ of the V_dis block, but far beyond 1ߪ from the ߤ of the F_only block. The M and ߪ of the FV_dis block were the highest and narrowest among all three blocks, followed by those of the V_dis block and the F_only block in sequence. Moreover, the ߤ of the predicted distribution for combing both cues was much smaller than that of the FV_dis block, but with the similar M and ߪ. The role of the individual force cue is not ignored entirely, as the [13] G. Rosati, F. Oscari, C. Pacchierotti, and D. Prattichizzo. Effects of kinesthetic and cutaneous stimulation during the learning of a viscous force field. IEEE Trans. Haptics, 7(4):251-263, 2014. [14] M. Rohde, C. J. van Dam, and M. O Ernst. Statistically optimal multisensory cue integration: a practical tutorial. Multisensory Research, 29(4-5):279-317, 2016. [15] J. M. Walker, A. A. Blank, P. A. Shewokis, and M. K. O'Malley. Tactile feedback of object slip facilitates virtual object manipulation. IEEE Trans. Haptics, 8(4):454–466, 2015.
https://openalex.org/W4362468288	https://periodicorease.pro.br/rease/article/download/8882/3444	Portuguese	null	CONTRIBUIÇÃO AO DESENVOLVIMENTO SUSTENTÁVEL DO JALAPÃO TOCANTINENSE	Revista Ibero-Americana de Humanidades, Ciências e Educação	2,023	cc-by	5,809	Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE doi.org/10.51891/rease.v9i3.8882 1Mestre em Recursos Hídricos e Saneamento Ambiental; Universidade Federal do Rio Grande do Sul- UFRGS,Porto Alegre-RS. CONTRIBUIÇÃO AO DESENVOLVIMENTO SUSTENTÁVEL DO JALAPÃO TOCANTINENSE CONTRIBUTION TO THE SUSTAINABLE DEVELOPMENT OF JALAPÃO TOCANTINENSE Belizário Franco Neto1 RESUMO: Este artigo tem como finalidade propor um elenco de diretrizes para planejar e orientar a ocupação ordenada e econômica dos grandes espaços vagos e ociosos da região do Jalapão em bases sustentáveis. Para tanto, foi feita a caracterização da região, relacionando as necessidades para esta ocupação com o potencial produtivo e com os requerimentos de uso sustentável de seus recursos naturais. O estudo conclui que as diretrizes propostas possuem amplas possibilidades de se planejar e orientar a ocupação ordenada e econômica do Jalapão em bases sustentáveis, mas a decisão da materialização das diretrizes propostas em atividades específicas e a velocidade de implementação destas, dependerão de ações coordenadas e sinérgicas do poder público estadual e municipal e da iniciativa privada, sendo necessário que o Governo Estadual exerça sua função indutora, estimulando o investimento e criando as condições para tal implementação. 350 Palavras-chave: Desenvolvimento sustentável. Jalapão. Cerrado. Palavras-chave: Desenvolvimento sustentável. Jalapão. Cerrado ABSTRACT: This article aims to propose a list of guidelines to plan and guide the orderly and economical occupation of large vacant and idle spaces in the Jalapão region on a sustainable basis. To this end, a characterization of the region was conducted, relating the needs for this occupation with the productive potential and with the requirements for the sustainable use of its natural resources. The study concludes that the proposed guidelines have ample possibilities for planning and guiding the orderly and economic occupation of Jalapão on a sustainable basis, but the decision to materialize the proposed guidelines into specific activities and the speed of their implementation will depend on coordinated and synergistic actions of the state and municipal public power and the private initiative, being necessary that the State Government exercises its inducing function, stimulating the investment and creating the conditions for such implementation. Keywords: Sustainable development. Jalapão. Cerrado. Keywords: Sustainable development. Jalapão. Cerrado. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE 1. INTRODUÇÃO O Jalapão tem sido foco da atenção de ambientalistas do país inteiro para a conservação da biodiversidade. Na condição de natureza, ressalta-se a importância de constituir uma área fundamental para a conservação do cerrado, além de se configurar num polo de ecoturismo e turismo de aventura de âmbito nacional e internacional (BRASIL, 2013). Na condição de sociedade, alerta-se para os riscos que podem ser desencadeados com a utilização predatória da base natural do cerrado, o que certamente ameaçará os resultados de uma utilização mais qualificada de seus atributos naturais e locacionais, a curto, médio e longo prazo. O Jalapão abriga um grande conjunto de áreas protegidas no cerrado e um dos maiores blocos de vegetação nativa remanescente no Brasil Central. Estudos oficiais de seleção de áreas prioritárias para a conservação da biodiversidade indicam que a região do Jalapão é considerada de importância biológica extremamente alta, MMA (1999, 2006 apud BRASIL, 2013, p. 18). Deste modo, o Jalapão é uma das áreas críticas e prioritárias para a conservação do cerrado e, por extensão, uma região chave para a conservação da biodiversidade global (BRASIL, 2013). 351 Para Santos Filho (2016, p. 91) as características do Cerrado – topografia plana e clima estável com regime equilibrado de chuvas – favorecem o potencial produtivo da região. O Bioma Cerrado é considerado o celeiro do mundo (MEDEIROS, 2007). Por outro lado, o cerrado é a maior, mais diversa e mais ameaçada savana do planeta, representando uma das 34 regiões globais para a conservação da biodiversidade, por aliar altos níveis de endemismo e alta pressão de perda de habitat, Myers et al (2000). Para PARRON et al (2008, p. 107) o uso sustentável é altamente recomendável, pois prevê o planejamento das operações em determinada área, garantindo a perenidade dos recursos naturais. De acordo com o Sistema Nacional de Unidades de Conservação da Natureza (SNUC) (BRASIL, 2002), o uso sustentável é “a exploração do ambiente de maneira a garantir a perenidade dos recursos ambientais renováveis e dos processos ecológicos, mantendo a biodiversidade e os demais atributos ecológicos, de forma socialmente justa e economicamente viável”. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Para Elkington (2001), desenvolvimento sustentável pode ser entendido como a harmonia entre a questão econômica e ambiental. 1. INTRODUÇÃO A ex-Primeira Ministra da Noruega Gro Harlem Brundtland, cita outra definição para o termo: “é a forma como as atuais gerações satisfazem as suas necessidades sem, no entanto, comprometer a capacidade de gerações futuras satisfazerem as suas próprias necessidades” Brundtland (1991 apud SCHARF, 2004, p.19). Camargo (2003, p. 43) traz um conceito bem mais completo e complexo: Camargo (2003, p. 43) traz um conceito bem mais completo e complexo: Em essência, o desenvolvimento sustentável é um processo de transformação no qual a exploração dos recursos, a direção dos investimentos, a orientação do desenvolvimento tecnológico e a mudança institucional se harmonizam e reforçam o potencial presente e futuro, a fim de atender às necessidades e aspirações humanas. Assim sendo, o Jalapão carece de uma ação proativa, por parte do poder público estadual e municipal, para evitar a exploração predatória dos recursos naturais e a aceleração do processo de degradação ambiental ao tempo que promova um desenvolvimento assentado nas atividades econômicas dinâmicas e sustentáveis que possam garantir a preservação da região. Dessa forma, este estudo tem como finalidade propor um elenco de diretrizes para planejar e orientar a ocupação ordenada e econômica dos grandes espaços vagos e ociosos da região em bases sustentáveis, com vistas ao incremento da produção e da produtividade, a geração de emprego, ao aumento da renda dos produtores e elevação da qualidade de vida da população. 352 2. CARACTERIZAÇÃO DA REGIÃO A região possui uma área territorial de 34.243,55 km2, que representa cerca de 12,34% do território estadual (IBGE, 2021a). Sua localização no extremo leste do Estado do Tocantins, permite limitação geográfica com os Estados da Bahia, Piauí e Maranhão. Constitui uma das dezoito (18) regiões administrativas em que o estado foi dividido com base em critérios fisiográficos, político-administrativos e socioeconômicos; para efeito de planejamento, ordenamento territorial, subsídio à elaboração de políticas públicas e formulação de estratégias operacionais de desenvolvimento regional (TOCANTINS, 2012) Os oito municípios que integram a região do Jalapão, Lagoa do Tocantins, Lizarda, Mateiros, Novo Acordo, Ponte Alta do Tocantins, Rio Sono, Santa Tereza do Tocantins e São Félix do Tocantins, totalizam uma população estimada em 34.648 habitantes, apenas 2,15% da população estadual (IBGE, 2021b), e revela a menor densidade demográfica média regional do estado com 1,01 habitantes/km2. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Figura 1. Localização da Região do Jalapão no Tocantins e no Brasil. Fonte: elaborado pela SEPLAN-DIGITS, com base no Atlas Tocantins (2012). 353 Figura 1. Localização da Região do Jalapão no Tocantins e no Brasil. Figura 1. Localização da Região do Jalapão no Tocantins e no Brasil. Fonte: elaborado pela SEPLAN-DIGITS, com base no Atlas Tocantins (2012). Figura 1. Localização da Região do Jalapão no Tocantins e no Brasil. Fonte: elaborado pela SEPLAN-DIGITS, com base no Atlas Tocantins (2012) g ç g J p Fonte: elaborado pela SEPLAN-DIGITS, com base no Atlas Tocantins (2012). O regime de chuvas prevalecente, com uma precipitação média anual no entorno de 1.550 mm, distribuídos em sua maioria de outubro a abril (7 meses) e a temperatura média anual é de 26 °C com variações durante o ano de mínima absoluta de 10,6 °C., que coincide com o período seco, e máxima absoluta de 39,6 °C. A média anual da umidade relativa do ar gira em torno de 60% a 70%, em termos gerais, favorecem o desenvolvimento agropecuário. (TOCANTINS, 2012). Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. 2. CARACTERIZAÇÃO DA REGIÃO Em relação aos recursos hídricos, a região que, em sua maior parte, integra a Bacia do Tocantins, conta com uma ampla e bem distribuída rede de rios e córregos permanentes, dos quais os mais importantes, pelo volume de água e extensão, são os rios Sono, Novo, Galhão, Prata, Soninho, Vermelho, Ponte Alta e Caracol (TOCANTINS, 2003). O recurso água é abundante e Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE o sistema fluvial mostra os processos gerados pelas águas canalizadas em termos da cristalinidade e pureza dos rios e da ausência das cargas de fundo e em suspensão. A cobertura vegetal predominante em quase toda a região é do tipo savana nas diversas formas que são características dos cerrados e campos brasileiros. Observa-se, ainda, o registro de matas ciliares ao longo das margens dos diversos rios e córregos que integram as pequenas bacias hidrográficas, em faixas relativamente estreitas em ambas as margens. Os cerrados apresentam-se tanto na categoria de cerrado denso como de cerrado ralo com a vegetação arbustiva e a densidade típicas de cada caso. A formação de campos, por sua vez, é dominada principalmente, pelas gramíneas nativas (TOCANTINS, 2012). O Jalapão abrange depressões resultantes de processos de recuo das escarpas da Serra Geral e da Chapada das Mangabeiras, onde podem ser observados alguns testemunhos da história natural regional como as Serras da Muriçoca, da Estiva, do Espírito Santo, da Jalapinha, entre outros, com a predominância de um relevo plano a suave ondulado, ocorrendo na interseção dessas serras (TOCANTINS, 2003). Os solos da região são, basicamente, provenientes da intemperização dos arenitos cretácicos da formação Urucuia. Em termos gerais, consistem em solos arenosos que se apresentam em duas categorias predominantes: Neossolos Quartzarênicos, que ocorrem na maior parte da região, e os Latossolos. Os solos Argissolos, Plintossolos e os Neossolos Litólicos, basicamente, complementam a configuração edáfica da região (TOCANTINS, 2012). 354 Para Fernandes et al (1982): Para Fernandes et al (1982): Englobam, sob a denominação de Formação Urucuia, todos os sedimentos cretáceos, inclusive a sua sequência basal, constituída de conglomerados, siltitos, argilitos e folhelhos que em mapeamentos anteriores do RADAMBRASIL, foram destacados como pertencentes à Formação Areado. De acordo com Lima (2000, p. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 2. CARACTERIZAÇÃO DA REGIÃO 1) os Neossolos Quartzarênicos têm nas microrregiões de Jalapão, Bico do Papagaio e Araguaína sua principal ocorrência, apresentam-se associados e com inclusões de Latossolos Vermelho-Amarelos, Latossolos Amarelos, Plintossolos Pétricos litoplínticos (Laterita Hidromórfica), Pétricos concrecionários (Solos Concrecionários) e Podzólicos (Argissolos). As unidades de conservação (estaduais e federais) localizadas no Jalapão estendem-se por cerca de 14,3 mil km², perfazendo 41,76% da região, sendo que as UCs de proteção integral cobrem 9,1 mil km² e as de uso sustentável totalizam 5,2 mil km². Estas áreas apresentam grandes Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE extensões de cerrado nativo e pouco impactadas em função das respectivas condições especiais de uso e ocupação (TOCANTINS, 2012). extensões de cerrado nativo e pouco impactadas em função das respectivas condições especiais de uso e ocupação (TOCANTINS, 2012). 3. DISCUSSÃO PARRON et al (2008, p. 263) dizem que: O cerrado constitui um enorme bioma, apresentando grande diversidade de recursos naturais e riqueza de conhecimentos de suas populações tradicionais. A região representa, ao mesmo tempo, a área de expansão da fronteira agrícola do país em sistemas de produção, nos quais predominam os monocultivos em grandes extensões de área, uso intensivo de insumos químicos, máquinas e irrigação. As questões e urgências ambientais estão emergindo há algumas décadas e, no cerrado, confronta-se, dia após dia, o dilema da conservação versus desenvolvimento agrícola. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 PARRON et al (2008, p. 263) dizem que: ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Santa Tereza Tocantins 11.356,04 10.969,81 297,44 10.371,77 9.844,42 São Félix Tocantins 14.022,52 13.860,33 86,52 11.121,13 9.938,75 Total Região 106.858,20 102.253,44 3.092,78 94.696,37 77.369,37 Fonte: IBGE (2017) inclui ICMS e o Fundo de Participação dos Municípios Rev inclui ICMS e o Fundo de Participação dos Municípios inclui ICMS e o Fundo de Participação dos Municípios A Tabela 1 dá margem aos seguintes comentários: (a) o Fundo de Participação dos Municípios (transferências do Governo Federal) constitui a principal fonte de receita da região; (b) as receitas tributárias municipais, em termos relativos, são irrelevantes; (c) as transferências constitucionais, em 2021, corresponderam, em média, a 95,69% das receitas orçamentárias dos municípios. Vale aqui ressaltar, para efeito comparativo, que o Jalapão ocupa 12,34% do território de um estado cuja economia depende, fundamentalmente, do setor agropecuário. A estrutura fundiária do Jalapão, analisada segundo as terras cadastradas no INCRA (Instituto de Colonização e Reforma Agrária), revela a existência de um pouco mais de 2.600 imóveis rurais ocupando uma extensão de 1.800.000 hectares, o que dá uma média de 700 hectares por imóvel. Considerando a área total da região, da ordem dos 3,4 milhões de hectares, as cifras precedentes indicam a existência de cerca de 1,6 milhões de hectares que, de alguma forma, encontram-se em situação irregular ou com domínio por definir-se. 356 A produção agropecuária está localizada, basicamente, nas áreas com melhor aptidão agrícola, ou seja, onde predominam os Latossolos. Nas manchas desses solos, entremeadas com areias quartzosas, que ocorrem nos municípios de São Félix do Tocantins, Mateiros, Lizarda e grande parte de Ponte Alta do Tocantins, as atividades produtivas, com exceção da soja e do milho, são realizadas por pequenos produtores, agrupados em comunidades, com base em uma agricultura do tipo familiar, sem o uso dos denominados insumos modernos e, portanto, com baixos índices de produção. Lima (2023, p. 975) poucos são os incentivos às famílias de zonas rurais para o desenvolvimento de atividades em campo, o que influência diretamente na baixa quantidade de famílias que realizam o cultivo familiar em suas propriedades. PARRON et al (2008, p. 263) dizem que: O cerrado constitui um enorme bioma, apresentando grande diversidade de recursos naturais e riqueza de conhecimentos de suas populações tradicionais. A região representa, ao mesmo tempo, a área de expansão da fronteira agrícola do país em sistemas de produção, nos quais predominam os monocultivos em grandes extensões de área, uso intensivo de insumos químicos, máquinas e irrigação. As questões e urgências ambientais estão emergindo há algumas décadas e, no cerrado, confronta-se, dia após dia, o dilema da conservação versus desenvolvimento agrícola. A região do Jalapão está inserida em um ambiente ecológico que lhe confere, no âmbito estadual, uma característica de singularidade entre as demais regiões do estado. O fator determinante dessa singularidade é o processo eólico de erosão e degradação ambiental, presente na região, redundando na predominância de categorias de solos frágeis, particularmente os Neossolos Quartzarênicos (Areias Quartzosas), na maior parte de seu território, tendo sido identificada como afetada por degradação ambiental, razão pela qual foi classificada e mapeada como “Área de Atenção Especial” (TOCANTINS, 2003). 355 A hidrografia e a relativa abundância de água no Jalapão sinalizam o grande potencial de agricultura irrigada, cuja prática poderá constituir-se em preponderante fator de desenvolvimento do agronegócio regional, desde que sejam adotadas tecnologias que assegurem a preservação ambiental e a conservação dos recursos naturais, particularmente, dos solos. Entretanto, o modestíssimo nível de desenvolvimento econômico dos municípios que compreendem a região fica demonstrado por sua mínima contribuição ao sistema tributário do estado. Em 2021 a arrecadação do ICMS na região alcançou tão somente a importância de R$ 2.933.860,43, equivalente a 0,07% da arrecadação estadual (TOCANTINS, 2021). Tabela 1 - Receitas e Despesas realizadas pelos municípios da Região do Jalapão, em 2017 (x 1000) R$ Municípios Receitas orçamentárias Transferências Receitas tributárias Despesas totais despesas correntes Lizarda 11.668,26 11.072,32 483,91 9.047,72 8.598,56 Novo Acordo 12.542,51 12.074,02 327,62 11.889,07 11.357,27 Ponte Alta do Tocantins 17.447,67 16.623,21 491,00 15.192,67 14.749,74 Lagoa do Tocantins 11.608,54 11.305,30 45,35 11.608,54 8.651,64 Mateiros 13.242,97 12.639,22 557,96 11.266,67 10.945,77 Rio Sono 14.969,69 13.709,23 1.186,94 14.198,80 13.221,97 1 - Receitas e Despesas realizadas pelos municípios da Região do Jalapão, em 2017 (x 1000) R$ Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 PARRON et al (2008, p. 263) dizem que: r outro lado, diz ainda, a autora: Por outro lado, diz ainda, a autora: A agricultura familiar sem dúvida é uma atividade capaz de aumentar a produção de alimentos, e ser economicamente viável, com práticas responsáveis com o meio ambiente, favorecendo ao mercado crescente de produtos saudáveis e frescos, fornecidos diretamente dos produtores. (Lima 2023, p. 974) Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Em relação à pecuária bovina, as informações sob análise parecem indicar ser esta uma das principais atividades econômica da região. É desenvolvida extensivamente, normalmente, para a produção de bezerros destinados a engorda em outras zonas do estado. O efetivo do rebanho da região, em 2021, era de 234.771 cabeças (IBGE, 2021c). Os principais cultivos regionais são soja, milho, mandioca e arroz. A produção rural, correspondente aos principais produtos da região pode ser observada na Tabela 2 abaixo. No município de Mateiros existem grandes projetos de produção de soja e milho com elevado nível tecnológico e altos índices de produtividade e de produção. Nos municípios de Rio Sono, Novo Acordo, Santa Tereza do Tocantins, Lagoa do Tocantins e parte do município de Ponte Alta do Tocantins, onde estão concentradas as áreas contíguas mais extensas de Latossolos, além da agricultura de tipo familiar típica da região, alguns projetos com bom grau de tecnificação, de tamanho médio, estão instalados e/ou em fase de instalação para a produção de grãos e fruticultura tropical. Tabela 2 - produção agrícola da região do Jalapão Produtos 2021 Área (há) Produção (ton) Rend (kg/ha) Região % s/ Estado Soja 74.093 184.892 5,04 2.495 Milho 9.786 43.513 2,88 4.446 Mandioca 574 9.066 3,41 15.794 Arroz 2.317 4.963 0,73 2.141 Feijão 1.324 878 1,67 663 Abacaxi 93 1.870 1,92 20.107 Melancia 69 1.350 0,93 19.565 Sorgo 366 980 1,28 2.722 Cana de açúcar 6 240 0,01 40.000 Fonte: IBGE (2021c) Rendimento frutos/há 357 Os Neossolos Quartzarênicos constituem solos frágeis com baixa capacidade de agregação de partículas devido aos reduzidos teores de argila e matéria orgânica, apresentando uma limitada capacidade de retenção de água, pouca espessura, de baixa fertilidade natural e com forte propensão à lixiviação. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 PARRON et al (2008, p. 263) dizem que: É de se assinalar que as recomendações de caráter técnico-científico relacionadas com o aproveitamento desses solos, que cobrem uma grande área da região, apontam para: (i) a necessidade de proteção permanente da cobertura vegetal do solo; (ii) a restrição ao uso de Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE mecanização; e (iii) uma maior aptidão do solo para o cultivo de plantas perenes (Embrapa, 2021a). Normalmente, ocorrem nos setores de menor altitude da região e sua topografia é plana ou ligeiramente ondulada. No caso do Jalapão, dadas as condições de precipitação, luminosidade, temperatura, umidade relativa do ar e de abundância de recursos hídricos, admite-se que os Neossolos Quartzarênicos poderiam ser aproveitados economicamente para a produção de algumas espécies frutícolas tropicais e para o florestamento. O teor de argila nos Neossolos Quartzarênicos da região varia segundo a maior ou menor incidência da erosão eólica em espaços específicos. A formação de dunas já é registrada em alguns setores mais afetados. Essa categoria de solos predomina nos municípios de Mateiros, São Félix do Tocantins, Lizarda e Ponte Alta do Tocantins, onde se pode observar grandes vazios espaciais em termos de população e exploração agropecuária. Os Latossolos da região apresentam-se basicamente nas formas de Vermelho-Amarelo e Amarelo, típicos da grande região dos cerrados brasileiros, Lima (2000, p. 1). São solos de textura média, geralmente profundos, ácidos e friáveis. O elevado grau de intemperismo dos Latossolos é refletido em valores muito baixos de elementos nutrientes, representados pela soma e saturação de bases. Por outro lado, apresentam boa drenagem interna, condicionada por elevada porosidade e homogeneidade de características ao longo do perfil e, em razão disto, elevada permeabilidade. Este fato os coloca, quando em condições naturais, como solos de razoável resistência à erosão de superfície (laminar e sulcos), (Embrapa, 2021b). 358 Ao contrário dos Neossolos Quartzarênicos, os Latossolos localizam-se, normalmente, nas áreas de altitude mais elevada e seu uso para fins de produção agrícola comercial está condicionado à correção da acidez com calcário, à adubação química e a práticas adequadas de manejo do solo. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. SSN 26 5 33 5 PARRON et al (2008, p. 263) dizem que: Enquanto o Turismo de Aventura é um seguimento do turismo que envolve atividades de aventura de caráter recreativo e não competitivo, com risco avaliado” (TURISMO DE AVENTURA, 2009). 359 PARRON et al (2008, p. 263) dizem que: Os Latossolos predominam em áreas contíguas, em maior extensão, nos municípios de Rio Sono, Novo Acordo, Santa Tereza do Tocantins, Lagoa do Tocantins e Ponte Alta do Tocantins, zona mais próxima ao Rio Tocantins onde estaria o potencial produtivo mais imediato da região (TOCANTINS, 2012). Ressalta-se que, ao longo dos rios e córregos, inclusive onde há predominância de Neossolos Quartzarênicos, normalmente despontam várzeas, cujos solos apresentam maior teor de argila e matéria orgânica e, consequentemente, revelam maior fertilidade. Embora não se disponha de dados precisos de análise desses solos e se ignore a proporção existente na região, admite-se que as várzeas, devidamente drenadas, poderiam servir de base para o fomento de produtos alimentares, , ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE evista Ibero- Americana de Humanidades, Ciências e Educação- REASE particularmente de grãos. Nessa eventualidade, faz-se necessária uma análise mais profunda das implicações ambientais decorrentes da drenagem desses solos para dar margem a um aproveitamento econômico. Os Plintossolos e Neossolos Litólicos, por suas características físico-químicas, não são, em termos gerais, recomendados para cultivos agrícolas. Observadas as especificidades locais, aparentemente oferecem condições para pastagens e florestamento. Enquanto ao Argissolos apresentam maior potencial para uso agrícola exigindo um manejo adequado com a adoção de correção, adubação e de práticas conservacionistas para o controle da erosão (Embrapa, 2021c). (Souza, 2022, p. 9) destaca que: (Souza, 2022, p. 9) destaca que: Atualmente, entre as sete (07) regiões turísticas do estado instituídas pelo Programa de Regionalização do Governo Federal, o Jalapão é conhecido como o principal destino turístico do Tocantins. Considerando as características físicas da região, destacam-se os segmentos de Ecoturismo e Turismo de Aventura. As belas paisagens constituem-se em um conjunto de atrativos turísticos de belezas singulares, como as veredas de capim dourado, as dunas de areias avermelhadas, as cachoeiras de águas cristalinas e fervedouros que encantam por suas peculiaridades. Essas características físicas contribuíram para a região já ter sido cenário de filmes, novelas, seriados e reality shows. O Ecoturismo, segundo as Diretrizes para a Política Nacional de Ecoturismo é: um segmento da atividade turística que utiliza de forma sustentável o patrimônio natural e cultural, incentiva sua conservação e busca a formação de uma consciência ambientalista através da interpretação do ambiente, promovendo o bem estar das populações envolvidas (BRASIL, 1994). Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 4. MATERIAIS E MÉTODOS O estudo para formulação das diretrizes grais foi desenvolvido com base na revisão e interpretação das informações disponíveis e na opinião técnica de especialistas em áreas críticas do conhecimento necessário para se promover o desenvolvimento sustentável da região. Nenhum estudo específico de viabilidade foi elaborado. As diretrizes tomaram a presente forma com a realização dos exercícios iniciais de pré-viabilidade, com os quais se tratou de formulá-las e determinar a conveniência da sua materialização em ações específicas e de se avançar nos estudos mais aprofundados que demonstrarão a viabilidade técnica, econômica, financeira, ambiental e institucional para a implementação destas. Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE No processo de sua formulação prevaleceu a preocupação de se dar a seu conteúdo um nível técnico e científico de consistência e coerência que guardasse relação com a vulnerabilidade e a dimensão que caracterizam o Jalapão. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 5. RESULTADOS A partir de uma visão holística e sistêmica, apresenta-se, abaixo, a possibilidade de planejamento da ocupação das terras e da produção, envolvendo pequenos produtores radicados nas comunidades locais e implantando um modelo empresarial para silvicultura e fruticultura tropical, através da organização de produtores e assistência técnica com qualidade e quantidade, bem como, da expansão do ecoturismo e do turismo de aventura, tendo em vista as características ímpares da região e as suas belezas naturais, considerando a disponibilidade de recursos naturais e as limitações ambientais. A estratégia de ocupação ordenada e econômica do Jalapão em bases sustentáveis está refletida nas seguintes diretrizes: A estratégia de ocupação ordenada e econômica do Jalapão em bases sustentáveis está refletida nas seguintes diretrizes: a) promover o desenvolvimento de estudos ambientais visando a elaboração de uma base cartográfica composta de mapeamentos temáticos, dados de levantamento fundiário (cartorial e ocupacional), dados das condições socioeconômicas das comunidades locais, instalação de estações meteorológicas e da identificação, dimensionamento e seleção de polos para validação de tecnologias; 360 b) promover a pesquisa científica e a validação de tecnologias em áreas de associações de Neossolos Quartzarênicos, Latossolos e solos Podzólicos direcionadas para silvicultura e fruticultura, a partir da: (i) implementação de um plano de pesquisas aplicadas e/ou validação de tecnologias, desdobrado em função dos tipos de solos e da necessidade de conhecimentos relacionados com a compatibilização das atividades de desenvolvimento com a conservação de recursos naturais e preservação da biodiversidade e de um sólido suporte técnico, de retaguarda, das instituições de ensino e pesquisa (estaduais e federais), do órgão estadual de assistência técnica e extensão rural e da Empresa Brasileira de Pesquisa Agropecuária - EMBRAPA; e (ii) implantação de módulos experimentais e unidades de observação em cada tipo de solo em parceria com produtores; Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE c) promover a implementação de ações ambientais a fim de que seja alcançada a harmonia entre a conservação e preservação da região com o seu desenvolvimento econômico para fins de produção comercial. 5. RESULTADOS d) fortalecer as prefeituras municipais da região a fim torná-las mais eficientes na administração dos recursos públicos, na prestação de serviços às populações rurais e urbanas e no gerenciamento do ecoturismo e do turismo de aventura compreendendo os seguintes segmentos: (i) base tributária municipal; (ii) implantação do serviço de educação e supervisão ambiental; (iii) construção e manutenção de estradas vicinais e de pequenas pontes; (iv) atividades potenciais de turismo sustentável e sua viabilidade para implementação; (v) capacidade de carga do local com vista à futuros investimentos; (vi) uso da terra como base para implementação de áreas tampão de desenvolvimento, de serviços, e de minimização de impactos socioambientais negativos; e (vii) infraestrutura para o suprimento de água e condições sanitárias, energia sustentável e facilidades de recepção turística; d) fortalecer as prefeituras municipais da região a fim torná-las mais eficientes na administração dos recursos públicos, na prestação de serviços às populações rurais e urbanas e no gerenciamento do ecoturismo e do turismo de aventura compreendendo os seguintes segmentos: (i) base tributária municipal; (ii) implantação do serviço de educação e supervisão ambiental; (iii) construção e manutenção de estradas vicinais e de pequenas pontes; (iv) atividades potenciais de turismo sustentável e sua viabilidade para implementação; (v) capacidade de carga do local com vista à futuros investimentos; (vi) uso da terra como base para implementação de áreas tampão de desenvolvimento, de serviços, e de minimização de impactos socioambientais negativos; e (vii) infraestrutura para o suprimento de água e condições sanitárias, energia sustentável e facilidades de recepção turística; e) Promover a ocupação das terras e fomentar a produção: 361 1) em áreas de associações de Neossolos Quartzarênicos baseados na implantação de atividades de silvicultura (florestamento) e de fruticultura tropical, com a implantação de um modelo empresarial com unidades de produção de tamanhos médio e grande; 2) em áreas de associações de Latossolos envolvendo pequenos produtores radicados nas comunidades locais que praticam a agricultura familiar, estimulando a sua organização, fornecendo assistência técnica de qualidade, fomentando à produção de alimentos básicos, com base na agricultura diversificada e fortalecendo os agricultores para o acesso as linhas de crédito rural existentes no país; e 3) em áreas de Latossolos em associações com Plintossolos e Agissolos localizados na parte oeste do Jalapão, orientada, preferencialmente, ao desenvolvimento da fruticultura, com elevado nível tecnológico, envolvendo médios e grandes produtores rurais. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. 5. RESULTADOS ISSN - 2675 – 3375 Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE f) promover a capacitação da população local para prestação de serviços de condutor de turismo local, monitor ambiental, garçom/garçonete, camareira e auxiliar de cozinha; g) promover workshops especialmente orientados para melhores práticas em relação a desenho arquitetônico e arquitetura de serviços, gerenciamento, conservação do meio ambiente e, especialmente, o uso de tecnologias apropriadas direcionados a investidores do setor de turismo; h) promover a educação ambiental e turística visando a elevação do nível de informações da população local sobre o valor de um turismo sustentável e a construção da consciência ambiental; i) ampliar a infraestrutura econômica - energia, armazenagem, transformação, transportes, e de prestação de serviços essenciais à qualidade de vida de seus habitantes – saneamento básico, destinação de resíduos sólidos, saúde, educação e segurança pública, de abrangência regional e municipal. CONSIDERAÇÕES FINAIS 362 As diretrizes propostas possuem amplas possibilidades para se planejar e orientar a ocupação ordenada e econômica do Jalapão em bases sustentáveis. Contudo, a materialização destas diretrizes em atividades específicas e a implementação destas, dependerão de ações coordenadas e sinérgicas do poder público estadual e municipal e da iniciativa privada, sendo necessário que o Governo Estadual exerça sua função indutora, estimulando o investimento e criando condições para tal implementação, que certamente contribuirá para que seja logrado o equilíbrio espacial no processo de desenvolvimento do Estado, pois com efeito, incorporaria uma região geograficamente bem localizada, ecologicamente frágil e até agora apenas assistida com ações isoladas, ao dinâmico processo de desenvolvimento do Estado. REFERÊNCIAS BRASIL. Instituto Chico Mendes de Conservação da Biodiversidade. Projeto Corredor Ecológico da Região do Jalapão. Brasília-DF, 2013. BRASIL. Instituto Chico Mendes de Conservação da Biodiversidade. Projeto Corredor Ecológico da Região do Jalapão. Brasília-DF, 2013. BRASIL. Ministério da Indústria, Comércio e Turismo. Ministério do Meio Ambiente. Diretrizes para uma política nacional de ecoturismo. Brasília, DF: EMBRATUR/IBAMA, 1994. BRASIL. Ministério da Indústria, Comércio e Turismo. Ministério do Meio Ambiente. Diretrizes para uma política nacional de ecoturismo. Brasília, DF: EMBRATUR/IBAMA, 1994. CAMARGO, Aspásia. Governança para o século 21. Meio ambiente no século, v. 21, n. 21, p. 307- 21, 2003. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE DE SOUZA, Maria Antônia Valadares; GRÁCIO, Héber Rogério; CANÇADO, Airton Cardoso. CONFLITO SOCIOAMBIENTAL DO TURISMO NO PARQUE ESTADUAL DO JALAPÃO-TO. Revista Baru-Revista Brasileira de Assuntos Regionais e Urbanos, v. 8, n. 1, p. 18, 2022. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 ELKINGTON, John. Canibais com Garfo e Faca. São Paulo: Makron Books, 2001 EMBRAPA(2021a). Empresa Brasileira de Pesquisa Agropecuária. Território Mara Sul Pernambucana. Disponível em:<https://www.embrapa.br/agencia-de-informacao- tecnologica/territorios/territorio-mata-sul-ernambucana/caracteristicas-doterritorio/recursos- naturais/solos/neossolos-quartzarenicos>. Acesso em: 6 mar. 2023. EMBRAPA(2021b). Empresa Brasileira de Pesquisa Agropecuária. Bioma Cerrado. Disponível em: <https://www.embrapa.br/agencia-de-informacao-tecnologica/tematicas/bioma- cerrado/solo/tipos-de-solo/latossolos>. Acesso em: 6 mar. 2023. EMBRAPA(2021c). Empresa Brasileira de Pesquisa Agropecuária. Disponível em: <https://www.embrapa.br/agencia-de-informacao-tecnologica/tematicas/solos- tropicais/sibcs/chave-do-sibcs/argissolos>. Acesso em: 6 mar. 2023. FERNANDES, P.E.C.A. et al. Geologia. Em: Departamento Nacional de Produção Mineral. Projeto 363 IBGE(2017). Instituto Brasileiro de Geografia e Estatística. Finanças Públicas. Disponível em: <https://cidades.ibge.gov.br/brasil/panorama>. Acesso em: 28 fev 2023. IBGE(2021b). Instituto Brasileiro de Geografia e Estatística. Cidades e Estados. População Estimada. Disponível em: < https://www.ibge.gov.br/cidades-e-estados/to/>. Acesso em: acesso 20 fev 2023 IBGE(2021c). Instituto Brasileiro de Geografia e Estatística. Produção Agrícola Municipal. Disponível em: <https://cidades.ibge.gov.br/brasil/panorama>. Acesso em: 28 fev 2023. IBGE(2021a). Instituto Brasileiro de Geografia e Estatística. Cidades e Estados. Área Territorial. Disponível em: < https://www.ibge.gov.br/cidades-e-estados/to/>. Acesso em: 27 fev 2023. LIMA, Antônio Agostinho C.; OLIVEIRA, Francisco Nelsieudes S.; DE AQUINO, Antônio Renes Lins. Aptidão agrícola dos solos do Estado do Tocantins. 2000. LIMA, Raquel Aparecida Mendes. CONTRIBUIÇÃO TOCANTINENSE PARA A AGRICULTURA FAMILIAR. Revista Ibero-Americana de Humanidades, Ciências e Educação, v. 9, n. 2, p. 972-997, 2023. MYERS, Norman et al. Biodiversity hotspots for conservation priorities. Nature, v. 403, n. 6772, p. 853-858, 2000. Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero-Americana de Humanidades, Ciências e Educação. São Paulo, v.9.n.03. mar. 2023. ISSN - 2675 – 3375 Revista Ibero- Americana de Humanidades, Ciências e Educação- REASE PARRON, Lucilia Maria et al. Cerrado: desafios e oportunidades para o desenvolvimento sustentável. Planaltina, DF: Embrapa Cerrados, 2008., 2008. 464p RADAMBRASIL. Folha SD. 23 Brasília. Rio de Janeiro, 1982. SANTOS FILHO, Jonas Irineu dos et al. Potencial do Matopiba na produção de aves e suínos. Revista de Política Agrícola, v. 25, n. 2, p. 90-102, 2016. SCHARF, Regina. Manual de Negócios Sustentáveis. São Paulo, Amigos da Terra, 2004. CHARF, Regina. Manual de Negócios Sustentáveis. São Paulo, Amigos da Terra, 200 TOCANTINS, Governo do Estado. Secretaria da Fazenda. Relatório de Arrecadação por Município e Atividade Econômica. Palmas-TO, 2021. TOCANTINS, Governo do Estado. Secretaria do Planejamento e da Modernização da Gestão Pública. Atlas do Tocantins. Palmas-TO,2012. TOCANTINS, Governo do Estado. Secretaria do Planejamento e Meio Ambiente. Plano de Manejo do Parque Estadual do Jalapão. Palmas-TO, 2003. TURISMO DE AVENTURA. – [Brasília]: o Ministério: Florianópolis: SEAD/UFSC, 368 p. 2009. 364
https://openalex.org/W4230793516	https://nordiskbarnehageforskning.no/index.php/nbf/article/download/57/57	Danish	null	De pædagogiske læreplaners reformering: Dokumentation, faglighed og måder at gøre pædagog	Nordisk barnehageforskning	2,013	cc-by	11,364	nordisk barnehageforskning 2012 Nordisk Børnehaveforskning Norrænar Leikskólarannsóknir Nordic Early Childhood Education Research Pohjoismainen Varhaiskasvatustutkimus Nordisk Förskoleforskning www.nordiskbarnehageforskning.no issn 1890-9167 vol 5 nr 12 sid 1–12 Nordisk Børnehaveforskning Norrænar Leikskólarannsóknir Nordic Early Childhood Education Research issn 1890-9167 tema markedsliberalisme, professionshøjskoler og pædagoguddannelse De pædagogiske læreplaners reformering Dokumentation, faglighed og måder at gøre pædagog Maja Plum Institut for Medier, Erkendelse og Formidling, Københavns universitet, Danmark Maja Plum Institut for Medier, Erkendelse og Formidling, Københavns universitet, Danmark et al., 2009, p. 11f). I Danmark udgør Pædagogi- ske læreplaner et eksempel på en sådan reform. I 2004 indførtes således de pædagogiske lærepla- ner som en national reform af førskole området. dagog på. En måde der organiserer og indhegner bestemte elementer af daginstitutionel hverdag som ’fagligt’ pædagogisk arbejde, mens en ræk- ke andre elementer og måder at gøre pædagog på udgrænses fra denne faglighedsdefinition. et al., 2009, p. 11f). I Danmark udgør Pædagogi- ske læreplaner et eksempel på en sådan reform. I 2004 indførtes således de pædagogiske lærepla- ner som en national reform af førskole området. En reform der pålægger den enkelte daginstituti- on at opsætte mål og beskrive metoder inden for seks læringstemaer, nemlig temaerne sprog; so- ciale kompetencer; personlige kompetencer; naturen og naturfænomener; kulturelle udtryks- former og værdier; krop og bevægelse (Socialmi- nisteren, 2004), samt ikke mindst dokumentere hvorledes disse mål bliver opnået. Reformen hvi- ler på en argumentation om, at et sådan krav om dokumentation netop ikke må forstås som en egentlig reformering i betydningen forandring af daginstitutionsområdet – dets traditioner og værdier. Kravet om dokumentation udgør inden for reformens egen argumentation ikke et pålæg om et bestemt indhold, men enghed for at skabe viden om det pædagogiske indhold og gennem denne viden gøre dette arbejde synligt samt på denne basis skabe såkaldt ’faglig’1 udvikling. At opsætte mål inden for de seks læringstemaer, be- skrive metoder og dokumentere, hvorledes disse opnås, betragtes således som en vidensproces, der skal udvikle pædagogernes faglighed ved at arbejdet bliver mere ’målrettet’ og ’systematisk’, og at den enkelte pædagog bliver mere ’bevidst’ og ’reflekteret’2. For at bryde med den vidensforståelse, der gør sig gældende i læreplanstiltagets egen argumen- tation, henter jeg inspiration fra en anden måde at forstå viden, hvori viden netop ikke er bundet til den enkelte pædagogs bevidsthed. Begrebet at gøre pædagog henter således inspiration fra Ak- tør-Netværks-Teorien (ANT). Min analytiske indgang indebærer her et brud, der ligger i selve skrivemåden. Den konsekvente understregning af at noget gøres er udtryk for dette brud. Gen- nem en ANT indgang sætter jeg fokus på pæda- gog, ikke som en fast størrelse, men som et fænomen der får sin karakter gennem de sam- menhænge eller forbindelser, hun3 indgår i. Der- ved bryder jeg med dominerende forestillinger om, hvad viden, faglighed og pædagog vil sige. Maja Plum Institut for Medier, Erkendelse og Formidling, Københavns universitet, Danmark Title: The Reform of Educational Plans: Documentation, professionalism and the enactment of nursery teaching. Abstract: In 2004 the demand of documentation in Danish pre-school was made compulsory and national through the reform of educational plans. One of the arguments was that documentation would enhance the ‘pro- fessional’ development of nursery teachers and make their daily work visible and recognized. Drawing on per- spectives from Actor-Network-Theory this article analyses the way in which documentation is enacted as a network of heterogeneous elements within the individual pre-school. It is argued that documentation, rather than developing ‘professionalism’, enacts the nursery teacher in a particular ‘professional’ way. Thus, it is analyzed what ‘professionalism’ becomes, and what is excluded – produced as outside of ‘professional work’. Keywords: Documentation; Educational reform; Governing; Actor-Network-Theory Email: mplum@hum.ku.dk p Peer-reviewed article: Sent to reviewers March 2011, Accepted May 2012, Preprint 22 October 2012, Published 16 April 2012 as a part of a special edition in Nordisk Barnehageforskning 2013. humankapital (Skolverket, 2004, p. 8ff; Østrem, et al., 2009, p. 11f). I Sverige har man i 1998 iværksat nationale læreplaner, og i Norge har man i 2006 reformeret den gældende rammeplan på daginstitutionsområdet med en yderligere vægtning af læring og dokumentation. I begge lande er området overført til henholdsvis Udan- nelsesministeriet og Kunnskapsdepartementet, og reformerne kan netop forstås som en ambiti- on om at gøre daginstitutioner til en del af det samlede uddannelsesforløb. Reformerne begrun- des således i forestillinger om livslang læring, bedre overgang til skolen samt højnelse af lan- dets økonomiske konkurrencekraft (NOU, 2011, p. 330f; Skoleverket, 2004, p. 8ff; Østrem, Som så mange andre lande er Danmark ramt af en reformiver på uddannelsesområdet. En re- formiver der gennem en række såkaldte kvali- tetssikringsmekanismer skal gøre det gennemsig- tigt, målbart og dermed muligt at styre, hvad der foregår i den enkelte institution (Ball, 2003; Hjort, 2002; Peters, 2009; Whitty, 2002; Aasen, 2003). Inden for de seneste 10–15 år har sådan- ne reformer også ramt førskoleområdet og i sti- gende grad reformeret dette område som sty- ringsanliggende. 3. Jeg bruger konsekvent betegnelsen ’hun’ om pædagogen. Helt konkret har alle de pædagoger, jeg har observeret, været af hunkøn. Der kunne foreta- ges en del kønsspecifikke elaboreringer i relation til, hvilke elementer der bindes ind i faglighedskategori- en. Dette er imidlertid ikke forfulgt i denne artikel. 1. Når jeg gennem artiklen senere gør brug af enkelt citationstegn (’) ifm. faglighed, er det for at markere, at jeg disse steder taler om faglighed ikke som analytisk begreb, men som en empirisk kategori, der er vævet ind i læreplanernes dokumentations- krav, og som således skabes og bliver udfyldt på en særlig måde gennem arbejdet med dokumentation. 2. Begreberne’ bevidst’, ’reflekteret’, ’systema- tisk’ og ’målrettet’ indgår således konsekvent i for- bindelse med ’faglighed’ i de forskellige dokumenter, der er udgivet i forbindelse med introduktionen af Lov om Pædagogiske læreplaner. En egentlig analyse af disse sammen-bindinger er lavet i (Plum, 2010). 1. Når jeg gennem artiklen senere gør brug af enkelt citationstegn (’) ifm. faglighed, er det for at markere, at jeg disse steder taler om faglighed ikke som analytisk begreb, men som en empirisk kategori, der er vævet ind i læreplanernes dokumentations- krav, og som således skabes og bliver udfyldt på en særlig måde gennem arbejdet med dokumentation. 2. Begreberne’ bevidst’, ’reflekteret’, ’systema- tisk’ og ’målrettet’ indgår således konsekvent i for- bindelse med ’faglighed’ i de forskellige dokumenter, der er udgivet i forbindelse med introduktionen af Lov om Pædagogiske læreplaner. En egentlig analyse af disse sammen-bindinger er lavet i (Plum, 2010). Maja Plum Institut for Medier, Erkendelse og Formidling, Københavns universitet, Danmark Nemlig som et område, hvor de offentlige ressourcer, der gives ud, må ses i for- hold til det udbytte førskole institutioner bidra- ger med i forhold til det samlede uddannelsesfor- løb og dermed i forhold til nationens fremtidige eforskning utgitt av høgskolen i oslo og akershus og nettverket barnehageliv nordisk barnehageforskning utgitt av høgskolen i oslo og akershus og nettverket barnehageliv MAJA PLUM 2 dagog på. En måde der organiserer og indhegner bestemte elementer af daginstitutionel hverdag som ’fagligt’ pædagogisk arbejde, mens en ræk- ke andre elementer og måder at gøre pædagog på udgrænses fra denne faglighedsdefinition. Dokumentation (be)viser eller udvikler således ikke en allerede eksisterende faglighed. Refor- meringen i læreplanernes krav om dokumentati- on må forstås som de ubemærkede bevægelser, der udspiller sig omkring, hvad der bindes ind i faglighedskategorien, når dokumentation gø- res, og hvad der i selvsamme bevægelse udskilles fra denne faglighedsbetegnelse. Bevægelser der således er med til at værdisætte og hierarkisere det pædagogiske arbejde, forstået således at nogle organiseringer, artefakter og måder at gø- re pædagog bliver sat lig med faglighed, mens andre organiseringer, artefakter og måder at gø- re pædagog udskilles fra denne kategori. Og dermed også den samfundsmæssige anerkendel- se der er forbundet med kategorien. et al., 2009, p. 11f). I Danmark udgør Pædagogi- ske læreplaner et eksempel på en sådan reform. I 2004 indførtes således de pædagogiske lærepla- ner som en national reform af førskole området. En reform der pålægger den enkelte daginstituti- on at opsætte mål og beskrive metoder inden for seks læringstemaer, nemlig temaerne sprog; so- ciale kompetencer; personlige kompetencer; naturen og naturfænomener; kulturelle udtryks- former og værdier; krop og bevægelse (Socialmi- nisteren, 2004), samt ikke mindst dokumentere hvorledes disse mål bliver opnået. Reformen hvi- ler på en argumentation om, at et sådan krav om dokumentation netop ikke må forstås som en egentlig reformering i betydningen forandring af daginstitutionsområdet – dets traditioner og værdier. Kravet om dokumentation udgør inden for reformens egen argumentation ikke et pålæg om et bestemt indhold, men enghed for at skabe viden om det pædagogiske indhold og gennem denne viden gøre dette arbejde synligt samt på denne basis skabe såkaldt ’faglig’1 udvikling. At opsætte mål inden for de seks læringstemaer, be- skrive metoder og dokumentere, hvorledes disse opnås, betragtes således som en vidensproces, der skal udvikle pædagogernes faglighed ved at arbejdet bliver mere ’målrettet’ og ’systematisk’, og at den enkelte pædagog bliver mere ’bevidst’ og ’reflekteret’2. 1. Dermed ikke sagt, at man med et professions- sociologisk begrebsapparat ikke kan diskutere flere af de samme aspekter, der analyseres i denne artikel på andre måder. Eksempelvis er der oplagte forbin- delser mellem læreplansreformens faglighedskategori og akademiseringen af pædagoguddannelsen med alt, hvad det indebærer af tvetydigheder omkring vi- denskabeliggørelse, legitimering og vidensformer (Dybbroe, 2005; Hjort, 2004; Molander & Terum, 2008). Maja Plum Institut for Medier, Erkendelse og Formidling, Københavns universitet, Danmark Med udgangspunkt i en netop afsluttet under- søgelse vil jeg i nærværende artikel analysere dette forhold mellem ’faglighed’, forstået som en kategori der står centralt i læreplansreformen, og det konkrete arbejde med læreplanernes do- kumentation. Mit argument er, at dokumentati- on, som en produktion af viden, ikke kan forstås som en (mere eller mindre korrekt) beskrivelse af det daginstitutionelle arbejde, men at doku- mentation i sig selv indebærer bestemte organi- seringer og iscenesættelser af dette arbejde. Det indebærer, hvad jeg kalder, en måde at gøre pæ- For det første behandler jeg ikke viden som udtryk for en mentalt kategoriseret afspejling af fænomener ’ude i virkeligheden’. Jeg diskuterer altså ikke dokumentation ud fra klassiske viden- skabsfilosofiske parametre om, hvorvidt det ud- gør en form for viden, der ideelt eller reelt for- mår at beskrive en daginstitutionel virkelighed. Jeg studerer, hvordan dokumentation som fæno- men gøres i en daginstitutionel hverdag, og hvordan denne gøren samtidig gør pædagog på ordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no DE PÆDAGOGISKE LÆREPLANERS REFORMERING 3 en bestemt måde. For det andet behandler jeg ikke faglighed som udtryk for en bestemt vi- densbase, uddannelse, værdier, kodekser eller andre professionsanalytiske begrebssætninger1. Jeg er ikke optaget af, hvad faglighed er, burde være eller hvordan det kan begrebssættes. Jeg studerer, hvordan ’faglighed’, forstået som en empirisk kategori, bliver til som en bestemt må- de at gøre pædagog gennem det at gøre doku- mentation. Sagt anderledes, den forbindelse mellem ’faglighed’ og dokumentation, der etab- leres inden for læreplansreformen, bruger jeg som en indgang til at studere, hvad ’faglighed’ bliver gennem arbejdet med dokumentation. eller givet en gang for alle, men som kontinuer- ligt må (gen)erobres ved at pædagogen er videns- skabende, hvilket i denne forvaltningsmæssige sammenhæng vil sige ved at sætte mål, opstille metoder og dokumentere (Plum, 2010). Op til vedtagelsen af lovforslaget om de pæda- gogiske læreplaner var der i politiske såvel som pædagogiske kredse en del debat om et sådant reformtiltag. Diskussioner der særligt tegnede sig som en diskussion om, hvorvidt tiltaget kun- ne forstås som en underminering af den danske børnehaveform med vægten på barnets leg og selvbestemmelse, dvs. som en indførsel af mere skolelignende former, herunder en større statslig styring af området og deri indskrænkning af pæ- dagogens faglige autonomi (jf. E. M. Andersen, 2004; Ellegaard & Stanek, 2004; Hansen, Bech, & Plum, 2004). Efter lovens indførsel synes der imidlertid at have været temmelig stille. Maja Plum Institut for Medier, Erkendelse og Formidling, Københavns universitet, Danmark Arbej- det med at dokumentere er blevet udskældt for at være for tidskrævende, og for at ressourcerne til det ikke slår til, men derudover bliver det un- derstreget fra såvel ministerielle evalueringer som fra fagforeningens side, at pædagogerne er glade for tiltaget, og at det fastholder deres fri- hed og styrker deres faglighed (MINFF, 2008; Rosen-krands, 2008). Som flere af de pædago- ger, jeg har talt med, udtrykker det, så kan man se læreplansreformen som et forvaltningsmæs- sigt tiltag, der for første gang ’handler om pæda- gogik’, eller som drejer sig om dokumentation af det ’bløde’ frem for det ’hårde’, her forstået som talmæssige variable. Lignende konklusioner er at finde i enkelte forskningsrapporter, der be- skæftiger sig med implementeringen af loven og dens krav om dokumentation. Her peges der på, at den lokale udformning og fortolkning af lære- planerne, og dokumentationen forbundet her- med, kan være med til at vise, hvad pædagoger egentlig laver, bringe den pædagogiske tavse vi- den i spil og derved give et fælles fagligt sprog, bevidsthed og heri muligheden for at tilegne sig en højere grad af professionel status (Krejsler & Holm-Pedersen, 2010; Olesen, 2008). ) 2. Når jeg taler om daginstitutionsområdet, refe- rerer jeg til såvel vuggestuer (0–3 år), børnehaver (3– 6 år) og integrerede institutioner (0–6 år). nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no 2. Valget af de tre institutioner er begrundet i forhold til lokalisering, nemlig storby, provins og landområde. Kriterier der relaterer sig til det større forskningsprojekt, som undersøgelsen er en del af. Efter de indledende observationer udvalgte jeg imid- lertid to af disse tre cases. Disse to udvælges ud fra deres forskellighed i arbejdet med dokumentation, dvs. de fungerer som to variation cases (Flyvbjerg, 1991, s.150). Den ene har således de seneste 10 år ar- bejdet med dokumentation og kvalitetssikring og bli- ver i forskellige forvaltningsmæssige sammenhænge trukket frem som det gode eksempel. Den anden er begyndt at arbejde med dokumentation i forbindelse med læreplanerne og har en mere pragmatisk og del- vist kritisk holdning til dette tiltag. Jeg udvælger altså disse to institutioner, fordi jeg tænker, de kan gøre det muligt for mig at se en forskel i, hvorledes der arbejdes med dokumentation, og dermed på hvilke måder daginstitutionel hverdag indrettes i relation hertil. I denne artikel kommer jeg imidlertid ikke ind på sådanne forskelle. LÆREPLANER OG DAGINSTITUTIONS- OMRÅDET I DANMARK for på denne måde at få indkredset organiseringer i den daginstitutionelle hverdag og få greb om disse i relation til de organiserin- ger, der gør sig gældende omkring arbejdet med at dokumentere denne hverdag. Konkret har jeg fulgt én pædagog af gangen og arbejdet med en række fokuspunkter for mine observationer, samt forskellige iagttagelsesmæssige indskærpel- ser såsom observationer af blikretning, mimik, gestikuleringer. Endvidere har jeg i perioder haft lydudstyr fastsat til den fulgte pædagog og i min efterfølgende transskribering skrevet dette ud i forhold til mine observationsnoter af bevægelser i rum, blik, mimik, gestikuleringer. Min status som observatør kan således kategoriseres som ikke-deltagende eller total observation (Cohen & Manion, 1994, p. 108ff; Krogstrup & Kris- tiansen, 1999, p. 101ff) i betydningen, at jeg var kendt netop som ’observatøren’ eller ’forskeren’ etableret via min brug af blokken og pennen, ofte siddende på en stol trukket lidt væk fra børn og pædagoger. Udover mine observationer har jeg foretaget interview med lederne i de tre insti- tutioner, samt med de 13 forskellige pædagoger jeg har fulgt i løbet af mit feltarbejde. Jeg har endvidere indsamlet forskelligt skriftligt og bil- ledligt materiale, som institutionerne har produ- ceret, samt selv taget billeder af indretninger og steder3. Jeg har i relation til disse forskellige kva- litative metodikker arbejdet med perspektiver fra Aktør-Netværks-Teorien. Jeg har således ikke i mere klassik kvalitativ forstand søgt at goger i tre forskellige institutioner2. Jeg har del- taget i morgenåbninger, eftermiddagsfrugt, planlagte projekter, ture, dage på stuen, møder, pauser mv. for på denne måde at få indkredset organiseringer i den daginstitutionelle hverdag og få greb om disse i relation til de organiserin- ger, der gør sig gældende omkring arbejdet med at dokumentere denne hverdag. Konkret har jeg fulgt én pædagog af gangen og arbejdet med en række fokuspunkter for mine observationer, samt forskellige iagttagelsesmæssige indskærpel- ser såsom observationer af blikretning, mimik, gestikuleringer. Endvidere har jeg i perioder haft lydudstyr fastsat til den fulgte pædagog og i min efterfølgende transskribering skrevet dette ud i forhold til mine observationsnoter af bevægelser i rum, blik, mimik, gestikuleringer. Min status som observatør kan således kategoriseres som ikke-deltagende eller total observation (Cohen & Manion, 1994, p. 108ff; Krogstrup & Kris- tiansen, 1999, p. 101ff) i betydningen, at jeg var kendt netop som ’observatøren’ eller ’forskeren’ etableret via min brug af blokken og pennen, ofte siddende på en stol trukket lidt væk fra børn og pædagoger. 1. Jeg bibeholder og fordansker her det engelske translationsbegreb, da den danske ‘oversættelse’ sy- nes at indebære en tilstand af før og efter (altså i dette tilfælde at pædagog er en ting først, men bliver noget andet gennem dokumentation). Med translationsbe- grebet søger jeg at undgå dette før og efter. Det hand- ler i stedet om, hvorledes pædagog gøres forskelligt gennem forskellige sammenbindinger (jf. senere ana- lytiske overvejelser). LÆREPLANER OG DAGINSTITUTIONS- OMRÅDET I DANMARK Udover mine observationer har jeg foretaget interview med lederne i de tre insti- tutioner, samt med de 13 forskellige pædagoger jeg har fulgt i løbet af mit feltarbejde. Jeg har endvidere indsamlet forskelligt skriftligt og bil- ledligt materiale, som institutionerne har produ- ceret, samt selv taget billeder af indretninger og steder3. Jeg har i relation til disse forskellige kva- litative metodikker arbejdet med perspektiver fra Aktør-Netværks-Teorien. Jeg har således ikke i mere klassik kvalitativ forstand søgt at og som en del heraf, at det pædagoger laver har en særlig tavs og blød karakter præget af en løs struktur og således mangel på systematik og skriftlighed. Det er netop disse mangler lærepla- nernes dokumentation, i reformens egen argu- mentation, kan være med til at rette op på, og som sådan skabe den annoncerede synlighed og faglige udvikling. Mit ærinde i denne artikel er ikke at diskutere, hvorvidt dokumentation ska- ber en øget faglighed. Tværtimod. Mit udgangs- punkt er, at arbejdet med dokumentation er knyttet tæt sammen med kategorien ’faglighed’, og at dokumentation på denne måde er med til at skabe, hvad ’faglighed’ bliver. Spørgsmålet er altså, hvad ’faglighed’ bliver, når dokumentati- on gøres. Det er dette grundlæggende spørgs- mål, jeg forfølger i denne artikel. I første om- gang vil jeg således præsentere de metodiske og analytiske præmisser for den undersøgelse, som artiklen tager udgangspunkt i. Herefter vil jeg med udgangspunkt i eksempler fra undersøgel- sen udfolde tre centrale begreber, jeg henter fra Aktør-Netværks-Teorien, nemlig at gøre, sam- menbinding og translatere1. Jeg viser på denne måde, hvordan dokumentation kan forstås som en sammenbinding af daginstitutionelle elemen- ter, hvorigennem pædagog gøres på en særlig måde. En måde hvorpå pædagog translateres som ’faglig’. Jeg peger herefter på andre former for vidensproduktion i den daginstitutionelle hverdag, hvorigennem pædagog translateres på andre måder, der ikke desto mindre netop ikke gøres som dokumentation, og således udgrænses fra faglighedskategorien. og som en del heraf, at det pædagoger laver har en særlig tavs og blød karakter præget af en løs struktur og således mangel på systematik og skriftlighed. Det er netop disse mangler lærepla- nernes dokumentation, i reformens egen argu- mentation, kan være med til at rette op på, og som sådan skabe den annoncerede synlighed og faglige udvikling. Mit ærinde i denne artikel er ikke at diskutere, hvorvidt dokumentation ska- ber en øget faglighed. Tværtimod. LÆREPLANER OG DAGINSTITUTIONS- OMRÅDET I DANMARK I Danmark blev daginstitutionsområdet2 en del af den administrative forvaltning i 1960’erne, hvilket blev fulgt af en voldsom ekspansion i an- tallet af institutioner. Som forvaltningsanliggen- de har fokus helt op til 1990’erne således pri- mært været på antallet af pladser og muligheden for opfyldelsen af den såkaldte pasningsgaranti. Den forvaltningsmæssige styring af området har haft udgangspunkt i en række talmæssige varia- ble og budgetmæssige poster, mens den nærmere varetagelse af dét, der blev refereret til som ’pæ- dagogisk indhold’, har været udlagt til varetagel- se af de såkaldt ’faglige’, nemlig pædagoger og andre børneprofessionelle (P. Ø. Andersen, 2005; Kampmann, 2004; Plum, 2010). Det er dette indhold, der i lov om pædagogiske lærepla- ner nu argumenteres for skal trækkes frem i lyset og gøres til styringsmæssigt omdrejningspunkt. Pædagogens ’faglighed’ bliver på denne måde opløst som en forvaltningsmæssig fast reference – en viden eller ekspert, der kan henvises til. I ste- det bliver pædagogens ’faglighed’ et forvalnings- mæssigt udviklingsobjekt, som netop ikke er fast I Danmark blev daginstitutionsområdet2 en del af den administrative forvaltning i 1960’erne, hvilket blev fulgt af en voldsom ekspansion i an- tallet af institutioner. Som forvaltningsanliggen- de har fokus helt op til 1990’erne således pri- mært været på antallet af pladser og muligheden for opfyldelsen af den såkaldte pasningsgaranti. Den forvaltningsmæssige styring af området har haft udgangspunkt i en række talmæssige varia- ble og budgetmæssige poster, mens den nærmere varetagelse af dét, der blev refereret til som ’pæ- dagogisk indhold’, har været udlagt til varetagel- se af de såkaldt ’faglige’, nemlig pædagoger og andre børneprofessionelle (P. Ø. Andersen, I disse vurderinger af reformens effekter og be- tydning ligger centrale forestillinger omkring pædagoger og pædagogisk faglighed gemt. Dels en forestilling om at der findes en sådan ting som pædagogisk faglighed. En faglighed der ikke desto mindre kan betragtes som ueksplici- teret og svag i form af offentlig anerkendelse – en semiprofession, som det er blevet kaldt inden for professionslitteraturen (Moos, Krejsler, & Laursen, 2008; Nørregård-Nielsen, 2006). Dels, 2. Når jeg taler om daginstitutionsområdet, refe- rerer jeg til såvel vuggestuer (0–3 år), børnehaver (3– 6 år) og integrerede institutioner (0–6 år). ordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no MAJA PLUM 4 goger i tre forskellige institutioner2. Jeg har del- taget i morgenåbninger, eftermiddagsfrugt, planlagte projekter, ture, dage på stuen, møder, pauser mv. LÆREPLANER OG DAGINSTITUTIONS- OMRÅDET I DANMARK rskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no DE PÆDAGOGISKE LÆREPLANERS REFORMERING 5 analysere den interne mening i en daginstitutio- nel hverdag gennem indlevelse eller forståelse (Denzin & Lincoln, 2003, p. 4f), men fremlæst hvad der meningsfuldt etableres som dokumen- tation gennem de konkrete elementer, ord, blik- retning og bevægelser, der bindes sammen. Jeg har med et ANT vokabular analyseret de hetero- gene elementer, der i sine konkrete sammenbin- dinger gør dokumentation, herunder translate- rer pædagogen som ’faglig’. on. Dvs. de artefakter, individer, kategorier, spørgsmål, der går ind i og bindes sammen i det at dokumentere hverdagen. Og ikke mindst, hvorledes pædagogen gøres i og igennem sådan- ne sammenbindinger. g Udgangspunktet i de heterogene sammenbin- dinger betyder således, at analysen ikke starter gennem på forhånd utpegede kvaliteter eller egenskaber bundet til forskellige forhold og fæ- nomener, som giver noget forrang eller forkla- ringskraft frem for noget andet (Callon & Law, 1997; Latour, 1999, p. 174f; Law, 1994, p. 5). Det handler ikke om, hvordan pædagoger udfø- rer dokumentation, men hvordan pædagog, pa- pir, kamera, børn, farveblyanter, navneskilte etc. bindes sammen og gør dokumentation. Pædago- gen studeres altså ikke som udgangspunkt for en handlen, men som en relationel sammenbinding, der gør, men også gøres i den dokumenterende sammenbinding. Dette opgør med individet – i dette tilfælde pædagogen – som det fortolkende eller handlende udgangspunkt indebærer på denne måde, at pædagog kan gøres forskelligt afhængig af den måde hun er bundet op på så at sige, eller rettere de forskellige artefakter, stem- mer, individer, spørgsmål, kategorier mv. hun er bundet sammen med. At gøre pædagog kan alt- så forstås som den konkrete sammenbinding, hvorigennem pædagogen så at sige træder frem som pædagog – den karakter hun får igennem den pågældende sammenbinding. Dette indebæ- rer, at andre former for sammenbindinger kan translatere de pågældende artefakter eller indivi- der på nye måder gennem nye sammenbindinger (Latour, 1999, p. 58; Law, 2003). At gøre og gø- res gennem dokumentation kan således betrag- tes som én måde at gives karakter af pædagog – en karakter, der qua læreplansreformens etable- rede sammenhæng mellem dokumentation og ’faglighed’, kan forstås som en ’faglig’ pædagog. Andre sammenbindinger og dermed organise- ringer og forbindelser af hverdagens elementer kan gøre pædagog på andre måder, der ikke des- to mindre ikke synes dokumentationsværdige, og således udgrænses fra faglighedskategorien. LÆREPLANER OG DAGINSTITUTIONS- OMRÅDET I DANMARK Mit udgangs- punkt er, at arbejdet med dokumentation er knyttet tæt sammen med kategorien ’faglighed’, og at dokumentation på denne måde er med til at skabe, hvad ’faglighed’ bliver. Spørgsmålet er altså, hvad ’faglighed’ bliver, når dokumentati- on gøres. Det er dette grundlæggende spørgs- mål, jeg forfølger i denne artikel. I første om- gang vil jeg således præsentere de metodiske og analytiske præmisser for den undersøgelse, som artiklen tager udgangspunkt i. Herefter vil jeg med udgangspunkt i eksempler fra undersøgel- sen udfolde tre centrale begreber, jeg henter fra Aktør-Netværks-Teorien, nemlig at gøre, sam- menbinding og translatere1. Jeg viser på denne måde, hvordan dokumentation kan forstås som en sammenbinding af daginstitutionelle elemen- ter, hvorigennem pædagog gøres på en særlig måde. En måde hvorpå pædagog translateres som ’faglig’. Jeg peger herefter på andre former for vidensproduktion i den daginstitutionelle hverdag, hvorigennem pædagog translateres på andre måder, der ikke desto mindre netop ikke gøres som dokumentation, og således udgrænses fra faglighedskategorien. ofte siddende på en stol trukket lidt væk fra børn og pædagoger. Udover mine observationer har jeg foretaget interview med lederne i de tre insti- tutioner, samt med de 13 forskellige pædagoger jeg har fulgt i løbet af mit feltarbejde. Jeg har endvidere indsamlet forskelligt skriftligt og bil- ledligt materiale, som institutionerne har produ- ceret, samt selv taget billeder af indretninger og steder3. Jeg har i relation til disse forskellige kva- litative metodikker arbejdet med perspektiver fra Aktør-Netværks-Teorien. Jeg har således ikke i mere klassik kvalitativ forstand søgt at METODISKE OG ANALYTISKE OVERVEJELSER Nærværende artikel tager udgangspunkt i et stu- die af daginstitutionel hverdag i relation til ar- bejdet med dokumentation. Fokus for studiet har været, hvad der sker i en daginstitutionel hverdag, når der skal dokumenteres, herunder hvordan der dokumenteres, og hvad der træder frem som dokumentationsværdigt. Over en peri- ode af et halvt år har jeg fulgt forskellige pæda- 1. Jeg bibeholder og fordansker her det engelske translationsbegreb, da den danske ‘oversættelse’ sy- nes at indebære en tilstand af før og efter (altså i dette tilfælde at pædagog er en ting først, men bliver noget andet gennem dokumentation). Med translationsbe- grebet søger jeg at undgå dette før og efter. Det hand- ler i stedet om, hvorledes pædagog gøres forskelligt gennem forskellige sammenbindinger (jf. senere ana- lytiske overvejelser). LÆREPLANER OG DAGINSTITUTIONS- OMRÅDET I DANMARK Herunder udgrænses fra den værdi, der er tillagt en sådan faglighedskategori. rer pædagogen som ’faglig’. ANT udgør en samlebetegnelse for en række studier og begrebsudviklinger, hvis begyndelse udpeges til slut 1970’erne – 1980’erne, og som fæstnes til navne som Bruno Latour, Michael Callon og John Law. Woolgar og Latours studie ’Laboratory life’ (1979) anses som klassikeren inden for ANT. Her studerer de produktionen af videnskabelige kendsgerninger ved at følge det, der sker i et amerikansk laboratorium. Pointen er, at det, der rendyrkes som en videnskabelig kendsgerning, er produkt af mus, mikroskoper, mænd i hvide kitler, pipetter, computere, printe- re, papir etc., som indgår i forskellige forbindel- ser til hinanden: De hvide kitler bindes til mæn- dene, der igen flytter sig og indgår i forbindelse med musene via en række remedier. Herefter flytter mand-hvid kittel-remedie med musevæv sig til andet sted i rummet, bindes sammen med mikroskop, pen, der noterer tal på papir, hen til en computer og ud på et stykke printet papir. Disse forskellige bevægelser og sammenbindin- ger mellem en række af elementer gentages for til sidst at komme ud på papir som et videnska- beligt faktum. Den videnskabelige kendsgerning er på den måde udtryk for et heterogent netværk af forbindelser eller sammenbindinger – det fremstår som en kendsgerning, men er udgjort af organiseringer på tværs af mus, papir, mænd mv. ANT-perspektivet indebærer på denne måde et opgør med repræsentationstesen, og heri fore- stillingen om videnskabelig erkendelse som ud- tryk for et mere eller mindre korrekt spejl af vir- keligheden. På denne måde bryder et ANT perspektiv med en forståelse af dokumentation som en mere eller mindre korrekt beskrivelse af pædagogisk arbejde eller en objektiv måling af den pågældende daginstitution. Det, der åbnes op for, er i stedet at studere, hvilke sammenbin- dinger der går ind i, og således gør dokumentati- ANT udgør en samlebetegnelse for en række studier og begrebsudviklinger, hvis begyndelse udpeges til slut 1970’erne – 1980’erne, og som fæstnes til navne som Bruno Latour, Michael Callon og John Law. Woolgar og Latours studie ’Laboratory life’ (1979) anses som klassikeren inden for ANT. Her studerer de produktionen af videnskabelige kendsgerninger ved at følge det, der sker i et amerikansk laboratorium. 3. Det er det samlede empiriske materiale, der danner grundlag for mine konklusioner. I denne arti- kel er det imidlertid observationerne, der gøres til ud- gangspunkt for analyse. LÆREPLANER OG DAGINSTITUTIONS- OMRÅDET I DANMARK Pointen er, at det, der rendyrkes som en videnskabelig kendsgerning, er produkt af mus, mikroskoper, mænd i hvide kitler, pipetter, computere, printe- re, papir etc., som indgår i forskellige forbindel- ser til hinanden: De hvide kitler bindes til mæn- dene, der igen flytter sig og indgår i forbindelse med musene via en række remedier. Herefter flytter mand-hvid kittel-remedie med musevæv sig til andet sted i rummet, bindes sammen med mikroskop, pen, der noterer tal på papir, hen til en computer og ud på et stykke printet papir. Disse forskellige bevægelser og sammenbindin- ger mellem en række af elementer gentages for til sidst at komme ud på papir som et videnska- beligt faktum. Den videnskabelige kendsgerning er på den måde udtryk for et heterogent netværk af forbindelser eller sammenbindinger – det fremstår som en kendsgerning, men er udgjort af organiseringer på tværs af mus, papir, mænd mv. ANT-perspektivet indebærer på denne måde et opgør med repræsentationstesen, og heri fore- stillingen om videnskabelig erkendelse som ud- tryk for et mere eller mindre korrekt spejl af vir- keligheden. På denne måde bryder et ANT perspektiv med en forståelse af dokumentation som en mere eller mindre korrekt beskrivelse af pædagogisk arbejde eller en objektiv måling af den pågældende daginstitution. Det, der åbnes op for, er i stedet at studere, hvilke sammenbin- dinger der går ind i, og således gør dokumentati- NÅR DOKUMENTATION GØRES [tager fat i hendes arm] hvor blev det af?”. I dette udklip gøres dokumentation. Der tages et billede, som hænges op hjemme i institutionen for på denne måde at (be)vise, hvorledes man har opnået målene om at skabe rum for fordybelse og dermed kan vise, hvorledes børn lærer om na- turen ved at eksperimentere, opdage, udforske og undersøge. I dette udklip gøres dokumentation. Der tages et billede, som hænges op hjemme i institutionen for på denne måde at (be)vise, hvorledes man har opnået målene om at skabe rum for fordybelse og dermed kan vise, hvorledes børn lærer om na- turen ved at eksperimentere, opdage, udforske og undersøge. p p , g g p g g Jeg vil indlede med et eksempel fra en dag i skoven. I følgende udklip er jeg taget med pæda- gogen Lisbeth i skoven. Det er anden dag i et skovprojekt, for hvilket der er opstillet en række mål og dokumentationspunkter. Det overordne- de mål for projektet er at ’lære gennem at opda- ge, undersøge, udforske og eksperimentere i na- turen’, og i forhold til de voksne er det målet, at ’de øver sig i at skabe læringsrum, der fremmer fordybelse’. Et mål der er beskrevet i institutio- nens læreplan. Fordybelse er således udvalgt som dokumentationspunkt. De har i personale- gruppen aftalt, at alle grupper skal rundt om de fire elementer (ild, jord, vand, luft), og mere spe- cifikt er temaet i dag, om forskellige ting kan fly- de. I går har de derfor prøvet, om kastanjer kun- ne flyde i et vandbassin hjemme på legepladsen, og undervejs på turen hen til skoven har Lisbeth spurgt de børn, hun gik sammen med, om for- skellige ting på deres vej, og om de tror sådanne ting kan flyde. Gruppen har her gjort holdt ved en bæk, og børn går og står omkring bækken, klatrer op og ned ad skråninger, samler grene, kaster med sten eller andet. g g Dokumentation som fænomen er imidlertid netop ikke blot et spørgsmål om det endelige bil- lede eller den beskrivelse, der knyttes til. At gøre dokumentation er i ovenstående tilfælde et fæ- nomen, der gøres som en sammenbinding mel- lem en række heterogene elementer såsom kame- raet, to pædagoger, bækken, glasskåret, pigen, spørgsmål, teorien om de mange intelligenser og de på forhånd opsatte mål. NÅR DOKUMENTATION GØRES NÅR DOKUMENTATION GØRES På tværs af de tre institutioner, jeg har været i, udføres dokumentation typisk i form af tre ting: For det første i form af billeder i relation til hvil- ke der skrives små beskrivelser; for det andet i 3. Det er det samlede empiriske materiale, der danner grundlag for mine konklusioner. I denne arti- kel er det imidlertid observationerne, der gøres til ud- gangspunkt for analyse. ordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no MAJA PLUM 6 form af udstillinger af børneprodukter ligeledes med beskrivelser; og for det tredje i form af evaluerings- eller observationsskemaer hvor et givent projekt vurderes ud fra på forhånd etable- rede spørgsmål eller der noteres iagttagelser om- kring det enkelte barn i forhold til seks eller syv opstillede kompetence- eller intelligenskategori- er (svarende til Howards Gardners Multiple In- telligens Teori eller læreplanstemaerne). Det er altså disse former for vidensproduktion, der om- tales som dokumentation. Følgende vil jeg kigge på denne vidensproduktion, netop ikke som en mere eller mindre korrekt beskrivelse, men som en sammenbinding af heterogene elementer. Ikke af mus, mænd og mikroskoper, men af farvebly- anter, blade, bøger, observationsskemaer, børn, pædagoger, kamera mv. Gennem et ANT- per- spektiv vil jeg altså spørge, hvordan dokumenta- tion gøres – hvilke former for sammenbindinger der er på spil, og hvorledes dette gør pædagog. bliver våd” etc. En pige har fundet et glasskår. Hun holder det op. Pige: ”Kan det her flyde?” Lisbeth går hen til hende, hånd på skulder, vender hende mod sig, siger: ”Neeej, hvad er det – et glasskår”, hun vender pigen igen og skubber hende let ud til kanten af bækken. Lisbeth: ”Prøv at holde det op, så Jon (pæda- gogmedhjælper, der står klar med kameraet) kan tage et billede.” Lisbeth (højt): ”Kan et glasskår flyde?” Pigen kaster glasskåret i van- det. Lisbeth: ”Nej!! [tager fat i hendes arm] hvor blev det af?”. bliver våd” etc. En pige har fundet et glasskår. Hun holder det op. Pige: ”Kan det her flyde?” Lisbeth går hen til hende, hånd på skulder, vender hende mod sig, siger: ”Neeej, hvad er det – et glasskår”, hun vender pigen igen og skubber hende let ud til kanten af bækken. Lisbeth: ”Prøv at holde det op, så Jon (pæda- gogmedhjælper, der står klar med kameraet) kan tage et billede.” Lisbeth (højt): ”Kan et glasskår flyde?” Pigen kaster glasskåret i van- det. Lisbeth: ”Nej!! NÅR DOKUMENTATION GØRES Pæda- gogen fremhæver forhold og artefakter til be- mærkelse (”Neeej et blad”), opford-rer til etablering af erfaringer (”kan et blad flyde?”), og etablerer udpegninger af det enkelte barn (hånd på pigens skulder), der skal spørges, af- prøve og således fremvise sin erfaring og dermed om hendes lærelystne natur har gjort sig en er- kendelse (”Nej! Hvor blev det af?”), dvs. om målet er nået. Gennem sine spørgsmål, håndspå- lægninger, opfordringer og fremhævelser laver pædagogen således netop indzoomninger på lin- je med de indzoomninger, kameraet foretager gennem billedet. Hun etablerer en entydighed omkring et mål og det enkelte barns progression eller erkendelse inden for kategorien naturintel- ligens i relation til dette mål. Lige så lidt som ka- meraet drejes mod de børn, der klatrer op af skrænten, ligeså lidt stiller pædagogen spørgs- mål til denne klatreaktivitet, fremhæver skræn- ten eller laver udpegelser af enkelte børn. lysten, aktiv og udforskende adfærd, der igen kan faciliteres gennem pædagogens opfordring til etablering af en egen erfaring og hendes frem- hævning af forhold, der i den forbindelse bør be- mærkes. Ligesom denne læringslystende adfærd, og heri afprøvning af glasskårets massefylde i forhold til vand, kan bindes op på en kategori omkring barnets intelligens i forhold til natur og naturfænomener. Gennem dokumentation gøres bækken-glasskåret-spørgsmålet på denne måde som et formålsbestemt miljø, hvori barnet udle- ver sin aktive og lærelystne natur. Ligeledes gø- res pædagog. I ovenstående udklip gør og gøres pædagog i en sammenbinding med kamera- spørgsmål-pige-bæk-glasskår-på forhånd opsat- te mål-teorien om de mange intelligenser. Pæda- gogen fremhæver forhold og artefakter til be- mærkelse (”Neeej et blad”), opford-rer til etablering af erfaringer (”kan et blad flyde?”), og etablerer udpegninger af det enkelte barn (hånd på pigens skulder), der skal spørges, af- prøve og således fremvise sin erfaring og dermed om hendes lærelystne natur har gjort sig en er- kendelse (”Nej! Hvor blev det af?”), dvs. om målet er nået. Gennem sine spørgsmål, håndspå- lægninger, opfordringer og fremhævelser laver pædagogen således netop indzoomninger på lin- je med de indzoomninger, kameraet foretager gennem billedet. Hun etablerer en entydighed omkring et mål og det enkelte barns progression eller erkendelse inden for kategorien naturintel- ligens i relation til dette mål. Lige så lidt som ka- meraet drejes mod de børn, der klatrer op af skrænten, ligeså lidt stiller pædagogen spørgs- mål til denne klatreaktivitet, fremhæver skræn- ten eller laver udpegelser af enkelte børn. TRANSLATIONEN AF PÆDAGOG Når dokumentation gøres, gøres pædagog altså på en særlig måde. Hun gives eller træder i ka- rakter som pædagog på en bestemt måde. Eller rettere, hun translateres afhængig af den sam- menbinding, hun indgår i. Dette gælder ikke kun for pædagogen, men som jeg vil illustrere med det næste udklip, for en hel række af de elemen- ter, der skabes forbindelser mellem. Jeg følger denne dag Rikke, der sammen med Eva (anden pædagog) er ved at gøre klar til såkaldt ‘stor- gruppe’. Emnet i dag er ‘venner’ – hvordan man bliver venner, og hvad der gør en til en god ven. Dette emne er igen skrevet ind inden for et over- ordnet mål, nemlig ’at blive opmærksom på hin- anden og øve sig i sociale færdigheder’, som er opstillet i institutionens læreplan. Rikke og Eva har taget puder frem og lagt dem i en rundkreds samt lagt en række navneskilte oven på puderne. Farver, papir, sakse er place- ret på et bord uden for rundkredsen. Rikke finder en bog frem og lægger den på bordet. Eva lægger et kamera samme sted. De gen- nemgår programmet for den næste time: hvor- dan de vil læse en bog og bringe emnet om venner på banen, så børnene kan diskutere dette. Eva skriver på sit papir ‘At være en god ven’, som et punkt der skal dokumenteres. Hun sidder endvidere med et stykke papir, hvorpå der står ’tegn til observation’. Her er markeret syv forskellige intelligenstyper speci- ficeret i relation til hvert enkelt barn. Skrænt og børn retoucheres så at sige væk fra dokumentation gennem sideløbende (ikke kame- raforbundne) irettesættelser, der holder det man- getydige væk – uden for dokumentation. Samti- dig med at pædagogen på denne måde gør, gøres hun gennem disse handlinger og de sammenbin- dinger, hun herved bliver en del af. Hendes ka- rakter af pædagog gøres netop gennem de på forhånd og i situationen foregående sammenbin- dinger og elementer, der er til stede. I den doku- menterende sammenbinding synes pædagog så- ledes at blive en, der etablerer entydighed, som agerer i relation til specifikke mål og faciliterer og vurderer det enkelte barns natur i relation til disse mål og kategorier. Pædagog gøres på denne måde som en formålsbestemt facilitator af bar- nets progression indenfor (i dette tilfælde) syv in- dre kompetenceområder. NÅR DOKUMENTATION GØRES Lad mig starte ved kameraet: Kameraet er taget med i skoven og ta- ges frem, idet de er nået til bækken, og tager bil- leder af børn, der søsætter ting. Kameraet bindes på denne måde ikke til rækken af børn, der på vejen til skoven skal over et lyskryds, til pædago- gen der tæller dem inden de går på tur eller til de børn der i ovenstående udklip kravler op af skrænten. Kameraet bindes sammen med bæk- ken og de børn, der søsætter ting, der kan flyde. Særligt aktiveres kameraet, idet pigen selv for- mulerer spørgsmålet om glasskåret, der kan fly- de. Pigens spørgsmål synes her at skabe en sam- menbinding mellem projektets mål om det udforskende og eksperimenterende barn, bæk- ken og glasskåret. Idet spørgsmålet binder pigen til glasskåret og bækken, bindes samtidig kame- raet og pædagogen. Billedet tages. Der prøves som forskellige ting kan flyde. Fremhævning af elementer sker typisk ved høj stemmeføring a la: ”Neeej et blad – kan et blad flyde? Prøv engang, synker det til bunds?” Kameraet er fundet frem, og der tages billeder af børn, der søsætter forskellige ele- menter. Samtidig render børnene ud til lidt dy- bere vand, og der sker samtidigt et væld af irettesættelser a la: ”nej ikke helt derud”, ”du Dokumentation gøres på denne måde som en sammenbinding mellem de på forhånd formule- rede mål, bækken, bladene, glasskåret, barnet og pædagogen. I denne sammenbinding indgår der samtidig forbindelser til forskellige lærings- teorier og ikke mindst elementet Teorien om de mange intelligenser: Pigens spørgsmål er således noget, der genkendes som udtryk for en lærings- nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no DE PÆDAGOGISKE LÆREPLANERS REFORMERING 7 lysten, aktiv og udforskende adfærd, der igen kan faciliteres gennem pædagogens opfordring til etablering af en egen erfaring og hendes frem- hævning af forhold, der i den forbindelse bør be- mærkes. Ligesom denne læringslystende adfærd, og heri afprøvning af glasskårets massefylde i forhold til vand, kan bindes op på en kategori omkring barnets intelligens i forhold til natur og naturfænomener. Gennem dokumentation gøres bækken-glasskåret-spørgsmålet på denne måde som et formålsbestemt miljø, hvori barnet udle- ver sin aktive og lærelystne natur. Ligeledes gø- res pædagog. I ovenstående udklip gør og gøres pædagog i en sammenbinding med kamera- spørgsmål-pige-bæk-glasskår-på forhånd opsat- te mål-teorien om de mange intelligenser. NÅR DOKUMENTATION GØRES dokumentation en bestemt organisering og re- duktion af det daginstitutionelle arbejde og en bestemt måde at træde i karakter som pædagog. Qua den sammenhæng, der i læreplansreformen etableres mellem dokumentation og ’faglighed’, bliver denne måde at organisere det daginstituti- onelle arbejde og gøres som pædagog dét, der sættes lig med eller kommer til at udfylde, hvad ’faglighed’ vil sige. Gennem dokumentation gø- res pædagogen således som ’faglig’. TRANSLATIONEN AF PÆDAGOG Som sådan indebærer […] Eva samler børn sammen og de entrerer rummet, mens Rikke sidder på lav stol i rund- kredsen (arme foldet omkring ben, blikket føl- ger barnet, som det kommer ind og studerer navneskiltene). Rikke siger, ”Flot Anders! Kunne du lige se, at dit navn så ud som An- ders.” (smiler og stryger ham over ryggen idet han sætter sig ved siden af hende, læner sit ho- nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no 8 MAJA PLUM ved mod hans). Hun siger, “Kunne du finde dit navn Nora? (arme omkring ben igen) Flot! Det var godt, så har du lært noget. Hvor var det godt, du kom med i storgruppen.” ved mod hans). Hun siger, “Kunne du finde dit navn Nora? (arme omkring ben igen) Flot! Det var godt, så har du lært noget. Hvor var det godt, du kom med i storgruppen.” gennem sammenbindinger etableret gennem øjenkontakt, den foroverlænede holdning, for- bindelserne mellem navneskilte og puder samt talemæssige udpegelser (fordeling af taletur). Der skabes sammenbindinger mellem barnet, bogen, navneskilte, temaet om venner gennem ekspliciteringer af forhold til bemærkelse (”Ja, så du kan genkende noget i bogen”), opfordrin- ger til afprøvning (”Kan du finde dit navn?”, Om man kan genkende noget i bogen), fremhæ- velse af erkendelser (”Flot! Det var godt, så har du lært noget” eller ”det var godt tænkt”), som således binder sig til opsatte mål og kategorise- ringer af barnets natur i forhold til syv intelligen- styper (løbende noter der gøres på skema). Pæ- dagogen gør og gøres altså her på linje med foregående eksempel gennem en række sammen- bindinger. Sammenbindinger der netop er tilste- de eller går ind i det at gøre dokumentation. Og som gør pædagogen, som en der etablerer en en- tydighed i relation til specifikke mål og faciliterer og vurderer det enkelte barns natur i relation til disse mål og kategorier. [ …] De sidder alle i cirklen, Rikke taler om den bog, hun holder i hånden, der hedder ‘rig- tige venner’ (løfter bogen, så forsiden kan ses rundt i rundkredsen). Hun begynder at læse højt. Eva tager løbende billeder og skriver på sit papir. Hun noterer ud for det enkelte barn ting såsom, om det kan vente på sin tur, om det kan svare, om det forstår spørgsmål, tør at stå frem, kunne finde sit navn etc. Noterne gø- res mest ift. sociale, personlige og sproglige in- telligens. TRANSLATIONEN AF PÆDAGOG Rikke læser, ”nu er jeg rigtig sur på dig!”. Anna (barn) kommenterer, at hun også har prøvet at være sur på en ven. Rikke siger, ”Ja, så du kan genkende noget i bogen.” Hun læser færdig og spørger de andre børn om de kan genkende noget. Emma (barn) siger, at man må ikke sige idiot i børnehaven. Rikke (bogen nu på hendes lår, foroverbøjet mod Emma, ser hende i øjnene) spørger, ”hvorfor må man ikke det?”. Emma siger, at hendes sto- resøster siger, at man ikke må. Rikke siger, ”Og så tænkte du lige, at når man ikke må derhjemme, så skal man heller ikke gøre det her. Det var godt tænkt!” (peger finger mod eget hoved). Emma siger noget mere, og Rikke fordeler taletur ved at sige deres navne højt (blik mod det pågældende barn, foroverlænet mod det). De taler om at blive sure på sine venner. Rikke afslutter med at sige, at så er det vigtigt at blive gode venner igen. g g Igennem denne dokumenterende sammenbin- ding translateres elementernes karakter. En bog kan tages frem, læses, afbrydes, siddes på skød omkring i løbet af dagen, men translateres på en særlig (dokumentaristisk) måde, idet den her bindes sammen med kameraet, observationsske- maet, de opsatte mål, et programfastsat forløb, en række navnefastsatte børn, to udvalgte pæ- dagoger og et aflukket rum. Bogen gøres her gennem børnene, der sidder i rundkreds, forsi- den der holdes op, talemæssig fremhævelse af specifikke pointer, spørgsmål der skal relatere disse pointer til barnets egne erfaringer og over- vejelser samt pædagogen, der nedfælder stikord på sine skemaer. Bogen er således på forhånd bundet ind i en konstellation, hvor dens temati- ske indhold bindes sammen med projektets tema ’venner’, hvor en sådan tematik gør det muligt at stille spørgsmål til børnenes egne erfaringer, og hvor svarene, måden de gives og udfoldes på, er etableret som på forhånd etablerede skue- pladser, hvorigennem barnets sociale, personlige og sproglige kompetencer kan vurderes og ar- bejdes med (om man kan vente på tur, om man tør stå frem og sige noget, om man kan forstå og svare på spørgsmål, hvorledes dette ræsonne- ment over spørgsmål udfolder sig mv.). nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no DE PÆDAGOGISKE LÆREPLANERS REFORMERING 9 Hvis man følger disse ikke-dokumenterende vidensproduktioner gennem et ANT-perspektiv gøres de som en anden form for sammenbinding end dokumentation. De indebærer andre former for organisering af det daginstitutionelle arbejde og translaterer pædagogen på andre måder, der således udgrænses fra faglighedskategorien. I følgende eksempel følger jeg pædagogen Sine, der skal ind og ’putte børn’, som det omtales. En aktivitet i relation til hvilken, der på et stykke papir er optegnet en række navne på børn, og hvor disse navne konsulteres og afstemmes i for- hold til protokollens registreringer af afleverede børn denne dag. res som et formålsbestemt element, hvorigen- nem barnets sproglige intelligens kan afprøves og vurderes (om de kan kende deres eget navn), og hvor børn kan fastsættes netop som enkelte individer, der kan faciliteres som sådan (spørges individuelt, blik mod det enkelte barn, kroppen, der læner sig mod pude-navneskilt-barn). Når dokumentation gøres som en sammenbin- ding af heterogene elementer, sker der således en translation af elementernes karakter. De holdes så at sige fast på en særlig måde: De organiseres og træder i karakter på særlig vis. Det er denne translation af elementet pædagog, der gør hende som ’faglig’. Qua den etablerede sammenbin- ding mellem dokumentation og ’faglighed’, er det i denne dokumenterende sammenbinding, at hun gives karakter af ’faglig. At der er tale om en translation betyder samtidig, at der er en række andre elementer og sammenbindinger, der i løbet af en daginstitutionel dag gør hende som pædagog. Men på en andre måder. Den do- kumenterende translation af pædagogen som ’faglig’ indebærer således både en indhegning af specifikke sammenbindinger som ’fagligt’ pæda- gogisk arbejde, mens en række andre sammen- bindinger, og heri måder at translatere pædagog på udgrænses fra denne faglighedsdefinition. Det er nogle af disse udgrænsninger, jeg følgende vil sætte fokus på. Sine går mod fællesrummet med en gruppe af børn foran sig. I rummet er der en anden pæ- dagog i gang med at tage madrasser ned fra lodrette hylder, markeret med navneskilte. Hun placerer dem i rækker på gulvet, fordeler navnefastsatte dyner og puder oven på. Da Sine og børnene kommer ind, lægger No- man sig på madras tæt på dør. Sine siger, ”Nej, ved du hvad, det er Kaspers plads.” Hun peger rundt på forskellige madrasser mens hun for- tæller børn, hvor de skal ligge. Hun tæller nav- ne på sine fingre. Børn er placeret på separate madrasser. DE PÆDAGOGISKE LÆREPLANERS REFORMERING Carl sætter sig op og begynder at pille ved gardin (griner). Sine siger ”Carl!”. Sine går mod døren, vender sig rundt siger, ”hvem har lavet lort?”, hun læner sig mod den nærmeste barnenumse og lugter. Den anden pædagog siger, at nogen har tabt en ble der. Sine griner, siger ”Add”. Hun går ud og kom- mer tilbage med en til dyne, siger ”Carl, du skal lægge dig ned. Jeg vil ikke have mere bal- lade.” Laura (barn) siger, “jeg vil ha’ min sut”. Sine siger, ”det kan du tro, jeg finder den til dig.” Hun går og kommer tilbage med sut, gi- ver den til Laura, stryger hende over hår. Hun putter dyne ned om Louise (barn), og vender hende rundt, så hun ser ind i væggen. Hun sætter sig mellem Laura og Carl. Carl laver en lyd, og hun løfter ham, vender ham rundt, så han ser ind mod væggen, putter dynen om- kring ham, og lægger sig ned med sit hoved ved siden af hans. Han piller ved gardinet, og Sine holder det fast, hun hvisker ”lad gardinet være.” Hun lægger hånd på Lauras hoved, stryger hendes hår. Laura og Carl sover. Det er nogle af disse udgrænsninger, jeg følgende vil sætte fokus på. TRANSLATIONEN AF PÆDAGOG Det sam- me gælder for puderne, der til dagligt bygges hule af, kastes med, ligges på, men som i sam- menbindingen med navneskiltene, målene, ob- servationsskemaet og det enkelte barn translate- I dette eksempel på dokumentation, der gøres, sker der en sammenbinding mellem en hel række af elementer: Puder, navneskilte, navnfastsatte børn, bog, pædagoger, kamera, opstillede mål, observationsskema, teorien om de mange intelli- genser. Igen er der således tale om, at dokumen- tation gøres som en sammenbinding mellem en række heterogene elementer, der dermed organi- seres på en bestemt måde, hvorigennem pædago- gen men også barnet får en specifik karakter: Gennem dokumentation gøres barnet som læ- rende, og pædagogen gøres som formålsbestemt facilitator af denne læringslystne natur – i nær- værende eksempel i relation til kategorier om- kring primært sociale, personlige og sproglige in- telligenser. På linje med foregående eksempel laves der her indzoomninger: Det enkelte barn træder frem netop som enkelt og navnfastsat ordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no DE PÆDAGOGISKE LÆREPLANERS REFORMERING DET UDGRÆNSEDE Igennem mit feltarbejde noterer jeg mig en række skriftliggørelser af forskellige forhold, som jeg imidlertid samtidig bemærker ikke omtales eller præsenteres som dokumentation. Igennem bøger eller på store tavler, fungerer der således proto- kolliske opgørelser over afleverede og hentede børn, markeringer af i hvilke rum, de pågælden- de børn befinder sig, lister over børns navne, der krydses af i forbindelse med eftermiddagsfrugt og indtagelse af væske eller lister med en række navne og tidsintervaller, hvori der sættes kryds, når det pågældende barn er blevet skiftet. Disse fastetablerede og systematiserede skriftliggørel- ser synes imidlertid ikke at kunne betragtes som dokumentation. De synes således netop ikke at figurere som den ’skriftlighed’ og ’systematik’, der menes at mangle ved pædagogfaget. Og der- med heller ikke som en vidensproduktion der kan gøre pædagoger ’bevidste’, ’reflekterede’ og dermed ’faglige’, sådan som der igennem lære- plansreformen ellers argumenteres for at doku- mentation kan. [...] Ude af fællesrummet noterer Sine på tav- len, de børn der ikke er faldet i søvn, ligesom nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no ordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no MAJA PLUM 10 det noteres hvor længe pågældende barn har sovet, idet han eller hun vågner. række af sammenbindinger. Hun tæller børn, fø- rer børn fra rum til rum, fordeler madrasser, for- deler børn, henter sut, henter dyne, aer, lugter, vender, irettesætter. Hun gøres som en, der på den ene side skal agere i forhold til det enkelte barns behov for sut, men omvendt netop skal agere i forhold til en masse af børn og en bestemt behovsfunktion – i dette tilfælde søvn. Hun skal således holde sammen på denne masse af elemen- tet barn, der alle skal sove. Når der gøres sove- tidsregistrering, gøres pædagogen altså gennem sammenbindinger, der netop ikke handler om kameraartefaktiske indzoomninger, men om op- gørelsesmæssigt overblik: hvilke børn skal hvor- hen, hvilke madrasser, puder, sutter, lugte, stem- mer skal være frembragt eller bragt til ophør i relation til den pågældende funktion (sove). I dette udklip gøres en vidensregistrering: Der er sat beskrivelser op for hvilke børn, der skal sove og det registreres om de har sovet, samt hvor længe dette har stået på. Der er således foretaget en opgørelse, der i relation til mængden af børn for hvem det er hensigten, at de skal sove, (be)vi- ser om denne hensigt er opfyldt, og i såfald hvor længe det enkelte barns søvn har varet. DET UDGRÆNSEDE Der er således foretaget en opgørelse, der i relation til mængden af børn for hvem det er hensigten, at de skal sove, (be)vi- ser om denne hensigt er opfyldt, og i såfald hvor længe det enkelte barns søvn har varet. Igen er det imidlertid ikke opgørelsen som sådan, der fra et ANT perspektiv er det interessante, men fæno- menet og den sammenbinding, der her foregår. relation til den pågældende funktion (sove). Daginstitutionel hverdag er fuld af sådanne translationer af børn og pædagog: Af børn der afkrydses, føres med let hånd på nakken til et andet rum, hen til sin stol for at spise, tages op på skød, holdes ind til kroppen. Fuld af pædago- ger der tjekker lister, noterer, stiller mad frem, i en samarbejdende koordinering fører tøj på bør- nekroppen der skal ud at lege eller ind at sove, skifter ble, stryger over hår, irettesætter, schh’er, holder på arme eller ben der bevæger sig under maden eller til samling. Det er alle disse aspekter – translationen af disse elementer i sin sammen- binding med barn og pædagog – der aldrig synes at træde frem som dokumentationsværdige. Det er disse måder at gøre pædagog, der ikke gøres dokumenterende, og som således udgrænses fra, hvad ’faglighed’ bliver. Med andre ord, doku- mentation synes at kalde bestemte organiserin- ger af det daginstitutionelle arbejde frem. Be- stemte måder at træde i karakter som pædagog, der kommer til at udfylde, hvad ’faglighed’ vil sige, samtidig med at en lang række organiserin- ger – og heri måder at handle, omgås og være med børn – netop ikke bliver en del af denne faglighedskategori. Det paradoksale synes såle- des at være, at der i danske daginstitutioner er mængder af artefakter, såsom bleer, skralde- spande, tallerkner, madrasser mv., som indgår i kontinuerlige sammenbindinger med pædagog og barn, men som i læreplanernes reformering udgrænses fra ’faglighed’. Ikke fordi de ikke er skriftlige eller systematiserede – de træder netop frem gennem opgørelser, lister, skemaer – men fordi de ikke synes at blive en del af dokumenta- tion. Argumentet i denne artikel er ikke, at disse elementer burde dokumenteres. nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no DET UDGRÆNSEDE Igen er det imidlertid ikke opgørelsen som sådan, der fra et ANT perspektiv er det interessante, men fæno- menet og den sammenbinding, der her foregår. At gøre sovetidsregistreringer er et fænomen, der gøres som en sammenbinding mellem en række heterogene elementer såsom protokollen, listen over børn, madrasser, dyner, puder, navneskilte, børn, to pædagoger, sut, bleer samt en række teorier om børns basale behov for mad, søvn, omsorg. Igennem denne sammenbinding transla- teres elementerne. Navn-barn-pude er i nærvæ- rende eksempel ikke (som i eksemplet med Rik- ke) en sammenbinding, hvorigennem barnets lærende natur kan beskues og vurderes. Navn- barn-pude er en sammenbinding, der fordeler børn på pladser med deres egne bakterier og lug- te. Pladser hvor pædagoger kan sætte sig (ved si- den af Carl), fysisk flytte rundt på børn (vende dem så de kigger ind mod væggen), ae dem (over håret) og sige de skal være stille. Barn(-navn-pu- de) bindes således ikke til pædagogen gennem blikket, den foroverlænede krop, de faciliterende spørgsmål eller ekspliciteringer af forhold til be- mærkelse. Barn er i denne sammenbinding ikke øjne, indre erfaringer og overvejelser, der skal drages frem. Barn er numser der kan lugte (Sines hoved mod barnets bagdel), hår der kan aes og krop, der kan fordeles og flyttes. At gøre sove- tidsregistrering gør således barn og pædagog på en ganske anden måde, end når der gøres doku- mentation. Det gør barn som del af en masse el- ler gruppe, der skal flyttes gennem forskellige rum eller funktioner for på denne måde at få op- fyldt en række behov – fra stuen hvor der spises, til toilettet hvor der skiftes eller tisses, til fælles- rummet, hvor der soves – og som kan have vari- ationer i disse behov: Nogle skal have sut, nogle aes, nogle vendes mod væggen. Som del af en masse gøres barnet samtidig som element, der kan bryde med den pågældende funktion (piller ved gardin, grine, tale mv. når der skal soves), og som således irettesættes til opfyldelse af sit be- hov. Pædagogen gøres ligeledes gennem denne I dette udklip gøres en vidensregistrering: Der er sat beskrivelser op for hvilke børn, der skal sove og det registreres om de har sovet, samt hvor længe dette har stået på. DET UDGRÆNSEDE Pointen er, at der med reformen indføres et krav om viden, der Daginstitutionel hverdag er fuld af sådanne translationer af børn og pædagog: Af børn der afkrydses, føres med let hånd på nakken til et andet rum, hen til sin stol for at spise, tages op på skød, holdes ind til kroppen. Fuld af pædago- ger der tjekker lister, noterer, stiller mad frem, i en samarbejdende koordinering fører tøj på bør- nekroppen der skal ud at lege eller ind at sove, skifter ble, stryger over hår, irettesætter, schh’er, holder på arme eller ben der bevæger sig under maden eller til samling. Det er alle disse aspekter – translationen af disse elementer i sin sammen- binding med barn og pædagog – der aldrig synes at træde frem som dokumentationsværdige. Det er disse måder at gøre pædagog, der ikke gøres dokumenterende, og som således udgrænses fra, hvad ’faglighed’ bliver. Med andre ord, doku- mentation synes at kalde bestemte organiserin- ger af det daginstitutionelle arbejde frem. Be- stemte måder at træde i karakter som pædagog, der kommer til at udfylde, hvad ’faglighed’ vil sige, samtidig med at en lang række organiserin- ger – og heri måder at handle, omgås og være med børn – netop ikke bliver en del af denne faglighedskategori. Det paradoksale synes såle- des at være, at der i danske daginstitutioner er mængder af artefakter, såsom bleer, skralde- spande, tallerkner, madrasser mv., som indgår i kontinuerlige sammenbindinger med pædagog og barn, men som i læreplanernes reformering udgrænses fra ’faglighed’. Ikke fordi de ikke er skriftlige eller systematiserede – de træder netop frem gennem opgørelser, lister, skemaer – men fordi de ikke synes at blive en del af dokumenta- tion. Argumentet i denne artikel er ikke, at disse elementer burde dokumenteres. Pointen er, at der med reformen indføres et krav om viden, der ordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no DE PÆDAGOGISKE LÆREPLANERS REFORMERING 11 skal vise og udvikle en pædagogisk faglighed, men at denne viden netop ikke kan forstås som en synliggørelse af noget allerede eksisterende. Dokumentation gør ’faglighed’: Det fremkalder så at sige bestemte sammenbindinger og måder at gøre pædagog som relevante, legitime og ak- tuelle – som udtryk for ’fagligt’ pædagogisk ar- bejde. DET UDGRÆNSEDE Dermed ikke sagt at de andre måder for- svinder, men at de fungerer som noget uden for ’faglighed’ – noget vi dermed principielt kan sætte folk uden en sådan faglighedstittel til, og hvor værdien af dette arbejde, og måden det ud- føres, dermed sættes uden for faglighedskatego- riens diskussion og anerkendelse. forstås som de ubemærkede processer, der ud- spiller sig, idet dokumentation gøres. Processer hvorigennem noget kaldes frem, bindes sammen, organiseres og gøres på særlige ’faglige’ måder, mens andre skriftliggørelser, artefakter, sammen- bindinger og måder at gøre pædagog ikke får ka- rakter af ’faglighed’. Mit forehavende er i den forbindelse ikke at argumentere for, at faglighed burde gøres, defineres eller forstås på en bestemt måde. Min pointe er, at der i læreplansreformens påståede synliggørelse og udvikling af det eksi- sterende foregår en værdisætning og en hierarki- sering af det pædagogiske arbejde. Processer hvorigennem der indhegnes og udgrænses, hvad pædagogisk ’faglighed’ vil sige. Som sådan handler de processer af dokumen- tation, der foregår hver dag, og som kan synes ganske uskyldige, at indebærer transformatio- ner, der går udover den enkelte daginstitutions bestræbelser på at synliggøre sit pædagogiske indhold. Transformationer der hænger sammen med en international uddannelsespolitisk dags- orden, hvor daginstitutionsområdets bidrag til landets human kapitale beholdning netop bliver nærværende i de måder, hvorpå det synes legi- timt eller muligt at træde i karakter som ’faglig’ (Plum, 2012). Med andre ord, de indzoomnin- ger og etableringer af entydighed, der gør pæda- gogisk ’faglighed’ som en formålsbestemt facili- tering af barnet som kompetenceopdelt og progressionsvilligt indre, må ses i relation til de forståelseshorisonter, der er indlejret i selve læ- replansreformen som styringstiltag. Ligeledes må udgrænsningen af de sammenbindinger mel- lem børnekroppe, pædagoger, sutter, bleer, lug- te, rum og de måder at handle, omgås og være med børn, der en del af sådanne organiseringer, ses som mere end et tilfældigt valg af dokumen- tationsfokus i den enkelte daginstitution. Snare- re må det ses som udtryk for en samfundsmæs- sig værdisætning af, hvad børn er – hvorledes de er værd at beskæftige sig med. Og dermed også, hvad der værdisættes og ikke værdisættes som del af en ’faglig’ pædagogisk opgave. Som sådan ligger læreplanernes egentlige reformering i de tilsyneladende uskyldige processer af at doku- mentere den pædagogiske ’faglighed’. AFSLUTTENDE BEMÆRKNINGER Research methods in education. London: Routledge. Denzin, N. K., & Lincoln, Y. S. (2003). Introduction: The discipline and practice of qualitative research. In N. K. Denzin & Y. S. Lincoln (Eds.), Strategies of qualitative inquiry (pp. 1–45). London: SAGE Publications. NOU Norges offentlige utredninger. (2011). Bedre integrering. Mål, strategier, tiltak. Oslo: Departe- mentenes servicesenter: Informasjonsforvaltning. Dybbroe, B. (2005). Uddannelse af pædgaoger som refleksivt moderniseringsprojekt. I T. R. Eriksen & A. M. Jørgensen (Eds.), Professionsidentitet i for- andring (pp. 228–243). Copenhagen: Akademisk Forlag. Nørregård-Nielsen, E. (2006). Pædagoger i skyggen: Om børnehavepædagogernes kamp for faglig aner- kendelse. Odense: Syddansk Universitetsforlag. Olesen, J. (Ed.). (2008). Når loven møder børns insti- tutioner. Copenhagen: Danish School of Education Press. Ellegaard, T., & Stanek, A. H. (Eds.). (2004). Lære- planer i børnehaven: Baggrund og perspektiver. Vejle: Kroghs Forlag. Peters, M. A. (2009). Governmentality, education and the end of neoliberalism? In M. A. Peters, A. C. Besley, M. Olssen, S. Maurer & S. Weber (Eds.), Governmentality: Studies in education (pp. xxvii– xlviii). Rotterdam: Sense. Flyvbjerg, B. (1991). Rationalitet og magt: Det kon- kretes videnskab. Aarhus: Akademisk forlag. Hansen, A. D., Bech, S. L., & Plum, M. (2004). Spil- let om læring: En diskursanalyse af brugen af læ- ring på dagtilbudsområdet. Copenhagen: Learning Lab Denmark. Plum, M. (2010). Dokumenteret faglighed: Analyser af hvordan 'pædagogisk faglighed' produceres gen- nem læreplanernes dokumentationsteknologi. Co- penhagen: University of Copenhagen. Hjort, K. (2002). Moderniseringen af den offentlige sektor. Roskilde: Roskilde Universitetsforlag. Plum, M. (2012). Humanism, administration and education: The demand of documentation and the production of a new pedagogical desire. Journal of Education Policy, 27,(4), 491–507. Hjort, K. (Ed.). (2004). De profesionelle: Forskning i professioner og professionsuddannelser. Roskilde: Roskilde Universitetsforlag. Kampmann, J. (2004). Societalization of childhood: New opportunities? New demands? In H. Brem- beck, B. Johansson & J. Kampmann (Eds.), Bey- ond the competent child (pp. 127–152). Roskilde: Roskilde University Press. Rosenkrands, U. (2008). Pædagogiske læreplaner: En bragende succes. Børn & Unge, 11. Skolverket. (2004). Förskola i brytningstid: En natio- nell utvärdering av förskolan. Stockholm: Skolver- ket. Krejsler, J., & Holm-Pedersen, P. (2010). Hvad der lignende en kontrakt med djævlen. Børn & Unge. Forskning, 7. Whitty, G. (2002). Making sense of education policy: Studies in the sociology and politics of education. London: SAGE. Krogstrup, H. K., & Kristiansen, S. (1999). Delta- gende observation, introduktion til en samfundsvi- denskabelig metode. Copenhagen: Hans Reitzel. Østrem, S., Bjar, H., Føsker, L. R., Hogsnes, H. AFSLUTTENDE BEMÆRKNINGER Med inspiration fra Aktør-Netværks-Teorien har jeg i denne artikel analyseret, hvorledes do- kumentation gøres. Med udgangspunkt i lære- plansreformens sammenkobling mellem doku- mentation og ’faglighed’ har jeg anskuet doku- mentation som et fænomen af heterogene ele- menter, der i sin sammenbinding translaterer ele- menternes karakter, og derved gør pædagog på særlig ’faglig’ vis. Jeg har således vist, hvorle- des dokumentation gøres gennem sammenbin- dinger af eksempelvis puder-navneskilte-navn- fastsatte børn-opsatte mål-observationsskema- er-kamera-bog-to pædagoger-teorien om de mange intelligenser. Sammenbindinger, hvori- gennem den daginstitutionelle hverdag organise- res på bestemte måde, og hvorigennem pædagog translateres på særlig vis. Gør hende ’faglig’: Hvor denne faglighedskategori bliver udtryk for en formålsbestemt facilitator af barnets progres- sion indenfor en række indre kompetenceområ- der. Igennem denne type af sammenbinding gøres pædagogen som ’faglig’ gennem etablere- ring af indzoomninger på det enkelte barn, der herved fastholder en entydighed i den daginstitu- tionelle kompleksitet. Med udgangspunkt i et eksempel på andre af de skriftlige vidensproduk- tioner, der foregår i en daginstitutionel hverdag, har jeg analyseret andre former for sammenbin- dinger og dermed translationer af pædagog. Må- der at gøre pædagog på der netop ikke synes at kunne gøres som dokumentation, og som således udgrænses fra faglighedskategorien. LITTERATUR Andersen, E. M. (2004). Læreplaner: Glem ikke bar- net. Børn & Unge, 13. Andersen, P. Ø. (2005). Pædagogens praksis. Copen- hagen: Hans Reitzel. Dokumentation kan som sådan netop ikke for- stås som en mere eller mindre korrekt viden om det pædagogiske indhold. Som noget der synlig- gør dette indhold og udvikler fagligheden. I ste- det kan de pædagogiske læreplaners reformering ordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no 12 MAJA PLUM MAJA PLUM 12 Ball, S. J. (2003). Class strategies and the education market: The middle classes and social advantage. London: RoutledgeFalmer. Law, J. (2003). Traduction/Trahision: Notes on ANT. Retrieved from www.comp.lancs.ac.uk/so- ciology/papers/law-traduction-Trahision.pdf MINFF. (2008). Evaluering af loven om pædagogi- ske læreplaner: Ministeriet for Familie- og For- brugsanliggender. Bekendtsgørelse om temaer og mål i pædagogiske læ- replaner BEK nr. 684 (2004). Callon, M., & Law, J. (1997). After the individual in society: Lessons on collectivity from science, tech- nology and society. Canadian Journal of Sociolo- gy, 22(2), 165–182. Molander, A., & Terum, L. I. (Eds.). (2008). Profe- sjonsstudier. Oslo: Universitetsforlaget. Moos, L., Krejsler, J., & Laursen, P. F. (2008). Rela- tionsprofessioner: Lærere, pædagoger, sygeplejer- sker, sundhedsplejersker, socialrådgivere og mellemledere. Copenhagen: Danmarks Pædagogi- ske Universitet. Cohen, L., & Manion, L. (1994). AFSLUTTENDE BEMÆRKNINGER D., Jansen, T. T., Nordtømme, S., et al. (2009). Alle teller mer: En evaluering av hvordan Rammeplan for barnehagens innhold og oppgaver blir innført, brukt og erfart. Tønsberg: Høgskolen i Vestfold. Latour, B. (1999). Pandora's hope: Essays on the re- ality of science studies. Cambridge, Mass. ; Lon- don: Harvard University Press. Aasen, P. (2003). What happened to social-demo- cratic progressivism in Scandinavia? Restructuring education in Sweden and Norway in the 1990s. In M. Apple (Ed.), The state and the politics of knowledge (sid. 109–147). New York: Routledge- Falmer. Latour, B., & Woolgar, S. (1979). Laboratory life: The construction of scientific facts (2nd ed.). Lon- don: Sage. Law, J. (1994). Organizing modernity: Blackwell. rskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no nordisk barnehageforskning 2012 5(12), 1–12 issn 1890-9167 www.nordiskbarnehageforskning.no
https://openalex.org/W3007627731	https://peerj.com/articles/8588v0.3/submission	English	null	Peer Review #1 of "Increased absorption in autonomous sensory meridian response (v0.1)"	null	2,020	cc-by	11,362	Manuscript to be reviewed Increased absorption in Autonomous Sensory Meridian Response Agnieszka Barbara Janik McErlean Corresp., 1 , Eleanor Jane Osborne-Ford 1 1 School of Science, Bath Spa University, Bath, United Kingdom Corresponding Author: Agnieszka Barbara Janik McErlean Email address: a.janikmcerlean@bathspa.ac.uk Background. Autonomous sensory meridian response (ASMR) is a cross-sensory phenomenon characterised by a static-like sensation which typically originates on the scalp and spreads throughout the body leading to a state of deep relaxation. It can be triggered by visual and auditory stimuli in real life, incidentally by various media and via intentionally created ASMR videos. Previously ASMR has been linked to a speciﬁc personality proﬁle and this study aimed to further elucidate individual diﬀerences associated with this phenomenon. Methods. To this eﬀect ASMR-Experiencers and age and gender matched controls were compared on measures of ﬂow, absorption and mindfulness. Results. This revealed that ASMR was associated with elevated absorption but no group diﬀerences were found with respect to the other constructs, suggesting that the ability to get deeply immersed with the current experience accompanied by loss of reﬂective awareness may be an important factor contributing to the experience of ASMR. PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Manuscript to be reviewed Manuscript to be reviewed Increased absorption in Autonomous Se Agnieszka Barbara Janik McErlean1 and E 1. School of Science, Bath Spa Unive Corresponding Author: Dr Agnieszka Janik McErlean School of Science Bath Spa University Bath UK E-mail: a.janikmcerlean@bathspa.ac.uk Tel: +44 (0)1225 876637 Abstract 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Background. Autonomous sensory meridian response (ASMR) is a cross-sensory phenomenon characterised by a static-like sensation which typically originates on the scalp and spreads throughout the body leading to a state of deep relaxation. It can be triggered by visual and auditory stimuli in real life, incidentally by various media and via intentionally created ASMR videos. Previously ASMR has been linked to a specific personality profile and this study aimed to further elucidate individual differences associated with this phenomenon. Methods. To this effect ASMR-Experiencers and age and gender matched controls were compared on measures of flow, absorption and mindfulness. Results. This revealed that ASMR was associated with elevated absorption but no group differences were found with respect to the other constructs, suggesting that the ability to get deeply immersed with the current experience accompanied by loss of reflective awareness may be an important factor contributing to the experience of ASMR. 25 26 27 28 29 30 31 32 33 34 35 Background. Autonomous sensory meridian response (ASMR) is a cross-sensory phenomenon characterised by a static-like sensation which typically originates on the scalp and spreads throughout the body leading to a state of deep relaxation. It can be triggered by visual and auditory stimuli in real life, incidentally by various media and via intentionally created ASMR videos. Previously ASMR has been linked to a specific personality profile and this study aimed to further elucidate individual differences associated with this phenomenon. Methods. To this effect ASMR-Experiencers and age and gender matched controls were compared on measures of flow, absorption and mindfulness. Results. This revealed that ASMR was associated with elevated absorption but no group differences were found with respect to the other constructs, suggesting that the ability to get deeply immersed with the current experience accompanied by loss of reflective awareness may be an important factor contributing to the experience of ASMR. Manuscript to be reviewed 25 26 27 28 29 30 31 32 33 34 35 Autonomous sensory meridian response (ASMR) is a multisensory phenomenon where auditory and visual stimuli such as whisper or personal attention trigger a pleasant, static-like 44 45 Autonomous sensory meridian response (ASMR) is a multisensory phenomenon where auditory and visual stimuli such as whisper or personal attention trigger a pleasant, static-like 44 45 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) tingling sensation which typically originates from the head and disperses throughout the body resulting in a relaxed state (Barratt & Davis, 2015). Although ASMR can be experienced in daily life, in recent years many ASMR-inducing YouTube channels have been created attracting large audiences who watch such videos to experience the sensation, relax, and fall asleep and in some cases to alleviate anxiety and stress (Barratt & Davis, 2015; Janik McErlean & Banissy, 2017; Poerio, Blakey, Hostler, & Veltri, 2018). Although the prevalence of ASMR is not known, the popularity of ASMR media suggests that the phenomenon might be widespread. 46 47 48 49 50 51 52 One line of evidence suggests that there are personality differences between people who experience ASMR and those who do not. Specifically, trait ASMR has been linked to i) increased openness to experience (Janik McErlean & Banissy, 2017; Fredborg, Clark, & Smith, 2017) which taps into one’s interest in novel experiences and propensity to fantasise (John, Naumann, & Soto, 2008), ii) heightened fantasising (Janik McErlean & Banissy, 2017), which reflects an ability to immerse oneself in a fictional reality (Davis, 1983), and iii) elevated mindfulness (Fredborg, Clark and Smith, 2018), which entails concentrating on the present moment (Brown & Ryan, 2003). These findings are interesting taking into account that for many individuals ASMR is triggered when focusing on the external triggers which resembles mindfulness practice (Fredborg et al., 2018) and that ASMR videos, which often entail role plays, require imaginatively transposing oneself into the virtual reality (Janik McErlean & Banissy, 2017). PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed conceptualised as a trait, whereby individuals with the so called autotelic personality (who are intrinsically motivated) are more susceptible to experiencing flow (Csikszentmihalyi, 2000). To date several measures have been developed which allow for measuring flow both as a state and a trait (e.g. Jackson & Eklund, 2002, 2004). Barratt and Davis (2015) suggested that ASMR is a ‘flow-like’ phenomenon achieved by watching others in a similar state. Interestingly, some of the most popular ASMR triggers, such as watching someone making expert hand movements, are typical examples of being in a state of flow (Janik McErlean & Banissy, 2017). Moreover, individuals with greater susceptibility to flow have been found to report a greater number of ASMR triggers highlighting a link between the two phenomena (Barratt & Davis, 2015). Furthermore, a positive association between flow and a newly developed ASMR measure has been recently reported (Roberts, Beath, & Boag, 2019). However, whether ASMR is associated with increased levels of flow both in terms of intensity and prevalence compared to the general population is currently not known. In addition, both Barratt and Davis (2015) and Roberts et al., (2019) used a modified version of the Flow State Scale (Jackson & Marsh, 1996) which only taps into the passive aspects of flow. While this measure is more likely to be relevant in the ASMR context which does not entail active engagement in activities, it does not capture the other core components of the flow experience such as the balance between challenge and skill. As such no study to date examined whether ASMR is associated with increased levels of flow using the global measure of this construct. 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 conceptualised as a trait, whereby individuals with the so called autotelic personality (who are intrinsically motivated) are more susceptible to experiencing flow (Csikszentmihalyi, 2000). To date several measures have been developed which allow for measuring flow both as a state and a trait (e.g. Jackson & Eklund, 2002, 2004). Barratt and Davis (2015) suggested that ASMR is a ‘flow-like’ phenomenon achieved by watching others in a similar state. Interestingly, some of the most popular ASMR triggers, such as watching someone making expert hand movements, are typical examples of being in a state of flow (Janik McErlean & Banissy, 2017). Manuscript to be reviewed Moreover, individuals with greater susceptibility to flow have been found to report a greater number of ASMR triggers highlighting a link between the two phenomena (Barratt & Davis, 2015). 70 71 72 73 74 75 76 77 78 Furthermore, a positive association between flow and a newly developed ASMR measure has been recently reported (Roberts, Beath, & Boag, 2019). However, whether ASMR is associated with increased levels of flow both in terms of intensity and prevalence compared to the general population is currently not known. In addition, both Barratt and Davis (2015) and Roberts et al., (2019) used a modified version of the Flow State Scale (Jackson & Marsh, 1996) which only taps into the passive aspects of flow. While this measure is more likely to be relevant in the ASMR context which does not entail active engagement in activities, it does not capture the other core components of the flow experience such as the balance between challenge and skill. As such no study to date examined whether ASMR is associated with increased levels of flow using the global measure of this construct. 79 80 81 82 83 84 85 86 87 88 The reported link between ASMR and flow is also interesting in the context of increased mindfulness among ASMR-experiencers (Fredborg et al., 2018). Mindfulness, similarly to flow, can be conceptualised as a trait and as a state. In Fredborg et al., (2018) study ASMR- 89 90 91 The reported link between ASMR and flow is also interesting in the context of increased mindfulness among ASMR-experiencers (Fredborg et al., 2018). Mindfulness, similarly to flow, can be conceptualised as a trait and as a state. In Fredborg et al., (2018) study ASMR- experiencers were found to score higher than controls in terms of trait mindfulness based on their scores on the Mindful Attention Awareness Scale which taps into one’s general disposition to pay attention to and be aware of the present moment (Brown & Ryan, 2003). They also scored higher 89 90 91 92 93 94 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) experiencers were found to score higher than controls in terms of trait mindfulness based on their scores on the Mindful Attention Awareness Scale which taps into one’s general disposition to pay attention to and be aware of the present moment (Brown & Ryan, 2003). Manuscript to be reviewed 53 54 55 56 57 58 59 60 61 62 63 64 ASMR has also been previously likened to a state of flow (Barratt & Davis, 2015), which can be measured both as a global construct and in terms of its underlying components including warped passing of time (Ross & Keiser, 2014), complete absorption (Jackson, Thomas, Marsh, & 65 66 67 ASMR has also been previously likened to a state of flow (Barratt & Davis, 2015), which can be measured both as a global construct and in terms of its underlying components including warped passing of time (Ross & Keiser, 2014), complete absorption (Jackson, Thomas, Marsh, & 65 66 67 Smethurst, 2001), and intense concentration when fully engaged in optimally challenging and intrinsically rewarding tasks (Csikszentmihalyi & Csikszentmihalyi, 1992). Flow has also been 68 69 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed They also scored higher 92 93 94 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) The purpose of this study was to investigate whether there are differences between ASMR- experiencers and controls in terms of flow, absorption and mindfulness. This study also aimed to elucidate the relationship between these constructs and ASMR characteristics such as intensity, pleasantness, and ASMR videos viewing habits. 119 120 121 122 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed on the Curiosity subscale of the state mindfulness measure called the Toronto Mindfulness Scale which suggests increased interest in one’s own inner experiences among the ASMR group (Lau et al., 2006). Although mindfulness and flow are similar in terms of concentrating on the present experience, they differ dramatically in terms of the role of self-awareness which is enhanced in mindfulness and diminished in flow where loss of self-consciousness is the central feature (Bishop, 2002; Bishop et al., 2004; Nakamura & Csikszentmihalyi, 2009). Moreover, although existing research provides support for the positive association between the global measure of flow comprising all underlying dimensions and mindfulness (Thienot et al., 2014), recent findings suggest that this is driven only by the control facet of flow which reflects a sense of agency and mastery over a task, whereas absorption which is another core underlying dimension of flow is in fact negatively related to mindfulness (Sheldon, Prentice, & Halusic, 2015). 95 96 97 98 99 100 101 102 103 104 105 Absorption as a key facet of flow is a trance-like state of consciousness characterised by an ability to fully focus one’s attention on a particular object or situation and to become 106 107 Absorption as a key facet of flow is a trance-like state of consciousness characterised by an ability to fully focus one’s attention on a particular object or situation and to become 106 107 perceptually engrossed with the current experience (Tellegen, 1981, 1982; Tellegen & Atkinson, 1974). It is also a stable personality trait typically measured with the Tellegen Absorption Scale (Tellegen & Atkinson, 1974). Absorption has also been linked to hypnotisability, imagery, day- dreaming, and openness to experience (Weilbel et al., 2010; Glisky, Tataryn, Tobias, Kihlstrom, & McConkey, 1991). Recently, an association between absorption propensity and ASMR has also been reported (Roberts et al., 2019). Taking into account existing findings and the 108 109 110 111 112 113 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed The purpose of this study was to investigate whether there are differences between ASMR- experiencers and controls in terms of flow, absorption and mindfulness. This study also aimed to elucidate the relationship between these constructs and ASMR characteristics such as intensity, pleasantness, and ASMR videos viewing habits. 119 120 121 122 Manuscript to be reviewed Manuscript to be reviewed gender and then within gender on age +/- 2 years. Majority was matched within +/- 1 year and four participants who identified as non-binary in terms of gender were matched within +/- 4 years. 143 144 145 gender and then within gender on age +/- 2 years. Majority was matched within +/- 1 year and four participants who identified as non-binary in terms of gender were matched within +/- 4 years. 143 144 145 years. In addition, data contributed by those ASMR-experiencers who stated they watch ASMR videos (N= 149, 121 female, 26 male, 2 other, age M = 26.22, SD = 8.12) was used in the correlational analysis which aimed to examine the relationship between absorption, flow and mindfulness and ASMR characteristics including pleasantness, intensity, number of videos watched in a single session and frequency of watching ASMR videos. Measures The study employed an anonymous survey which was administered via Bristol Online Survey and was approved by the Ethics Reviewers at Bath Spa University. All participants were asked to provide an electronic consent prior to taking this survey. Tellegen Absorption Scale (TAS) employed in this study is a widely used measure of absorption (Tellegen & Atkinson, 1974). It consists of 34 questions and the participants are asked to indicate the degree to which they agree with each statement such as ‘While watching a movie, TV show or a play, I may become so involved that I may forget about myself and my surroundings and experience the story as if it were real and as if I were taking part in it’ on a scale from 1 (disagree strongly) to 5 (agree strongly). A total score is calculated by summing the scores for each question. Internal consistency of TAS in this study was α = .93. Mindful Attention and Awareness Scale (MAAS) is an established measure of mindfulness (Brown & Ryan, 2003). It consists of 15 items such as ‘I could be experiencing some emotion and not be conscious of it until sometime later’. Participants are asked to indicate how frequently they have such experiences on a scale from 1 (almost always) to 6 (almost never). Manuscript to be reviewed PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Materials & Methods 123 p Participants were recruited via websites dedicated to ASMR and among Psychology students, who were offered credits for their participation. Participants were asked to indicate whether they would classify themselves as controls or ASMR-experiencers based on the description of the phenomenon (as per Janik McErlean & Banissy, 2017). To further verify that participants in the ASMR group were genuine ASMR-experiencers they were asked to provide a series of answers related to their ASMR experience (see Questionnaire). E.g. they were asked to indicate the intensity of their ASMR when engaging with various stimuli. If a participant indicated they did not experience ASMR in response to any of the popular triggers they were excluded from the analysis. There were only two such individuals, who also happened to be outliers in terms of their age. They were excluded from the analysis. A priori power analysis revealed that suggested sample size to conduct MANOVA was 232 participants in order to obtain statistical power at 0.90 level (Effect size = 0.06, α = 0.05). Through opportunity sampling, a total of 316 participants completed the survey (59% = ASMR-experiencers). In order to reduce the imbalance between the group sizes and to match the two groups in terms of age and gender 124 ASMR-experiencers (92 female, 30 male, 2 other, age M = 21.79, SD = 3.36) and 124 controls (92 female, 30 male, 2 other, age M = 21.40, SD = 3.13) were selected from the total sample prior to data analysis resulting in the final sample of 248 individuals which was used to compare the groups on absorption, flow and mindfulness. These participants were first matched on 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 Manuscript to be reviewed The final MAAS score is calculated by averaging all individual answers. The internal consistency of MAAS in this study was α = .79. 167 168 The final MAAS score is calculated by averaging all individual answers. The internal consistency of MAAS in this study was α = .79. 167 168 consistency of MAAS in this study was α = .79. Flow Questionnaire: Firstly, participants were presented with the description of flow (as per Csikszentmihalyi & Csikszentmihalyi, 1988; Csikszentmihalyi, 1990) based on which they indicated whether they have ever had similar experiences which allows for classifying individuals into those capable of experiencing flow and those who are not. In order to avoid false positive answers those who reported to experience flow were also asked to state what activities they were engaged in when having such experiences. Subsequently, they were asked to complete the Flow Experiences Scale (FES; Schwartz & Waterman, 2006) which consists of eight statements corresponding to those originally outlined by Csikszentmihalyi and Csikszentmihalyi (1988) and Csikszentmihalyi (1990) which inquire about different aspects of the experience of flow including e.g. ‘I lose track of time’ to which participants provide a rating on a scale from 1- not at all characteristic of me to 7- very characteristic of me. A total score reflecting the extent to which flow is experienced is calculated by summing the scores for each statement. Question 2 is reverse scored. The internal consistency for FES in this study was α = .74. 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 Flow Questionnaire: Firstly, participants were presented with the description of flow (as per Csikszentmihalyi & Csikszentmihalyi, 1988; Csikszentmihalyi, 1990) based on which they indicated whether they have ever had similar experiences which allows for classifying individuals into those capable of experiencing flow and those who are not. In order to avoid false positive answers those who reported to experience flow were also asked to state what activities they were engaged in when having such experiences. Subsequently, they were asked to complete the Flow Experiences Scale (FES; Schwartz & Waterman, 2006) which consists of eight statements corresponding to those originally outlined by Csikszentmihalyi and Csikszentmihalyi (1988) and Csikszentmihalyi (1990) which inquire about different aspects of the experience of flow including e.g. Manuscript to be reviewed 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 In addition, data contributed by those ASMR-experiencers who stated they watch ASMR videos (N= 149, 121 female, 26 male, 2 other, age M = 26.22, SD = 8.12) was used in the correlational analysis which aimed to examine the relationship between absorption, flow and mindfulness and ASMR characteristics including pleasantness, intensity, number of videos watched in a single session and frequency of watching ASMR videos. 146 147 148 149 150 The study employed an anonymous survey which was administered via Bristol Online Survey and was approved by the Ethics Reviewers at Bath Spa University. All participants were asked to provide an electronic consent prior to taking this survey. 153 154 155 Tellegen Absorption Scale (TAS) employed in this study is a widely used measure of absorption (Tellegen & Atkinson, 1974). It consists of 34 questions and the participants are asked to indicate the degree to which they agree with each statement such as ‘While watching a movie, TV show or a play, I may become so involved that I may forget about myself and my surroundings and experience the story as if it were real and as if I were taking part in it’ on a scale from 1 (disagree strongly) to 5 (agree strongly). A total score is calculated by summing the scores for each question. Internal consistency of TAS in this study was α = .93. 156 157 158 159 160 161 162 Mindful Attention and Awareness Scale (MAAS) is an established measure of mindfulness (Brown & Ryan, 2003). It consists of 15 items such as ‘I could be experiencing some emotion and not be conscious of it until sometime later’. Participants are asked to indicate how frequently they have such experiences on a scale from 1 (almost always) to 6 (almost never). 163 164 165 166 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be review Manuscript to be reviewed Instead three separate independent samples t-tests were conducted to compare the groups on the three personality traits. Comparison of ASMR-Experiencers and Controls on Personality Traits Three independent-samples t-tests revealed: i) a non-significant group difference t(104) = - 0.990, p = .324, d = 0.19, 95% CI [-4.807; 1.606] between ASMR-experiencers (M=36.49, SD=8.16) and controls (M=38.09, SD=8.19) in terms of flow, ii) a non-significant group difference on mindfulness t(246) = -0.715, p = .475, d = 0.09, 95% CI [-0.273; 0.127] between ASMR-experiencers (M=3.20, SD=0.80) and controls (M=3.28, SD=0.79), iii) a statistically significant group difference on absorption t(246) = 4.995, p < .001, d = 0.63, 95% CI[8.979; 20.252] with ASMR-experiencers (M=108.74, SD=23.80) scoring higher than controls (M=94.22, SD=21.95) (see Table 1 for Descriptive Statistics). Furthermore, chi-square analysis revealed that the proportion of individuals who reported to experience flow was significantly greater [ = 6.590, p = .010] among ASMR-Experiencers (51%) compared to controls (34%). 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 Comparison of ASMR-Experiencers and Controls on Personality Traits Three independent-samples t-tests revealed: i) a non-significant group difference t(104) = - 0.990, p = .324, d = 0.19, 95% CI [-4.807; 1.606] between ASMR-experiencers (M=36.49, SD=8.16) and controls (M=38.09, SD=8.19) in terms of flow, ii) a non-significant group difference on mindfulness t(246) = -0.715, p = .475, d = 0.09, 95% CI [-0.273; 0.127] between ASMR-experiencers (M=3.20, SD=0.80) and controls (M=3.28, SD=0.79), iii) a statistically significant group difference on absorption t(246) = 4.995, p < .001, d = 0.63, 95% CI[8.979; 20.252] with ASMR-experiencers (M=108.74, SD=23.80) scoring higher than controls (M=94.22, SD=21.95) (see Table 1 for Descriptive Statistics). Manuscript to be reviewed ‘I lose track of time’ to which participants provide a rating on a scale from 1- not at all characteristic of me to 7- very characteristic of me. A total score reflecting the extent to which flow is experienced is calculated by summing the scores for each statement. Question 2 is reverse scored. The internal consistency for FES in this study was α = .74. 169 170 171 172 173 174 175 176 177 178 179 180 181 182 ASMR Questionnaire: In addition, ASMR-experiencers completed the self-designed ASMR questionnaire which was based on items previously used in ASMR literature (Barratt & Davis, 2015; Janik McErlean & Banissy, 2017; Fredborg et al., 2018). Specifically, participants were asked: 1) whether they watch ASMR videos (Yes/No), 2) why they watch ASMR videos (open ended), 3) how often they watch ASMR videos: never, less than once a month, 2-3 times a month, 2-3 times a week, daily, 4) whether they require specific conditions to achieve ASMR (open ended), 5) to indicate the intensity of common triggers such as crisp sounds, personal attention etc. using a scale from 0 (no tingles) to 6 (the most intense tingles) – ASMR intensity score was calculated by averaging the scores across all listed triggers as per Fredborg et al. 183 184 185 186 187 188 189 190 191 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) ASMR Questionnaire: In addition, ASMR-experiencers completed the self-designed ASMR questionnaire which was based on items previously used in ASMR literature (Barratt & Davis, 2015; Janik McErlean & Banissy, 2017; Fredborg et al., 2018). Specifically, participants were asked: 1) whether they watch ASMR videos (Yes/No), 2) why they watch ASMR videos (open ended), 3) how often they watch ASMR videos: never, less than once a month, 2-3 times a month, 2-3 times a week, daily, 4) whether they require specific conditions to achieve ASMR (open ended), 5) to indicate the intensity of common triggers such as crisp sounds, personal attention etc. using a scale from 0 (no tingles) to 6 (the most intense tingles) – ASMR intensity score was calculated by averaging the scores across all listed triggers as per Fredborg et al. Manuscript to be reviewed (2018), 6) whether the intensity of these triggers varies from session to session (Yes/No); 7) to indicate how pleasurable ASMR is from 1- Quite Uncomfortable to 5 - Quite pleasurable. Open ended questions were coded prior to the analysis and the total intensity score was computed by averaging the scores for all listed triggers. 192 193 194 195 (2018), 6) whether the intensity of these triggers varies from session to session (Yes/No); 7) to indicate how pleasurable ASMR is from 1- Quite Uncomfortable to 5 - Quite pleasurable. Open ended questions were coded prior to the analysis and the total intensity score was computed by averaging the scores for all listed triggers. 192 193 194 195 indicate how pleasurable ASMR is from 1- Quite Uncomfortable to 5 - Quite pleasurable. Open ended questions were coded prior to the analysis and the total intensity score was computed by averaging the scores for all listed triggers. Results Data Inspection and Assumptions Testing The data was normally distributed and no outliers were identified as all values fell within +/- 2.5 SD from the mean. However, as only 106 out of 248 (43%) participants reported to experience flow it was not possible to conduct MANOVA due to the insufficient sample size. Instead three separate independent samples t-tests were conducted to compare the groups on the three personality traits. Comparison of ASMR-Experiencers and Controls on Personality Traits Three independent-samples t-tests revealed: i) a non-significant group difference t(104) = - 0.990, p = .324, d = 0.19, 95% CI [-4.807; 1.606] between ASMR-experiencers (M=36.49, SD=8.16) and controls (M=38.09, SD=8.19) in terms of flow, ii) a non-significant group difference on mindfulness t(246) = -0.715, p = .475, d = 0.09, 95% CI [-0.273; 0.127] between ASMR-experiencers (M=3.20, SD=0.80) and controls (M=3.28, SD=0.79), iii) a statistically significant group difference on absorption t(246) = 4.995, p < .001, d = 0.63, 95% CI[8.979; 20.252] with ASMR-experiencers (M=108.74, SD=23.80) scoring higher than controls (M=94.22, SD=21.95) (see Table 1 for Descriptive Statistics). Furthermore, chi-square analysis revealed that the proportion of individuals who reported to experience flow was significantly greater [ = 6.590, p = .010] among ASMR-Experiencers (51%) compared to controls (34%). 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 experience flow it was not possible to conduct MANOVA due to the insufficient sample size. PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed 183 184 185 186 187 188 189 190 191 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) (2018), 6) whether the intensity of these triggers varies from session to session (Yes/No); 7) to indicate how pleasurable ASMR is from 1- Quite Uncomfortable to 5 - Quite pleasurable. Open ended questions were coded prior to the analysis and the total intensity score was computed by averaging the scores for all listed triggers. Results Data Inspection and Assumptions Testing The data was normally distributed and no outliers were identified as all values fell within +/- 2.5 SD from the mean. However, as only 106 out of 248 (43%) participants reported to experience flow it was not possible to conduct MANOVA due to the insufficient sample size. Instead three separate independent samples t-tests were conducted to compare the groups on the three personality traits. Comparison of ASMR-Experiencers and Controls on Personality Traits Three independent-samples t-tests revealed: i) a non-significant group difference t(104) = - 0.990, p = .324, d = 0.19, 95% CI [-4.807; 1.606] between ASMR-experiencers (M=36.49, SD=8.16) and controls (M=38.09, SD=8.19) in terms of flow, ii) a non-significant group difference on mindfulness t(246) = -0.715, p = .475, d = 0.09, 95% CI [-0.273; 0.127] between ASMR-experiencers (M=3.20, SD=0.80) and controls (M=3.28, SD=0.79), iii) a statistically significant group difference on absorption t(246) = 4.995, p < .001, d = 0.63, 95% CI[8.979; 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 Manuscript to be reviewe Manuscript to be reviewed Manuscript to be reviewed Manuscript to be reviewed Furthermore, chi-square analysis 204 205 206 207 208 209 210 211 212 213 Comparison of ASMR-Experiencers and Controls on Personality Traits 205 Comparison of ASMR-Experiencers and Controls on Personality Traits Three independent-samples t-tests revealed: i) a non-significant group difference t(104) = - 0.990, p = .324, d = 0.19, 95% CI [-4.807; 1.606] between ASMR-experiencers (M=36.49, SD=8.16) and controls (M=38.09, SD=8.19) in terms of flow, ii) a non-significant group difference on mindfulness t(246) = -0.715, p = .475, d = 0.09, 95% CI [-0.273; 0.127] between ASMR-experiencers (M=3.20, SD=0.80) and controls (M=3.28, SD=0.79), iii) a statistically significant group difference on absorption t(246) = 4.995, p < .001, d = 0.63, 95% CI[8.979; 20.252] with ASMR-experiencers (M=108.74, SD=23.80) scoring higher than controls (M=94.22, SD=21.95) (see Table 1 for Descriptive Statistics). Furthermore, chi-square analysis revealed that the proportion of individuals who reported to experience flow was significantly greater [ = 6.590, p = .010] among ASMR-Experiencers (51%) compared to controls (34%). 205 206 207 208 209 210 211 212 213 214 215 216 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) An additional analysis was conducted using the whole data set rather than just the subsample of ASMR-experiencers and control participants who were matched for age and gender. This analysis revealed qualitatively similar results to the ones reported above (see Supplemental Results). Correlations between ASMR features and personality traits Pearson’s correlations employed to examine the relationship between absorption, flow and mindfulness and ASMR characteristics revealed a positive relationship between: i) absorption and the intensity of ASMR (r(149) = .180, p = .028), ii) intensity and frequency of watching ASMR videos (r(149) = .354, p < .001), iii) intensity and pleasantness (r(149) = .389, p < .001) and iv) pleasantness and frequency of watching ASMR videos (r(149) = .233, p = .004. No other correlations were statistically significant (Table 2; Figure 1). Discussion 250 The purpose of this study was to further elucidate whether ASMR is associated with wider differences in terms of personality traits. Current results demonstrate that ASMR-experiencers show elevated absorption compared to age and gender matched controls, but no group differences were found in terms of mindfulness or flow. This suggests that ASMR-experiencers display increased readiness for experiential involvement and heightened ability to become fully engaged with the current experience (Tellegen, 1981; Jamieson, 2005). This is interesting taking into account that being immersed in the virtual reality appears to be a key component of technologically-mediated ASMR and that increased absorption has been previously linked to a more immersive virtual reality experience (Baños et al., 1999). Moreover, increased absorption has been associated with elevated openness to experience (Weilbel, Wissmath, & Mast, 2010; Glisky et al., 1991) which is also heightened among ASMR-experiencers (Janik McErlean & Banissy, 2017; Fredborg et al., 2017). ASMR has also been linked to another trait relevant to technologically-mediated ASMR i.e. increased fantasizing (Janik McErlean & Banissy, 2017) which reflects heightened propensity to become imaginatively involved in a fictional reality. As such, the combination of increased absorption, openness to experience and fantasizing may 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 The purpose of this study was to further elucidate whether ASMR is associated with wider differences in terms of personality traits. Current results demonstrate that ASMR-experiencers show elevated absorption compared to age and gender matched controls, but no group differences were found in terms of mindfulness or flow. This suggests that ASMR-experiencers display increased readiness for experiential involvement and heightened ability to become fully engaged with the current experience (Tellegen, 1981; Jamieson, 2005). This is interesting taking into account that being immersed in the virtual reality appears to be a key component of 251 252 253 254 255 256 257 technologically-mediated ASMR and that increased absorption has been previously linked to a more immersive virtual reality experience (Baños et al., 1999). Moreover, increased absorption has been associated with elevated openness to experience (Weilbel, Wissmath, & Mast, 2010; Glisky et al., 1991) which is also heightened among ASMR-experiencers (Janik McErlean & Banissy, 2017; Fredborg et al., 2017). ASMR has also been linked to another trait relevant to technologically-mediated ASMR i.e. increased fantasizing (Janik McErlean & Banissy, 2017) which reflects heightened propensity to become imaginatively involved in a fictional reality. Manuscript to be reviewed Summary of responses to ASMR Questionnaire The main motivation behind watching ASMR videos was to relax (71%), followed by to fall 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 Manuscript to be review Summary of responses to ASMR Questionnaire The main motivation behind watching ASMR videos was to relax (71%), followed by to fall asleep (60.1%), to experience ASMR (54.1%), to improve mood, especially in relation to anxiety and depression (12.8%), and to help concentrate on work or a task (6.8%). On average, participants watched 3.3 videos per session. 42.3% reported to watch ASMR videos daily, followed by to 2-3 times a week (30.9%), 2- 3 times a month (19.5%), and less than once a month (7.4%). 49% of participants stated that they needed specific conditions to experience ASMR. This included: a quiet room (71.23%), dim lighting (30.14%), to be alone or have no distractions (26.03%), to be wearing headphones (20.55%), to be in bed or lying down (13.7%), to have the room at a specific temperature (12.33%) to be in a real life setting (8.22%), comfort (8.22%), which is largely consistent with the findings of Barratt, Spence, & Davis (2017). In terms of the intensity on average whispering (M = 4.03) was rated to produce the most intense 230 231 232 233 234 235 236 237 238 239 240 241 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed ASMR, followed by personal attention (M = 3.65), crisp sounds (M = 3.57), paying attention to detail, concentrating on something, slowly performing mundane actions or explaining something (M = 3.43), role-play (M = 3.15), hair brushing (M = 3.04) and lastly people eating (M = 1.34). 94% reported that the intensity of ASMR varies from session to session. The majority of participants (60.4%) gave ASMR the highest rating in terms of how pleasurable it is, followed by 35.6% rating ASMR as mildly pleasant, 2.7% as neutral, one participant reported ASMR to be quite uncomfortable (0.7%) and another one as mildly uncomfortable (0.7%). 242 243 244 245 246 247 248 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) contribute to the individual likelihood of experiencing ASMR. Moreover, when considering those individuals who consume ASMR media to alleviate stress and anxiety it is possible that for them becoming fully absorbed in ASMR videos may act as a form of distraction from psychological distress which is similar to the well documented effect of virtual reality gaming and other immersive methods as effective pain reduction strategies (Dahlquist et al., 2007; Jameson, Trevena, & Swain, 2011). 266 267 268 269 270 271 contribute to the individual likelihood of experiencing ASMR. Moreover, when considering those individuals who consume ASMR media to alleviate stress and anxiety it is possible that for them becoming fully absorbed in ASMR videos may act as a form of distraction from psychological distress which is similar to the well documented effect of virtual reality gaming and other immersive methods as effective pain reduction strategies (Dahlquist et al., 2007; Jameson, Trevena, & Swain, 2011). 266 267 268 269 270 271 The current study also found that ASMR-experiencers did not differ from controls in terms of mindfulness, which is consistent with our results of increased absorption among ASMR- experiencers and the reports of conceptual incompatibility between the two constructs (Brown & Ryan, 2003; Sheldon et al., 2015). Moreover, the correlational analysis employed in this study revealed no association between mindfulness and absorption further suggesting little overlap between these psychological variables. It therefore appears that it is the total immersion in the experience accompanied by a loss of self-awareness, which are core features of absorption, rather than sustained consciousness of the current moment characteristic of mindfulness that are features relevant to the ASMR experience. However, it is of note that these results are inconsistent with previous findings of increased mindfulness in ASMR (Fredborg et al., 2018). As both studies employed the same scale and participant recruitment method it is not clear why different results were obtained. Although the sample size in this study was not as large as in the other one, it was sufficiently large as determined by a priori power analysis. However, it appears that the control participants sampled by Fredborg et al., (2018) scored particularly low on MAAS (M = 3.02, SD = 1.17) compared to the control participants in the present study (M = 3.28, SD = 0.79), which might explain the difference in results between the two studies. Moreover, both the current and Fredborg et al., (2018)’s results pertaining to the control sample are much lower than originally reported by the MAAS authors who found average MAAS scores 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 The current study also found that ASMR-experiencers did not differ from controls in terms of mindfulness, which is consistent with our results of increased absorption among ASMR- experiencers and the reports of conceptual incompatibility between the two constructs (Brown & Ryan, 2003; Sheldon et al., 2015). Moreover, the correlational analysis employed in this study revealed no association between mindfulness and absorption further suggesting little overlap between these psychological variables. It therefore appears that it is the total immersion in the experience accompanied by a loss of self-awareness, which are core features of absorption, rather than sustained consciousness of the current moment characteristic of mindfulness that are features relevant to the ASMR experience. However, it is of note that these results are inconsistent with previous findings of increased mindfulness in ASMR (Fredborg et al., 2018). As both studies employed the same scale and participant recruitment method it is not clear why different results were obtained. Although the sample size in this study was not as large as in the other one, it was sufficiently large as determined by a priori power analysis. However, it appears that the control participants sampled by Fredborg et al., (2018) scored particularly low on MAAS (M = 3.02, SD = 1.17) compared to the control participants in the present study (M = 3.28, SD = 0.79), which might explain the difference in results between the two studies. Discussion 250 As such, the combination of increased absorption, openness to experience and fantasizing may 258 259 260 261 262 263 264 265 Manuscript to be reviewed PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) to be M =3.85 (SD = 0.68) among undergraduate students and M = 3.97 (SD = 0.64) in a community sample (Brown & Ryan, 2003). It is not clear what drives this discrepancy in findings between the three studies. As such, the potential link between ASMR and mindfulness should be further explored in future research. 290 291 292 293 to be M =3.85 (SD = 0.68) among undergraduate students and M = 3.97 (SD = 0.64) in a community sample (Brown & Ryan, 2003). It is not clear what drives this discrepancy in findings between the three studies. As such, the potential link between ASMR and mindfulness should be further explored in future research. 290 291 292 293 Furthermore, current results show similar levels of flow between ASMR-experiencers and controls. This is consistent with existing literature demonstrating an association between the global flow construct and mindfulness (Thienot et al., 2014) and current findings of no group differences on either of these traits. This is also in line with the original conceptualisation of the construct which emphasises that active engagement in activities is necessary for the experience of flow (Csikszentmihalyi & Csikszentmihalyi, 1992) and research showing that flow has been most commonly related to activities that are effortful and based on challenge, skill and intrinsic motivation (Mauri, Cipresso, Balgera, Villamira, & Riva, 2011) whilst the experience of ASMR is typically elicited through passively watching videos or by observing people in real life who engage in ASMR-inducing activities. Therefore, ASMR appears to only reflect the passive aspects of flow, which are more consistent with the construct of absorption (Jackson et al., 2001). This is also in line with Barratt and Davis (2015) and Roberts et al., (2019) studies who measured flow in terms of its passive component which was found to be positively associated with the ASMR experience. Moreover, these results also fit with previous reports of elevated fantasising and imaginative involvement in ASMR (Janik McErlean & Banissy, 2017) and the fact that fantasy engagement and imagination are key characteristics of absorption but not flow (Nakamura & Csikszentmihalyi, 2014). However, it is of note that substantially more ASMR- experiencers reported to experience flow (51%) compared to controls (34%) suggesting that while the extent to which both groups experience flow is similar, flow appears to be more prevalent among ASMR-experiencers. Moreover, both the current and Fredborg et al., (2018)’s results pertaining to the control sample 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 Moreover, both the current and Fredborg et al., (2018)’s results pertaining to the control sample are much lower than originally reported by the MAAS authors who found average MAAS scores 288 289 Moreover, both the current and Fredborg et al., (2018)’s results pertaining to the control sample are much lower than originally reported by the MAAS authors who found average MAAS scores 288 289 Moreover, both the current and Fredborg et al., (2018)’s results pertaining to the control sample are much lower than originally reported by the MAAS authors who found average MAAS scores 288 289 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed experience flow when engaged in sport, music, art or when concentrating on a work-related task which are typical examples of flow-inducing activities (Csikszentmihalyi & Rathunde, 1993) and only five individuals reported to experience flow when consuming ASMR media. This suggests that ASMR group appears to experience flow more readily than controls but rarely in an ASMR context. However, it is of note that although previous research is quite inconsistent when it comes to the prevalence of flow, the percentage of individuals identifying with the experience in this study (both ASMR-experiencers and controls) is comparably low. For instance, Moneta (2012) reports that about two-thirds of the general population are capable of experiencing flow while some of the earlier studies suggest the experience to be universal with prevalence rates of 97% among university students (Massimini, Csikszentmihalyi, & Delle Fave, 1988). Although, Han (1988) reports that only 33% of elderly Korean immigrants in America identified with the experience suggesting potential effects of age and culture. 314 315 316 317 318 319 320 321 322 323 324 325 experience flow when engaged in sport, music, art or when concentrating on a work-related task which are typical examples of flow-inducing activities (Csikszentmihalyi & Rathunde, 1993) and only five individuals reported to experience flow when consuming ASMR media. This suggests that ASMR group appears to experience flow more readily than controls but rarely in an ASMR context. However, it is of note that although previous research is quite inconsistent when it comes to the prevalence of flow, the percentage of individuals identifying with the experience in this study (both ASMR-experiencers and controls) is comparably low. For instance, Moneta (2012) reports that about two-thirds of the general population are capable of experiencing flow while some of the earlier studies suggest the experience to be universal with prevalence rates of 97% among university students (Massimini, Csikszentmihalyi, & Delle Fave, 1988). Although, Han (1988) reports that only 33% of elderly Korean immigrants in America identified with the experience suggesting potential effects of age and culture. 314 315 316 317 318 319 320 321 322 323 324 325 This study also found that absorption was positively correlated with the intensity of ASMR suggesting that this trait is linked to the extent to which ASMR is experienced. However, no association was found between flow and mindfulness and ASMR features, further demonstrating that these traits are not related to ASMR. PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Interestingly, the majority of ASMR sample reported to 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 Manuscript to be reviewed 75% of the ASMR sample whose data was used for comparison purposes and overall 80% of the entire ASMR sample in this study (not just the ones matched for age and gender) reported to watch ASMR-videos. However, we have found that ASMR-experiencers who watch ASMR videos (M = 108.31, M = 27.27) and ASMR-experiencers who do not watch such videos (M = 108.04, SD = 25.679) reported exactly the same level absorption suggesting that our results pertain also to those ASMR-experiencers who do not engage with ASMR media. 338 339 340 341 342 343 In addition, consistently with previous research this study has found that the key reasons for watching ASMR videos were to relax, fall asleep, to experience ASMR, and to help with stress or anxiety (Barratt & Davis, 2015; Janik McErlean & Banissy, 2017). In addition, 6.7% of participants reported to play ASMR videos in the background while studying or working. Moreover, 49% of participants stated that they needed specific conditions, such as a quite space with few distractions, in order to experience ASMR which is consistent with the previous studies (Barratt & Davis, 2015). 344 345 346 347 348 349 350 Manuscript to be reviewed In addition to this, our results suggest that those who experience ASMR more intensely find it more pleasurable and also engage with ASMR media more frequently. However, it is important to note that due to the correlational nature of this research it is not possible to conclude that the trait of absorption predisposes people to experience ASMR or that the ability to experience ASMR increases the propensity for absorption. It is also possible that both absorption and ASMR are underlined by a third variable such as openness to experience. 326 327 328 329 330 331 332 333 334 335 This study also found that absorption was positively correlated with the intensity of ASMR suggesting that this trait is linked to the extent to which ASMR is experienced. However, no association was found between flow and mindfulness and ASMR features, further demonstrating that these traits are not related to ASMR. In addition to this, our results suggest that those who experience ASMR more intensely find it more pleasurable and also engage with ASMR media more frequently. However, it is important to note that due to the correlational nature of this research it is not possible to conclude that the trait of absorption predisposes people to experience ASMR or that the ability to experience ASMR increases the propensity for absorption. It is also possible that both absorption and ASMR are underlined by a third variable such as openness to experience. 326 327 328 329 330 331 332 333 334 335 Furthermore, although one may suspect that the self-selection bias may have contributed to the current results of increased absorption among ASMR-experiencers compared to controls as 336 337 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Furthermore, although one may suspect that the self-selection bias may have contributed to the current results of increased absorption among ASMR-experiencers compared to controls as 336 337 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Conclusions 351 In summary, our findings contribute to the existing literature documenting individual differences associated with ASMR by showing that absorption proneness may be an important factor contributing to the experience of ASMR. This study also shows that ASMR is not linked to the constructs of flow and mindfulness. 352 353 354 355 In summary, our findings contribute to the existing literature documenting individual differences associated with ASMR by showing that absorption proneness may be an important factor contributing to the experience of ASMR. This study also shows that ASMR is not linked to the constructs of flow and mindfulness. 352 353 354 355 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Manuscript to be reviewed Baños, R., Botella, C., García-Palacios, A., Villa, H., Perpiñá, C., & Gallardo, M. (1999). 358 Psychological variables and reality judgment in virtual environments: The roles of absorption and dissociation. CyberPsychology & Behavior, 2(2), 143-148. 359 360 Barratt, E. L., & Davis, N. J. (2015). Autonomous Sensory Meridian Response (ASMR): a flow- like mental state. PeerJ, 3, e851. 361 362 Barratt, E., L., Spence, C., & Davis, N. J. (2017). Sensory determinants of the autonomous 363 sensory meridian response (ASMR): understanding the triggers. PeerJ, 5:e3846. 364 Bishop, S.R. (2002). What do we really know about mindfulness-based stress reduction? Psychosomatic Medicine, 64(1), 71-83. 365 366 Bishop, S.R. (2002). What do we really know about mindfulness-based stress reduction? 365 Psychosomatic Medicine, 64(1), 71-83. 366 Bishop, S.R., Lau, M., Shapiro, S., Carlson, L., Anderson, N.D., Carmody, J., Segal, Z.V., 367 Abbey, S., Speca, M., Velting, D., & Devins, G. (2004). Mindfulness: a proposed operational definition. Clinical Psychology: Science and Practice, 11(3), 230-241. 368 369 definition. Clinical Psychology: Science and Practice, 11(3), 230-241. 69 Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfulness and its role in 370 Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfuln 70 psychological well-being. Journal of Personality and Social Psychology, 8 71 psychological well-being. Journal of Personality and Social Psychology, 84, 822 - 848. 371 Csikszentmihalyi, M. (1988). The future of flow. In M. Csikszentmihalyi & I. S. 372 Csikszentmihalyi, M. (1988). The future of flow. In M. Csikszentmihalyi & 372 Csikszentmihalyi (Eds.), Optimal experience: Psychological studies of flow in consciousness 373 (pp. 364–383). Cambridge; New York: Cambridge University Press. 374 Csikszentmihalyi, M. (1990). Flow: The psychology of optimal experience. New York: Harper & Row. 375 376 Csikszentmihalyi, M. (2000). Beyond boredom and anxiety: Experiencing flow in work and play. San Francisco: Jossey-Bass. First published in 1975. 377 378 Csikszentmihalyi, M., & Csikszentmihalyi, I.S. (1992). Optimal Experience: Psychological Studies of Flow in Consciousness. Cambridge: Cambridge University Press. 379 380 Csikszentmihalyi, M., & Rathunde, K. (1993). The measurement of flow in everyday life. 381 Nebraska Symposium on Motivation, 40, 57-97. 382 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Manuscript to be reviewed Dahlquist, L. M., McKenna, K. D., Jones, K. K., Dillinger, L., Weiss, K. E., & Ackerman, C. S. (2007). Active and passive distraction using a head-mounted display helmet: Effects on cold 383 384 Dahlquist, L. M., McKenna, K. D., Jones, K. K., Dillinger, L., Weiss, K. E., & Ackerman, C. S. (2007) Active and passive distraction using a head mounted display helmet: Effects on cold 383 384 Dahlquist, L. M., McKenna, K. D., Jones, K. K., Dillinger, L., Weiss, K. E., & Ackerman, C. S. 383 (2007). Active and passive distraction using a head-mounted display helmet: Effects on cold pressor pain in children. Health Psychology, 26(6), 794-801. 384 385 Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a 386 multidimensional approach. Journal of personality and social psychology, 44(1), 113 -126. Fredborg, B., Clark, J., & Smith, S. D. (2017). An examination of personality traits associated with Autonomous Sensory Meridian Response (ASMR). Frontiers in Psychology, 8, 247. Fredborg, B.K., Clark, J.M., & Smith, S.D. (2018). Mindfulness and autonomous sensory meridian response (ASMR).’ PeerJ, 6:e5414. 387 388 389 390 391 multidimensional approach. Journal of personality and social psychology, 44(1), 113 -126. 387 multidimensional approach. Journal of personality and social psychology, 44(1), 113 -126. Fredborg, B., Clark, J., & Smith, S. D. (2017). An examination of personality traits associated with Autonomous Sensory Meridian Response (ASMR). Frontiers in Psychology, 8, 247. Fredborg, B.K., Clark, J.M., & Smith, S.D. (2018). Mindfulness and autonomous sensory 387 388 389 390 Fredborg, B., Clark, J., & Smith, S. D. (2017). An examination of personality traits associated 388 with Autonomous Sensory Meridian Response (ASMR). Frontiers in Psychology, 8, 247. Fredborg, B.K., Clark, J.M., & Smith, S.D. (2018). Mindfulness and autonomous sensory meridian response (ASMR) ’ PeerJ 6:e5414 389 390 391 with Autonomous Sensory Meridian Response (ASMR). Frontiers in Psychology, 8, 247. 389 Fredborg, B.K., Clark, J.M., & Smith, S.D. (2018). Mindfulness and autonomous sensory meridian response (ASMR).’ PeerJ, 6:e5414. 390 391 Fredborg, B.K., Clark, J.M., & Smith, S.D. (2018). Mindfulness and autonomous sensory meridian response (ASMR).’ PeerJ, 6:e5414. 390 391 meridian response (ASMR).’ PeerJ, 6:e5414. 391 Glisky, M.L., Tataryn, D.J., Tobias, B.A., Kihlstrom, J.F., & McConkey, K.M. (1991). 392 Glisky, M.L., Tataryn, D.J., Tobias, B.A., Kihlstrom, J.F., & McConkey, K.M. (1991). 392 Absorption, openness to experience, and hypnotizability. Journal of Personality and Social Psychology, 60(2), 263-272. 393 394 Han, S. (1988). Manuscript to be reviewed Jackson, S.A., Thomas, P.R., Marsh, H.W., & Smethurst, C.J. (2001). Relationship between flow, self-concept, psychological skills and performance. Journal of Applied Sport 406 407 Jackson, S.A., Thomas, P.R., Marsh, H.W., & Smethurst, C.J. (2001). Relationship between 406 Jackson, S.A., Thomas, P.R., Marsh, H.W., & Smethurst, C.J. (2001). Relationship between flow, self-concept, psychological skills and performance. Journal of Applied Sport Psychology, 13(2), 129-153. 406 407 408 flow, self-concept, psychological skills and performance. Journal of Applied Sport Psychology 13(2) 129 153 407 408 flow, self-concept, psychological skills and performance. Journal of Applied Sport Psychology, 13(2), 129-153. 407 408 Jameson, E., Trevena, J., & Swain, N. (2011). Electronic gaming as pain distraction. Pain Research and Management, 16(1), 27-32. 409 410 Jamieson, G.A. (2005). The modified tellegen absorption scale: a clearer window on the structure and meaning of absorption. Australian Journal of Clinical and Experimental Hypnosis, 33(2), 119-139. 411 412 413 structure and meaning of absorption. Australian Journal of Clinical and Experimental Hypnosis, 33(2), 119-139. 412 413 Janik McErlean, A.B. and Banissy, M.J. (2017) ‘Assessing individual variation in personality and empathy traits in self-reported autonomous sensory meridian response.’ Multisensory Research, 30(6), 601-613. 414 415 416 John, O. P., Naumann, L. P., & Soto, C. J. (2008). Paradigm shift to the integrative big-five trait taxonomy: history, measurement, and conceptual issues. In: O. P. John, R. W. Robins, & L. A. Pervin (Eds.), Handbook of personality: theory and research (pp. 114-58). New York: Guilford Press. 417 418 419 420 Lau, M. A., Bishop, S. R., Segal, Z. V., Buis, T., Anderson, N. D., Carlson, L., ... & Devins, G. (2006). The Toronto mindfulness scale: Development and validation. Journal of clinical 421 422 Lau, M. A., Bishop, S. R., Segal, Z. V., Buis, T., Anderson, N. D., Carlson, L., ... & Devins, G. 421 Lau, M. A., Bishop, S. R., Segal, Z. V., Buis, T., Anderson, N. D., Carlson, L., ... & Devins, G. (2006). The Toronto mindfulness scale: Development and validation. Journal of clinical psychology, 62, 1445-1467. doi: 10.1002/jclp.20326 421 422 423 (2006). The Toronto mindfulness scale: Development and validation. Journal of clinical psychology, 62, 1445-1467. doi: 10.1002/jclp.20326 422 423 (2006). The Toronto mindfulness scale: Development and validation. Journal of clinical 422 psychology, 62, 1445-1467. doi: 10.1002/jclp.20326 423 Massimini, F., Csikszentmihalyi, M., & Delle Fave, A. (1988). Flow and biocultural evolution. 424 In M. Csikszentmihalyi & I. S. Manuscript to be reviewed The relationship between life satisfaction and flow in elderly K 95 Han, S. (1988). The relationship between life satisfaction and flow in elderly Korean immigrants. In M. Csikszentmihalyi & I. S. Csikszentmihalyi (Eds.), Optimal experience: Psychological studies of flow in consciousness. Cambridge; New York: Cambridge University Press. 395 396 397 398 immigrants. In M. Csikszentmihalyi & I. S. Csikszentmihalyi (Eds.), Optimal experience: Psychological studies of flow in consciousness. Cambridge; New York: Cambridge University Press. 396 397 398 Jackson, S. A., & Eklund, R. C. (2002). Assessing flow in physical activity: The Flow State 399 Jackson, S. A., & Eklund, R. C. (2002). Assessing flow in physical activity: The Flow State Scale-2 and Dispositional Flow Scale-2. Journal of Sport and Exercise Psychology, (24), 133–150. 399 400 401 Scale-2 and Dispositional Flow Scale-2. Journal of Sport and Exercise Psychology, (24), 133–150. 400 401 Jackson, S. A., & Eklund, R. C. (2004). The Flow Scales manual. Morgantown, WV: Fitness 402 Jackson, S. A., & Eklund, R. C. (2004). The Flow Scales manual. Morgantown, WV: Fitness Information Technology, Inc. 402 403 Information Technology, Inc. 403 Jackson, S.A., Marsh, H.W., (1996). Development and validation of scale to measure optimal experience: flow state scale. Journal of Sport and Exercise Psychology 18 (1), 17-35. 404 405 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Moneta, G. B. (2012). On the measurement and conceptualization of flow. In S. Engeser (Ed.), Advances in flow research (pp. 23–50). New York: Springer. 430 431 Moneta, G. B. (2012). On the measurement and conceptualization of flow. In S. Engeser (Ed.), 430 Moneta, G. B. (2012). On the measurement and conceptualization of flow. In S. Engeser (Ed.), Advances in flow research (pp. 23–50). New York: Springer. 430 431 Advances in flow research (pp. 23–50). New York: Springer. 431 Advances in flow research (pp. 23–50). New York: Springer. 431 Nakamura, J., & Csikszentmihalyi, M. (2009). Flow theory and research. In S. Lopez & C. R. Snyder (Eds.), Oxford handbook of positive psychology (2nd ed., pp. 89–105). New York, NY: Oxford University Press. 432 433 434 Nakamura, J., & Csikszentmihalyi, M. (2014). The concept of flow. In Flow and the 435 foundations of positive psychology (pp. 239-263). Springer, Dordrecht. 436 Poerio, G. L., Blakey, E., Hostler, T. J., & Veltri, T. (2018). More than a feeling: Autonomous 437 Poerio, G. L., Blakey, E., Hostler, T. J., & Veltri, T. (2018). More than a feeling: Autonomous sensory meridian response (ASMR) is characterized by reliable changes in affect and physiology. PloS one, 13(6), 1-18. 437 438 439 sensory meridian response (ASMR) is characterized by reliable changes in affect and physiology. PloS one, 13(6), 1-18. 438 439 Roberts, N., Beath, A., & Boag, S. (2019). Autonomous sensory meridian response: scale 440 Roberts, N., Beath, A., & Boag, S. (2019). Autonomous sensory meridian response: scale development and personality correlates. Psychology of Consciousness: Theory Research, and Practice, 6(1), 22-39. https://doi.org/10.1037/cns0000168 440 441 442 development and personality correlates. Psychology of Consciousness: Theory Research, and Practice, 6(1), 22-39. https://doi.org/10.1037/cns0000168 441 442 Ross, S.R., & Keiser, H.N. (2014). Autotelic personality through a five-factor lens: individual differences in flow propensity. Personality and Individual Differences, 59, 3-8. 443 444 Schwartz, S.J., & Waterman, A.S. (2006). Changing interests: a longitudinal study of intrinsic 445 Schwartz, S.J., & Waterman, A.S. (2006). Changing interests: a longitudinal study of intrinsic motivation for personality salient activities. Journal of Research in Personality, 40(6), 1119- 1136. 445 446 447 motivation for personality salient activities. Journal of Research in Personality, 40(6), 1119- 1136. 446 447 Sheldon, K. M., Prentice, M., & Halusic, M. (2015). The experiential incompatibility of 448 mindfulness and flow absorption. Social Psychological and Personality Science, 6(3), 276- 283. 449 450 Tellegen A. (1982). Manuscript to be reviewed Csikszentmihalyi (Eds.), Optimal experience: Psychological studies of flow in consciousness (pp. 288–306). 425 426 Mauri, M., Cipresso, P., Balgera, A., Villamira, M., & Riva, G. (2011). Why is Facebook so successful? Psychophysiological measures describe a core flow state while using Facebook. Cyberpsychology Behaviour and Social Networking, 14(12), 723-731. 427 428 429 successful? Psychophysiological measures describe a core flow state while using Facebook. Cyberpsychology Behaviour and Social Networking, 14(12), 723-731. 428 429 Cyberpsychology Behaviour and Social Networking, 14(12), 723-731. 429 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Manuscript to be reviewed Brief manual for the Differential Personality Questionnaire. Unpublished manuscript, University of Minnesota. Tellegen, A. (1987, October). Discussion: Hypnosis and absorption. Paper presented at the 38th annual meeting of the Society for Clinical and Experimental Hypnosis, Los Angeles. 451 452 453 454 Manuscript to be review Manuscript to be reviewed Tellegen, A., & Atkinson, G. (1974). Openness to absorbing and self-altering experiences ("absorption"), a trait related to hypnotic susceptibility. Journal of Abnormal Psychology, 83, 268-277. Tellegen, A. (1981). Practicing the two disciplines for relaxation and enlightenment: Comment on “Role of the feedback signal in electromyographic biofeedback: The relevance of attention” by Quails and Sheehan. Journal of Experimental Psychology: General, 110, 217– 226. 455 456 457 458 459 460 461 Thienot, E., Jackson, B., Dimmock, J., Grove, J. R., Bernier, M., & Fournier, J. F. (2014). 462 Development and preliminary validation of the mindfulness inventory for sport. Psychology of Sport and Exercise, 15, 72–80. 463 464 Weilbel, D., Wissmath, B., & Mast, F.W. (2010). Immersion in mediated environments: the role of personality traits. Cyberpsychology, Behaviour and Social Networking, 13(3), 251-256. 465 466 468 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Manuscript to be reviewed Table 1(on next page) Means and Standard Deviations for ASMR-Experiencers and Controls on FES, MAAS an TAS Note: FES = Flow Experiences Scale, MAAS = Mindful Attention and Awareness Scale, TAS Tellegen Absorption Scale. N (MAAS, TAS): ASMR=124, Control = 124; N (FES): ASMR = 63 Control = 43. Table 1(on next page) Table 1(on next page) Means and Standard Deviations for ASMR-Experiencers and Controls on FES, MAAS and TAS Means and Standard Deviations for ASMR-Experiencers and Controls on FES, MAAS and TAS Note: FES = Flow Experiences Scale, MAAS = Mindful Attention and Awareness Scale, TAS = Tellegen Absorption Scale. N (MAAS, TAS): ASMR=124, Control = 124; N (FES): ASMR = 63, Control = 43. Means and Standard Deviations for ASMR-Experiencers and Controls on FES, MAAS and TAS Note: FES = Flow Experiences Scale, MAAS = Mindful Attention and Awareness Scale, TAS = Note: FES = Flow Experiences Scale, MAAS = Mindful Attention and Awareness Scale, TAS = Tellegen Absorption Scale. N (MAAS, TAS): ASMR=124, Control = 124; N (FES): ASMR = 63, Control = 43. PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Manuscript to be reviewed Table 2(on next page) Table 2(on next page) Manuscript to be reviewed 1 Table 1 2 Means and Standard Deviations for ASMR-Experiencers and Controls on FES, MAAS and TAS 3 ASMR Control Variable Mean SD Mean SD FES 36.49 8.16 38.09 8.19 MAAS 3.20 0.80 3.28 0.79 TAS 108.74 23.80 94.22 21.95 4 Note: FES = Flow Experiences Scale, MAAS = Mindful Attention and Awareness Scale, TAS = 5 Tellegen Absorption Scale. N (MAAS, TAS): ASMR=124, Control = 124; N (FES): ASMR = 6 63, Control = 43. 7 8 9 10 11 1 Table 1 2 Means and Standard Deviations for ASMR-Experiencers and Controls on FES, MAAS and TAS 3 ASMR Control Variable Mean SD Mean SD FES 36.49 8.16 38.09 8.19 MAAS 3.20 0.80 3.28 0.79 TAS 108.74 23.80 94.22 21.95 4 Note: FES = Flow Experiences Scale, MAAS = Mindful Attention and Awareness Scale, TAS = 5 Tellegen Absorption Scale. N (MAAS, TAS): ASMR=124, Control = 124; N (FES): ASMR = 6 63, Control = 43. 7 andard Deviations for ASMR-Experiencers and Controls on FES, MAAS and TAS p , , 5 Tellegen Absorption Scale. N (MAAS, TAS): ASMR=124, Control = 124; N (FES): ASMR = 6 63, Control = 43. PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Table 2(on next page) Correlations between ASMR intensity, number of videos watched in a single session, frequency of watching ASMR videos, pleasantness of ASMR and the scores on TAS, MAAS and FES. Note: N (Intensity, Frequency, Pleasantness, TAS ad MAAS) = 149. N (Number) = 143. N (FES) = 72. TAS = Tellegen Absorption Scale, MAAS = Mindful Attention and Awareness Scale, FES = Flow Experiences Scale. p < .05, p < .01, p < .001. PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Manuscript to be reviewed Manuscript to be reviewed Figure 1 Figure 1 Manuscript to be reviewed 1 Table 2 2 Correlations between ASMR intensity, number of videos watched in a single session, frequency 3 of watching ASMR videos, pleasantness of ASMR and the scores on TAS, MAAS and FES. Variable Intensity Number Frequency Pleasantness TAS MAAS FES Intensity - -.005 .354* .389*** .180* .033 .207 Number -.005 - .111 -.030 -.061 -.040 -.232 Frequency .354* .111 - .233 -.009 -.072 .017 Pleasantness .389* -.030 .233 - .156 -.042 .168 TAS .180* -.061 -.009 .156 - -.027 .156 MAAS .033 -.040 -.072 -.042 -.027 - .173 FES .207 -.232 .017 .168 .156 .173 - 4 Note: N (Intensity, Frequency, Pleasantness, TAS ad MAAS) = 149. N (Number) = 143. N 5 (FES) = 72. TAS = Tellegen Absorption Scale, MAAS = Mindful Attention and Awareness 6 Scale, FES = Flow Experiences Scale. p < .05, p < .01, *p < .001. 7 etween ASMR intensity, number of videos watched in a single session, frequency SMR videos, pleasantness of ASMR and the scores on TAS, MAAS and FES. ( y, q y, , ) ( ) 5 (FES) = 72. TAS = Tellegen Absorption Scale, MAAS = Mindful Attention and Awareness 6 Scale, FES = Flow Experiences Scale. p < .05, p < .01, *p < .001. 7 8 PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020) Figure 1 A) Relationship between ASMR Intensity and absorption (TAS), mindfulness (MAAS), and ﬂow (FES) based on z scores, B) Relationship between ASMR Intensity and TAS based on original scores, C) Relationship between ASMR Intensity and MAAS based on original scores, D) Relationship between ASMR Intensity and FES based on original scores. PeerJ reviewing PDF \| (2019:09:41043:2:0:NEW 2 Jan 2020)
https://openalex.org/W4289884157	https://link.springer.com/content/pdf/10.1007/s10643-022-01376-5.pdf	English	null	Parents’ Perspectives of Family Engagement with Early Childhood Education and Care During the COVID-19 Pandemic	Early childhood education journal	2,022	cc-by	9,791	Introduction short- and long-term impacts on children and their families (Goldfeld et al., 2022; Schmeer et al., 2021). A critical com ponent to promoting children’s resilience during adverse or traumatic events is for children to experience sensitive, responsive caregiving from the important adults in their lives (Bartlett & Vivrette, 2020). Quality early childhood education and care (ECEC) can provide a buffer against stressful home environments, particularly for those children experiencing less sensitive and responsive adult interactions in the home (Berry et al., 2016; Davies et al., 2021). Thus, during a pandemic it is feasible that ECEC settings may fur ther protect children against the potential negative impacts of home isolation and unstable home environments. Although the health risk of exposure to the coronavirus disease is relatively low for young children (Bhopal et al., 2021; Sinha et al., 2020), responses to the pandemic may result in long term negative effects on children’s wellbeing and development. Mitigation measures of confinement and social distancing have been shown to reduce transmission of the disease, but these pandemic responses have often created stressful family environments, with potentially detrimental Abstract The COVID-19 pandemic has created significant challenges for Early Childhood Education and Care (ECEC) services and families, impacting family access to services and their communication and engagement with educators. This study aimed to examine parents’ perspectives of family engagement with ECEC services during the pandemic. Primary caregivers in Victoria at the time of recruitment (September–November 2020) were invited to participate. Of the 66 participants who completed an online survey, 25 also took part in semi-structured video call or phone interviews; qualitative findings from these interviews are reported in this paper. Four key themes were conceptualised using a reflexive thematic approach: (1) disruptions to ECEC access and attendance impacting on family routines and relationships, and child development; (2) barriers to family engagement; (3) ECEC educators’ support of families and children during the pandemic; and (4) increased parental appreciation of the ECEC profession. Findings revealed that disruptions to ECEC access and routines during the pandemic adversely impacted family engagement, and child learning and social-emotional wellbeing for some families. These were aggravated by other stressors, including increased parental responsibilities in the home, financial and health concerns, and changed work conditions. Findings also demonstrated successful methods used by educators to maintain communication and connections with families. Importantly, parents expressed increasing appreciation of the profession and an increased awareness of the value of family involvement in children’s learning. Learnings regarding strategies for effective and alternative ways of engaging families are discussed. Keywords Early childhood education and care · COVID-19 · Family engagement · Educator-parent partnerships Parents’ Perspectives of Family Engagement with Early Childhood Education and Care During the COVID-19 Pandemic Accepted: 5 July 2022 / Published online: 5 August 2022 © The Author(s) 2022 2 Genetics, Murdoch Children’s Research Institute, Melbourne, Australia 1 Melbourne Graduate School of Education, The University of Melbourne, 3010 Ground Floor Kwong Lee Dow Building, 234 Queensberry St, Melbourne, VIC, Australia Early Childhood Education Journal (2023) 51:1279–1289 https://doi.org/10.1007/s10643-022-01376-5 Early Childhood Education Journal (2023) 51:1279–1289 https://doi.org/10.1007/s10643-022-01376-5 Theoretical Frameworks and Defining Family Engagement with ECEC The current study is informed by Bronfenbrenner’s ecologi cal model, whereby child development is viewed as a com plex system of relationships affected by multiple levels of the child’s surrounding environment, from the immediate context of family and ECEC/school to broad cultural values, laws, and customs (Bronfenbrenner, 1977); Guy-Evans, 2020). Within this system, it is the microsystem, which includes the immediate environmental surrounds of the child such as family and early childhood education, that is thought to be the most influential (Bronfenbrenner & Ceci, 1994; Guy-Evans, 2020). In this way, children’s develop ment is influenced by the interactions between the child and their primary caregivers (family and educators), but also by the connections and partnerships between these caregivers. This study is also informed by Kim and Sheridan’s fam ily engagement model (Kim & Sheridan, 2015). Two cen tral dimensions of family engagement proposed by Kim & Sheridan (2015) include: (1) the structural dimension (ECEC and family efforts working side by side to support children’s learning and development); and (2) the relational dimension (a focus on continuing communication, con nections, and interactions between families and educators) (Jeon et al., 2020). The structural dimension can include parent involvement in activities in ECEC settings, which may provide the opportunity to observe practices used by educators to promote children’s learning and development which parents can then implement in the home (Barnett et al., 2020). Communication is an important aspect of the relational dimension of family engagement and includes educators’ practices and strategies used to engage families (Eadie et al., 2017; Christenson, 2003). Building educator-family partnerships is especially important during stressful global events when parent-child relationships may be compromised. However, risk mitiga tion measures during the COVID-19 pandemic meant that ECEC services faced numerous communication challenges, such as reduced opportunities for face-to-face contact with children and families. While the pandemic resulted in the increased use of digital forms of communication – which research suggests can be beneficial in terms of making the work of educators and children’s learning more visible to families (Oke et al., 2021) - there are limitations to the use of online learning for very young learners (Hu et al., 2021), as well as the issue of the digital divide, whereby certain groups are at risk of being excluded from this form of interaction (Barnett et al., 2021; Shaik, 2022). The Importance of Parent Engagement with ECEC and Family-Educator Partnerships A recent meta-analysis demonstrated that educators’ use of practices of engagement (e.g., sending home information about the child) and par ent engagement in the ECEC centre (such as volunteering) were indirectly associated with kindergarten academic read iness through increases in the quality of parent engagement in home learning activities (Barnett et al., 2020). Strong parent-educator relationships can be particularly important for disadvantaged families; for example, a study involving low-income parents and their preschool children demon strated a positive association between parents’ perception of parent-educator communication, and the home learning environment (Lin et al., 2019). In Australia, the importance of strong family-educator partnerships is also reflected in policy frameworks such as the Australian Government’s Early Years Learning Framework (EYLF). The Framework describes the principles, practices and outcomes that sup port and enhance young children’s learning and states that partnerships with families “are based on the foundations of understanding each other’s expectations and attitudes and building on the strength of each other’s knowledge” (EYLF, 2009, p.13). documents, recent evidence suggests that Australian educa tors may require increased support and training to improve their skills and confidence in working with families (Mur phy et al., 2021). Gaining a better understanding of parents’ perspectives of engagement with ECEC and what they view as facilitating factors and barriers to engagement, may inform improvements to family engagement practices, thereby enhancing family–ECEC relationships, and sup porting children’s learning and development across home and ECEC settings (Vuorinen, 2021). Theoretical Frameworks and Defining Family Engagement with ECEC Further, in an Australian qualitative study with educators during the pandemic, educators reported that digital technology was important for keeping services and families connected dur ing the pandemic, but this was contingent on families hav ing access and capacity to be online (Murray et al., 2021). 1 3 The Importance of Parent Engagement with ECEC and Family-Educator Partnerships 1 Melbourne Graduate School of Education, The University of Melbourne, 3010 Ground Floor Kwong Lee Dow Building, 234 Queensberry St, Melbourne, VIC, Australia Children develop and learn within the contexts of the home environment, ECEC services, and community. The 1 3 Early Childhood Education Journal (2023) 51:1279–1289 1280 home environment, which includes the nature of parent- child interactions in and around the home, is an important predictor of child learning and development (Lehrl et al., 2020). Positive parenting behaviours and quality parent- child relationships can strengthen family bonds and support child learning and wellbeing (Perrin et al., 2016). In addi tion, strong partnerships between early childhood educators and families can boost parents’ confidence and capability to support young children’s learning. A recent meta-analysis demonstrated that educators’ use of practices of engagement (e.g., sending home information about the child) and par ent engagement in the ECEC centre (such as volunteering) were indirectly associated with kindergarten academic read iness through increases in the quality of parent engagement in home learning activities (Barnett et al., 2020). Strong parent-educator relationships can be particularly important for disadvantaged families; for example, a study involving low-income parents and their preschool children demon strated a positive association between parents’ perception of parent-educator communication, and the home learning environment (Lin et al., 2019). In Australia, the importance of strong family-educator partnerships is also reflected in policy frameworks such as the Australian Government’s Early Years Learning Framework (EYLF). The Framework describes the principles, practices and outcomes that sup port and enhance young children’s learning and states that partnerships with families “are based on the foundations of understanding each other’s expectations and attitudes and building on the strength of each other’s knowledge” (EYLF, 2009, p.13). home environment, which includes the nature of parent- child interactions in and around the home, is an important predictor of child learning and development (Lehrl et al., 2020). Positive parenting behaviours and quality parent- child relationships can strengthen family bonds and support child learning and wellbeing (Perrin et al., 2016). In addi tion, strong partnerships between early childhood educators and families can boost parents’ confidence and capability to support young children’s learning. The Current Study This qualitative study aimed to explore parents’ experiences of engagement with ECEC services during the COVID- 19 pandemic in Australia. The State of Victoria went into Stage 3 lockdown on July 8, 2020, with Metropolitan While the importance of building strong family-educator partnerships is clearly articulated in research and policy 1 3 1281 Early Childhood Education Journal (2023) 51:1279–1289 Table 1 Interview Topic Guide Topic guide key area Example interview questions ECEC attendance and involvement during the pandemic • How frequently was your child attending ECEC before the pandemic and how has this changed since the first lockdown in March? • What involvement have you had with your ECEC service and educators during the pan demic? How have you felt about this? COVID-19 impact on ECEC-family communication • What types of communication methods, for example telephone or email, has your ECEC service used with you and your family during the pandemic? - In what way, if any are these communication methods different from before the pandemic? - Which communication methods do you find most useful? - What type of information has ECEC been communicating with you during the d i b f h l h d f opic Guide Example interview questions • How frequently was your child attending ECEC before the pandemic and how has this changed since the first lockdown in March? • What involvement have you had with your ECEC service and educators during the pan demic? How have you felt about this? • What types of communication methods, for example telephone or email, has your ECEC service used with you and your family during the pandemic? - In what way, if any are these communication methods different from before the pandemic? - Which communication methods do you find most useful? - What type of information has ECEC been comm nicating ith o d ring the Victoria moving into Stage 4 lockdown on August 2, 2020, which resulted in ECEC services being closed to all fami lies except essential workers and vulnerable children for a period of two months (noting this lockdown lasted almost 16 weeks in total). This provided a unique opportunity to capture families’ experiences at a time when many fami lies had limited access to ECEC and services were required to navigate alternative methods of engagement. COVID-19 impact on ECEC expectations and connectedness Participants - If no, what resources or communication might have been helpful to support your child’s learning during lockdown? Participants who had taken part in an online survey exam ining family engagement with ECEC during the COVID- 19 pandemic were invited to take part in this study. Survey participants were recruited primarily via Facebook and the research team’s ECEC networks (September-November 2020). Primary caregivers of young children (aged 1–6 years) living in Victoria, Australia who were enrolled in ECEC services at the time of recruitment were invited to take part. Upon completion of the online survey, partici pants were asked if they would be willing to take part in a one-to-one interview to further discuss their experiences of engagement with ECEC during the pandemic. Of the 66 participants who completed the survey, 37 parents (56.1%) agreed to be contacted for an interview. For the purpose of this paper, we are reporting on the qualitative data captured through the interviews. All participants were provided with a participant information statement and provided informed consent. Ethical approval for the study was obtained from the University of Melbourne Human Research Ethics Com mittee (#2057564). might have been helpful to support your child’s learning during lockdown? • What, if any, expectations have your ECEC service or educators placed on you in relation to your child’s learning at home? • Have you had any expectations of your ECEC service in terms of supporting your child’s learning and development during the pandemic, and if so what are they? • Do you think your perception of and con nectedness to ECEC is different during the pandemic compared to pre-pandemic? n • Has the pandemic impacted on your child’s interactions or connections with other family members, for example not seeing grandparents or other family members who had spent regular time with your child prior to the pandemic? • Has the pandemic impacted on your relation ship or interactions with your child, and if so, in what way? • What, if any, expectations have your ECEC service or educators placed on you in relation to your child’s learning at home? • Have you had any expectations of your ECEC service in terms of supporting your child’s learning and development during the pandemic, and if so what are they? an option due to being in Stage 4 lockdown. ECEC support for children’s learning and development Participants A topic guide of interview questions was used by the interviewer, which included prompts to encourage more detailed responses from parents. Table 1 highlights the key areas and example questions covered by the topic guide. All interviews were audio recorded with participant consent. COVID-19 impact o family relationships The Current Study The study aimed to examine parents’ varying experiences engaging with ECEC services, and to explore the challenges, suc cesses, and learnings to carry forward.i • How frequently was your child attending ECEC before the pandemic and how has this changed since the first lockdown in March? • What involvement have you had with your ECEC service and educators during the pan demic? How have you felt about this? • What types of communication methods, for example telephone or email, has your ECEC service used with you and your family during the pandemic? - In what way, if any are these communication methods different from before the pandemic?i - In what way, if any are these communication methods different from before the pandemic?i The specific research questions were: - Which communication methods do you find most useful? 1) What was the impact of the pandemic on ways families engage with ECEC services? - What type of information has ECEC been communicating with you during the pandemic, e.g., about fees, health and safety, individual information about your child, story time? 2) What were parents’ perceived facilitating factors and barriers to engagement with ECEC services during the pandemic? 3) How did ECEC services support families and children with learning at home during the pandemic? • Has your ECEC service provided any resources or assistance to help you support your child’s learning at home during the pandemic, for example ideas on activities to do at home? • Has your ECEC service provided any resources or assistance to help you support your child’s learning at home during the pandemic, for example ideas on activities to do at home? Methods - If yes, what resources have they provided and have you used them? How have you used the resources? How useful have you found these resources? COVID-19 impact on ECEC-family communication ECEC attendance and involvement during the pandemic Analyses to ECEC services and personal factors brought on by the pandemic (e.g., changes to employment arrangements, such as working from home).f A reflexive thematic approach was used to analyse inter view transcripts (Braun & Clarke, 2019). The approach was inductive (not hypothesis driven), semantic (focused on what parents are saying rather than trying to determine the assumptions underpinning what they are saying) and critical realist (focused on reporting an assumed reality evident in the data, i.e., reporting experiences, meanings and the real ity of participants).i The multiplicity of contextual factors affecting each fam ily had implications for their capacity to engage with ECEC. Parents who experienced multiple disruptions to family rou tines, which were compounded by a lack of other external supports for childcare (such as family members), responded with an increased awareness of their need for ECEC. These parents were often juggling multiple responsibilities (such as fulltime care, remote learning for children, and working from home) and reported increased stressors brought about by mitigation measures: “we both were working full time the whole time, and you’re trying to fit eight hours of working a day when your child’s awake from 7:00 till 8:00 which means you’re working overnight” (P4). For another parent, prolonged lack of ECEC access had a negative impact on their attitude towards engaging in activities with their child: “But after the eight-week lockdown, I just do not want to do much more role playing. I’m just so over it. I think she sees the strain there, where at the beginning it was really fun and cute and now it’s just really annoying” (P17). The multiplicity of contextual factors affecting each fam ily had implications for their capacity to engage with ECEC. The first author began with data familiarization, reading and re-reading each transcript and noting initial impres sions. Next, initial codes generated by the first author were refined by the second and third authors before identifying initial themes. Themes were conceptualised from clustering similar codes together and organised around a central con cept or idea to capture shared meaning. To ensure analytic rigour, the thematic analysis process encompassed multiple interpretations of the data from people with diverse exper tise with themes progressively refined via discussion and consensus amongst the lead researchers. j y y g ( ) Conversely, some families reported being much less impacted by mitigation measures and restricted access to ECEC. Parent Concern for Children’s Wellbeing and Development Another impact of changes to ECEC access and attendance during the pandemic was concern amongst parents about children’s social-emotional wellbeing, learning, and devel opment. While some parents voiced concern more gener ally about children missing out on learning and preparation for school transition, the most common area of concern was children’s social-emotional development. Parents relayed concerns about the impact of social isolation on their child and the lack of opportunities for interactions with peers, both at home and in ECEC services (with reduced numbers of children in attendance). One parent described how their child “cries because none of his friends are there” (P10). Data Collection Parents who agreed to take part in an interview were con tacted via email and given the option of doing the inter view by phone or video call (Zoom). Face-to-face was not 1 3 Early Childhood Education Journal (2023) 51:1279–1289 1282 Analyses For these parents, factors such as living in a rural area which was not as greatly impacted by ECEC restric tions, as well as having childcare support from external family members contributed to less disruptions to routines and families’ experiences: “Because we couldn’t access childcare, he was spending more time with my mum… he’s gotten to spend more time with her when he normally wouldn’t” (P13). Despite differences in parental experiences and responses to the impact of disruptions during the pan demic, an overall effect observed similarly across all parent participants was increased parental stress. Results Of the 37 survey participants who consented to an interview, 25 completed a one-to-one interview with the first author. All participating parents were mothers. On average, parents who took part were from less disadvantaged socioeconomic areas and most parents had completed an undergraduate or postgraduate degree. Children of participants ranged in age from 13 months to five years and attended a mix of long day care services and stand-alone kindergartens, with one child attending a family day care service. Four key themes were conceptualised from analyses of the interview data: (1) impact of disruptions to ECEC access for children and families; (2) barriers to family engagement during the pandemic; (3) ECEC supports for children and families during the pandemic; and (4) increased parental appreciation of the ECEC profession. 1 3 Theme 2: Barriers to Family Engagement During the Pandemic Parents also expressed concerns about children’s emo tional adjustment when they returned to ECEC services after a period of non-attendance due to the pandemic. One parent reported a regression in their child’s social skills: Disruptions to ECEC Access and Differing Parental Responses The COVID-19 pandemic and associated mitigation mea sures across Victoria led to multiple changes to families’ ECEC access and attendance. While parents reported a range of experiences, a key issue identified was the disrup tion to family routines caused by changes in families’ access Some parents also reported concerns about the impact of reduced contact with wider family members for children 1 1 3 Early Childhood Education Journal (2023) 51:1279–1289 1283 during the pandemic. As described by one parent, the child “just didn’t understand where people had disappeared to” and “essentially, all her relationships with family members became screen-based and missing out on the physical con tact that you would otherwise have with a little kid” (P23). Related to this concern were parents’ comments about how children were not engaging well with electronic means of communication with family members and friends. Children were described as “want(ing) to run around, not participate in the conversation” (P18), or being “sick of it and just, ‘I want to see grandma in real life. I don’t want to talk on the iPad” (P25). during the pandemic. As described by one parent, the child “just didn’t understand where people had disappeared to” and “essentially, all her relationships with family members became screen-based and missing out on the physical con tact that you would otherwise have with a little kid” (P23). Related to this concern were parents’ comments about how children were not engaging well with electronic means of communication with family members and friends. Children were described as “want(ing) to run around, not participate in the conversation” (P18), or being “sick of it and just, ‘I want to see grandma in real life. I don’t want to talk on the iPad” (P25). we were always in this sort of half on, half off zone… I feel like the kids would have felt like we were always preoccu pied, which we probably were (P22). Another parent observed that the additional stressors caused by the pandemic had led to increased parental stress, with negative impacts on the parent-child relationship: “I actually find I’ve probably got a lot less patience with her now. I find myself getting more frustrated with her and a lot more short with her than I think I otherwise would have” (P12). Reduced or Removed Opportunities for Face-to-Face Contact Between Families and ECEC Services If we were at the playground and him and another kid were going to the swing at the same time, he’d get very aggressive, which is not like him at all. . the social anxiety of leaving has sort of started up again for both kids. . I don’t think it’s very great to just be relying on me, I like that they rely on other adults and stuff as well. (P9) A major barrier to families’ engagement with ECEC ser vices during the pandemic was the elimination of oppor tunities for informal face-to-face conversations and parent observations of their child within their ECEC environment during pick-up and drop-off times. These opportunities were valued by parents and described as the “main way we used to engage” (P5). Face-to-face interactions were also key to providing parents with an increased understanding of their child’s experiences at ECEC services as well as their sense of community. However, as lamented by one parent: Other parents reported their children’s disinclination and anxieties about returning to ECEC: At this age, it’s crucial to be socialising physically with kids…I do worry a little bit that he’s going to be super anx ious about going back into large groups again because we’ve been at home for so long. (P1) It went from in the yard pick up, they’re all running around and to have a bit of a chat. . there’s parents hang ing around. . to suddenly having zero information... because you can’t see into the room, you can’t see anything so just have no sense of, it’s like a black hole that you’re sending child in. (P5) Loss of Connections with Familiar Educators We’ve got a WhatsApp parent group too so there’s a lot being exchanged, ideas. . and we set-up a lot of joint phone calls. . they had some Zoom parties, everyone getting on and listening to a story and then that sparked its own set of connections”. (P5) Changes to attendance at ECEC services created barriers for some children and families in accessing or maintaining con nections with the educators with whom children were most familiar. This loss of familiar relationships had a negative impact on family engagement and child wellbeing: Impact on Family Relationships A positive effect that emerged from disruptions to ECEC attendance was strengthened sibling and parent-child rela tionships for some families. In some cases, forced home isolation led to parents appreciating the amount of quality time they had with their child; “it’s probably strengthened our relationship and something I needed…as a mum to be like, okay, they’re only little once, stop going, stop trying to do everything” (P11). Some parents also commented on the bonding observed between siblings from more time spent at home. One parent commented: While some parents recognised that measures to limit face-to-face interactions were a constraint of the pandemic for ECEC services, other parents reported experiences of limited engagement with ECEC educators once attendance had resumed. One parent stated, “it’s become a contactless, transactional kind of point” (P14), while another pointed out: There’s a lot less interaction now. Again, that’s contrib uting to not knowing the carers as well. You know the key ones, who are the room leaders or whatever, but the other staff that you used to chat with particularly at the end of the day, to get to know them, you’re not getting that chance because you can’t hang around. (P24) I think it’s lovely the way my kids have bonded. They have just become little best friends just because of the amount of time they spent together (P10). While some parents felt the restrictions had a positive impact on their parent-child relationship, it is important to acknowledge that for other parents the stress of juggling working from home with caring for and home-schooling children was extremely challenging. This led to some par ents feeling guilty for not investing in quality time with their children due to being occupied with other responsibilities: “Instead of being able to enjoy our time together, I feel like Reduced opportunities for information exchange and parental engagement during face-to-face conversations with educators highlighted the need for services to find alterna tive ways to maintain these conversations with families during the pandemic. However, there was variability in ser vices’ capacity to pivot to online modes of communication, 1 3 Early Childhood Education Journal (2023) 51:1279–1289 1284 Theme 3: ECEC Supports for Children and Families During the Pandemic I think it would’ve just been nice if he could’ve spoken to the two educators that he’s most connected with. . having some kind of connection with them would have been nice, even like a recorded video that they could just send home of them doing something in the room. (P3) Loss of Opportunities for Interactions with Other Families with some services more reliant on face-to-face communi cation. As articulated by one parent: with some services more reliant on face-to-face communi cation. As articulated by one parent: Families reported loss of community and opportunities to interact with other families at their service, creating another barrier to ECEC engagement. As one parent commented: “that’s kind of lost now. You don’t really get to speak to many parents anymore”. (P11) Another parent described “strug gling” in the “new normal. . where you don’t get to be a part of a community as much” (P8), while another reported the need to establish alternative modes of communication to maintain a sense of community: “we are going to have to think about ways to change the methods of communication. I guess, to adapt to the fact that we don’t really get to talk to the parents very much anymore”. (P8) It is a culture of in-person communication, which is great when it’s working but pretty much impossible now. I think one of the challenging things they do with our service, is that they tend to do a lot of face-to-face communication, which is really nice. . But they’re not very digital… we would quite like them to pivot to something like Storypark now because otherwise we have so little assessment of the day. (P8)f Changes to drop-off and pick-up routines often meant parents were not able to enter their child’s ECEC room to assist in settling them during the pandemic. This too, had negative implications for family engagement. As one parent described: A phenomenon that emerged from the loss of opportuni ties for face-to-face interactions was the increase in fami lies’ use of digital modes of communication, with several being organised by families. One parent described con necting with other families at their ECEC service through parent-initiated online modes: It was harder for me. . because I was used to going into the room with him and getting him a little bit settled, and then sort of waiting five, ten minutes and then leaving, but to sort of have to hand him over at the door and he’s obviously in distress. (P18) 1 3 Differing Levels and Varying Types of Support The levels of support for home learning and approaches to communication with families varied widely across ECEC services, creating diverse experiences for families. While some parents reported receiving regular communication and “check-ins” from their ECEC service (e.g., emails, newsletters, and phone calls), online learning sessions (e.g., digital reading sessions), and resources for home learning (e.g., craft and learning activity packs for children to com plete at home), other parents reported receiving little or no contact or resources from their service. The diversity in the levels and types of support parents received during the pan demic may reflect the differing ways ECEC services coped with the challenges of the pandemic, their digital fluency, and the differing infrastructure, funding and supports avail able to services. While some services appeared to be able to pivot relatively quickly to using remote modes of engage ment, for other services, it was evident that the transition to digital modes of communication and learning was a major challenge. For example, one parent commented that staff at her ECEC service lacked confidence in “writing and using computers and everything”. (P23) The negative impact of lack of contact with services was felt most keenly by vulnerable families. In the case of one family of a child with a disability who experienced decreased ECEC attendance, feelings of disconnection and exclusion were reported: It was hurtful to be forgotten about, or to have had no contact from them, and to not receive any resources or help or contact for such a long period of time that we no lon ger felt part of the community and it’s hard not to see, that it’s just yet, another example of subtle discrimination and exclusion that people with disability experience…. . (P23)f Changes to room groupings, staff rotations, and turnover during the pandemic also disrupted children’s connections with familiar educators, negatively impacting on their well being. One parent commented on the child’s response to changes in room groupings and educators: “It’s been harder. There’s days where he says, ‘Oh, I don’t want to go to child care. I just want to stay home with you’”. (P25) 1 3 1 3 Early Childhood Education Journal (2023) 51:1279–1289 1285 The varying levels and types of support provided by ECEC services for home learning also had mixed degrees of participation and usefulness for families. Personalised, Two-Way Communication to Support Home Learning I think in a way the kinder got dragged into the virtual environment out of necessity, but it’s been fantastic now that they’re back to learning face-to-face, they’re still using it to send us pictures of what they’re doing, which is lovely and to communicate… I find it really simple and direct and much more easy for me as well as a parent who was often not around to pick up and drop off…now I feel like I know what’s going on a bit more at kinder. (P25)f Parents also reflected on the effectiveness of various strat egies for remote communication and home learning. Suc cessful ways of maintaining communication with families to promote family engagement and support home learning included: sending regular emails; making phone calls or video calls to see how families and children were doing; posting photos, videos and stories of what was happening at ECEC services via digital platforms such as “Storypark” or “Facebook”; posting videos of educators reading to children or online ‘live’ Zoom reading sessions; organising com munal online events (e.g., virtual sessions for celebrating Father’s day); using apps for notifying parents about home learning activities planned by educators (e.g., “EDUCA”) and posting activities for home learning on online plat forms (e.g. “Explore”); sharing learning and development plans with parents; and sending home resources and activity packs for children. On the other hand, less effective strategies for remote communication and home learning included: generalised emails, newsletters, and phone calls from centres about policy/health and safety updates which lacked a personal touch; adult-focused phone calls which were either not pur poseful or at a busy time for parents; home resource packs that were not engaging, personalised or authentic for chil dren (P1); and videos of educators reading to children in the centre for families to watch at home. For instance, one parent (P20) described how it was difficult to hear or see the teacher as she read to the children who were onsite, as the video recording would be disrupted by the sounds of other children in the room. Videos of educators demonstrating activities that parents could replicate at home with children were also met with limited success. Differing Levels and Varying Types of Support While some families welcomed and engaged with the home learning resources provided by their service, other families were less engaged due to reasons such as inconvenient timing (P13), a general lack of parent time (P15), and an inability to access resources required to do the activities provided by their ECEC service. However, there were other parents who reported benefiting from ideas and strategies provided by ECEC services via digital platforms: very informed, and it was just like virtual kinder basically. (P25) very informed, and it was just like virtual kinder basically. (P25) ( ) Overall, strategies that were successful in engaging families remotely and supporting children’s learning from home were those that that focused on fostering genuine con nections with families and children, rather than relaying information. Communication with families that was regular and timely was well-received by parents, as well as open communication channels, whereby educators remained accessible by phone and digital modes to families, adopt ing two-way modes of communication. One parent (P24) described how the ECEC service offered parents the abil ity to book Zoom meetings with educators, while another parent commented on the benefits of the Dojo app, which appeared to enhance communication and engagement, both during and post lockdowns: The watercolors, that was amazing, because I haven’t actually tried that. She absolutely loved it. So that was actu ally life changing. She does watercolors a lot now because of that. (P17) Personalised, Two-Way Communication to Support Home Learning One parent stated: One parent’s description of positive virtual family engagement highlights the importance of regular contact and personal interactions, as well as establishing clear expectations: We did do activities with our kids but I didn’t have a lot of patience, unless they seemed like they would be really easy and absorbing. . it’s not the same environment as the day care centre. . they just behave differently at daycare, espe cially with the activities. (P8) They conducted weekly Zoom sessions in small groups. They provided videos and tips and suggestions of activities every day. . they had a phone call with the parents at the start of each lockdown period to explain to us what was going to happen, to listen to any concerns…just to set the tone and make sure we knew what the expectations were at each side… We had an app that we were communicating through—ClassDojo—so they would post tasks or videos or suggestions, or just have regular contact daily. And then they would organize small group sessions of around three kids plus the two educators once a week for about 25 min. . I felt there was a really strong connection there and we felt Challenges of Engaging Young Children Through Digital Modes of Learning One of the key challenges to ECEC engagement during the pandemic identified by parents was the difficulty in engag ing young children through digital modes of learning: “It’s that digital thing, like it’s hard to engage with a screen. . she 1 1 3 1286 Early Childhood Education Journal (2023) 51:1279–1289 didn’t really participate as much as I think she does when she’s in person” (P12); and “she doesn’t enjoy the remote, the skyping. . She’ll get into it for ‘Hi. Yay, I get to see your face’ kind of thing, and then they try and read her a story, but she’s just not into it. . She wants the physical touch of them”. (P11) Parents’ comments in relation to children’s lack of interest in remote learning and preference for physi cal, face-to-face activities reflected the importance of the close physical proximity of educators for young learners. what the educators were offering in terms of, I guess, really considered learning experiences”. Some parents reflected not only on the importance of ECEC for their children’s “learning and discovering” (P1) but also spoke about the benefits of ECEC for their own wellbeing: “they went back and it was good for me. It was good for them… my mood and my productivity was so much better.” (P11) Theme 4: Parental Appreciation for the ECEC Profession Parents’ appreciation of the ECEC profession was reflected in their comments acknowledging the value of ECEC in society, and the challenges faced by the sector during the pandemic: “I think it’s been incredibly tough, and I really hope that at a policy and a government level, there could be some recognition of the fact that this service, the sector is the absolute lifeblood of our society.” (P4) Another parent acknowledged that it is “one of the most difficult professions to be in right now.” (P1) Parents also noted that despite the immense challenges and stress faced by educators, educa tors demonstrated dedication to supporting the children and families in their care. Some parents reflected that the pandemic had highlighted the work of educators in not only providing care for children but also playing a key role in supporting children’s learning and development, acknowledging “a deep appreciation for the fact that they’re educators, not babysitters”. (P4) Discussion The current study used a qualitative approach to explore family engagement with ECEC during the COVID-19 pandemic from the perspectives of parents living in Victo ria during one of the state’s strictest lockdown periods. In addressing the first research question relating to the impact of the pandemic on the ways in which families engaged with ECEC services, findings revealed that in many cases, dis ruptions to ECEC access and routines adversely impacted family engagement and parents’ views of their child’s learning and development, in particular, children’s social- emotional wellbeing. These were aggravated by other stressors, including increased parental responsibilities in the home and changed work conditions. The adverse impacts observed by parents on children’s social-emotional wellbe ing were thought to be due to a lack of peer interactions and isolation from ECEC daily interactions, resulting in a dependence on immediate family members. Recent stud ies have reported the impact of the pandemic on children’s social-emotional wellbeing due to a lack of opportunity for social interactions (Stites et al., 2021; Egan et al., 2021). Evidence suggests that ECEC settings can provide children with increased opportunities for social-emotional learning and strengthen kindergarten/school readiness via respon sive, sensitive daily educator-child interactions (Blewitt et al., 2018). Interactions with peers from an early age have Appreciation of Educators’ Efforts to Provide Ongoing Support to Families and Children Parents reflected on their appreciation for the lengths that educators had gone to in supporting families and children during the pandemic, particularly given the pressures on the sector. Parents also expressed gratitude in relation to educa tors’ commitment to ensuring the health and wellbeing of children, as well as their families: “Throughout the whole situation, they’ve always, always come back to saying that their primary concern is about the welfare of not just the children in their care at the present time, but the welfare of families who have been really seriously impacted by the pandemic.” (P6) 1 3 Benefits of ECEC for Families and Children However, this also highlights that for those who were already experiencing socioeconomic hardship and vulnerability, there was an increased risk of mental health problems and strained family relationships (Evans et al., 2020), underscoring the importance of access to ECEC and family support services during the pandemic. y pp g p A major impact of the pandemic on how ECEC ser vices engaged with families was the need for services to find alternative ways of communicating and connecting with families, outside of traditional face-to-face methods of engagement. Indeed, a major barrier to engagement with ECEC (research question 2) for many families was the loss of opportunities for face-to-face conversations with educa tors, especially when digital forms of communication were a challenge to implement or had not been widely used prior to the pandemic. This proved particularly challenging when a major strength of an ECEC service was delivering face- to-face communication, but this form of engagement was suddenly no longer feasible. Another barrier to engagement was the use of remote, online activities, which proved chal lenging with very young children. Face-to-face contact and interaction with educators and peers for very young children has been identified as critical to their learning and devel opment (Barnett et al., 2021). Previous research exploring parental beliefs about online learning for young children reported parents believed that online learning is less effec tive than traditional early childhood education environments and lacks the social interactions required to engage young children (Dong et al., 2020). Educators have also reported that online learning lacks the interaction and play-based learning opportunities for young children, but creating online learning communities could enhance parent engage ment with ECEC, enabling educators to support children’s learning and development in the home (Hu et al., 2021).i It is evident that successful family engagement approaches are built on strong relationships and tailored to meet the individual needs of children and families. However, it should also be acknowledged that this approach encom passes a number of challenges, not least the increased bur den on educators, who are already experiencing high levels of burnout and poor wellbeing (Berger et al., 2022; Eadie et al., 2021). For educators to provide quality and consis tent support for children and families, their own wellbeing needs to be supported. Benefits of ECEC for Families and Children In some cases, additional efforts by educators to connect with families during lockdowns increased the visibility of children’s learning and parents’ engagement in their child’s learning at home. One parent described the value of observ ing the educator’s interactions with her child during a Zoom call: “I was sitting in on his Zoom calls and I knew they were learning about opposites. When we were reading a book later, I could say, “Oh, there’s the ‘up’ shape. What is the opposite of the ‘up’ shape? Oh, it’s the down shape.“ You could filter that learning into your playtime. I wouldn’t have known that he’d just learnt that concept if I hadn’t been around and had that communication. (P25) Acknowledging the important contribution educators make to supporting children’s learning and development, another parent noted, “We didn’t think that we could match 1 3 Early Childhood Education Journal (2023) 51:1279–1289 1287 children (e.g., maintaining established relationships with familiar educators) can promote family engagement. Adopt ing methods for two-way conversations with families, rather than one-way delivery of more general information was also more effective. This is in line with Kim and Sheridan’s (2015) family engagement model, where it is evident that for families and educators to work together to promote chil dren’s learning and development, the relational dimension of engagement needs to be established, that is, continued communication, connections and interactions between fam ilies and educators. By establishing trusting relationships between parents and educators, continuity across the home and ECEC regarding children’s learning and development is more likely to occur. Findings also align with Bronfen brenner’s ecological model (1977), whereby the connec tions and partnerships that occur within the microsystem, between family and early childhood educators influence children’s learning and development. also been identified as important for social-emotional devel opment, with research demonstrating that social competence and positive relations with peers from an early age may be protective against later psychosocial problems (Egan et al., 2021; Hay, 2005). Despite the negative impacts of the pandemic mitigation measures on social interactions, the increase in quality time spent at home allowed for some families to strengthen their family bonds, in terms of both sibling and parent-child rela tionships. It may be that protective factors, such as employ ment and financial stability, and good health, contributed to some families having positive experiences during the stress ful periods of lockdowns. Benefits of ECEC for Families and Children Therefore, effective family engage ment approaches must account for educators’ wellbeing and needs as well as the needs of the children and families in their care. References Barnett, M. A., Paschall, K. W., Mastergeorge, A. M., Cutshaw, C. A., & Warren, S. M. (2020). Influences of parent engagement in early childhood education centers and the home on kindergarten school readiness. Early Childhood Research Quarterly, 53, 260–273. Barnett, W. S., Grafwallner, R., & Weisenfeld, G. G. (2021). Corona pandemic in the United States shapes new normal for young chil dren and their families. European Early Childhood Education Research Journal, 29(1), 109–124. Bartlett, J. D., & Vivrette, R. (2020). Ways to promote children’s resilience to the COVID-19 pandemic. Child Trends. Available online: https://www.childtrends.org/publications/ways-to-pro mote-childrens-resilienceto-thecovid-19-pandemic (accessed on 30 May 2020). Berger, E., Quinones, G., Barnes, M., & Reupert, A. (2022). Early childhood educators’ psychological distress and wellbeing during the COVID-19 pandemic. Early Childhood Research Quarterly, 60, 298–306. Limitations and Future Directions The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. digital platforms in ECEC settings. While educators may feel knowledgeable regarding how to work with families, a major challenge lies in converting their knowledge into practice to improve partnerships with families (Murphy et al., 2021). Professional learning for educators that focuses on building effective, collaborative family partnerships and supporting the home learning environment should also be a priority. The pandemic has shone a spotlight on the critical role ECEC plays, not only in providing care for children, but importantly, the vital role educators play in supporting chil dren’s learning and development and parental workforce participation. Findings from this study show it does appear parents’ perspectives of the role of ECEC is shifting, with a greater understanding not only of the ways in which educa tors promote children’s learning and development but also the contribution that parents can make to children’s learn ing. This has implications for greater family involvement in children’s learning and experiences within ECEC programs. Government investment and public messaging is needed to further highlight the visibility of the vital work of educators and to enhance the value of the sector in society (Eadie et al., 2021). Conclusions This study provides important insights into barriers and facilitators to family engagement with ECEC in the con text of COVID-19, and learnings for strengthening family- ECEC partnerships beyond the pandemic. The pandemic has highlighted the importance of ECEC for families, as well as the need for ECEC services to work with families to find personalised methods of engagement that meet their varying contexts and needs.. However, to do this success fully, the sector needs to be supported and resourced to con fidently use digital methods modes of communication, and to engage with families collaboratively to establish com munication practices that foster personalized, ongoing, and authentic engagement. Berry, D., Blair, C., Willoughby, M., Garrett-Peters, P., Vernon-Fea gans, L., Mills-Koonce, W. R., & Family Life Project Key Inves tigators. (2016). Household chaos and children’s cognitive and socio-emotional development in early childhood: Does childcare play a buffering role? Early childhood research quarterly, 34, 115–127. Bhopal, S. S., Bagaria, J., Olabi, B., & Bhopal, R. (2021). Children and young people remain at low risk of COVID-19 mortality. The Lancet Child & Adolescent Health, 5(5), e12–e13. https://doi. org/10.1016/S2352-4642(21)00066-3 Blewitt, C., Morris, H., Nolan, A., Jackson, K., Barrett, H., & Skouteris, H. (2018). Strengthening the quality of educator-child interac tions in early childhood education and care settings: A concep tual model to improve mental health outcomes for preschoolers. Early Child Development and Care. https://doi.org/10.1080/0300 4430.2018.1507028ll Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic anal ysis. Qualitative research in sport exercise and health, 11(4), 589–597. Acknowledgements We would like to thank all of the parents who contributed their time to take part in this project. This project was supported by the Research in Effective Education in Early Childhood (REEaCh) Research Hub, which was established in 2019 through the generosity of the Leaper Foundation. Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American psychologist, 32(7), 513. p p y g ( ) Bronfenbrenner, U., & Ceci, S. J. (1994). Nature-nurture reconcep tualised: A bio-ecological model. Psychological Review, 10(4), 568–586. Funding Open Access funding enabled and organized by CAUL and its Member Institutions Davies, C., Hendry, A., Gibson, S. P., Gliga, T., McGillion, M., & Gonzalez-Gomez, N. (2021). Early childhood education and care (ECEC) during COVID‐19 boosts growth in language and execu tive function.Infant and child development, 30(4), e2241. Limitations and Future Directions This study has a number of limitations. First, participants were recruited via online advertising, so only those who have access to the internet could take part and participants were self-selecting. In addition, parents were all mothers, from English speaking backgrounds, from less disadvan taged socioeconomic areas and had completed high levels of education. Experiences of ECEC engagement during the pandemic may have been quite different for families expe riencing socioeconomic disadvantage or from different lan guage or cultural backgrounds. While technology use was highlighted as an effective strategy for communicating with families and supporting the home learning environment, this approach may create further barriers for families with out digital access or capacity (Barnett et al., 2021; Murray et al., 2021). Further research exploring effective approaches for educators to build relationships with families who lack either the access or capacity to engage in digital forms of engagement is much needed, given the rise in the use of Importantly, findings also demonstrated successful meth ods used by educators to maintain communication and engagement with families, and support children’s learning and development at home (research question 3). Findings from this study showed that communication that is strongly focused on making genuine connections with families and 1 3 1288 Early Childhood Education Journal (2023) 51:1279–1289 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. Declaration Dong, C., Cao, S., & Li, H. (2020). Young children’s online learning during COVID-19 pandemic: Chinese parents’ beliefs and atti tudes. Children and youth services review, 118, 105440. Conflict of interest The authors declare that there is no conflict of in terest. 1 3 Early Childhood Education Journal (2023) 51:1279–1289 1289 development. School Effectiveness and School Improvement, 31(1), 1–6. Eadie, P., Tayler, C., & Stark, H. (2017). Every toddler talking final report. Melbourne, Australia: University of Melbourne. https://www.education.vic.gov.au/Documents/about/research/ Every%20Toddler%20Talking%20Final%20Report.pdf Lin, J., Litkowski, E., Schmerold, K., Elicker, J., Schmitt, S. A., & Purpura, D. J. (2019). Parent–Educator communication linked to more frequent home learning activities for preschoolers. Child & Youth Care Forum, 48(5), 757–772. Eadie, P., Levickis, P., Murray, L., Page, J., Elek, C., & Church, A. (2021). Early childhood educators’ wellbeing during the COVID- 19 pandemic. Early Childhood Education Journal, 49(5), 903–913 Murphy, C., Matthews, J., Clayton, O., & Cann, W. (2021). Partnership with families in early childhood education: Exploratory study. Australasian Journal of Early Childhood, 46(1), 93–106. Egan, S. M., Pope, J., Moloney, M., Hoyne, C., & Beatty, C. (2021). Missing early education and care during the pandemic: The socio- emotional impact of the COVID-19 crisis on young children. Early Childhood Education Journal, 49(5), 925–934. Murray, L., Eadie, P., Church, A., Levickis, P., Page, J., & Forthcom ing (2021). Wellbeing during the COVID-19 Pandemic: Perspec tives of Australian Early Childhood Educators. In J. Pattnaik & M.R. Jalongo (Ed). The Impact of COVID-19 on Early Childhood Education and Care: International Perspectives, Challenges, and Responses, Springer Nature. Accepted Oct Evans, S., Mikocka-Walus, A., Klas, A., Olive, L., Sciberras, E., Karantzas, G., & Westrupp, E. M. (2020). From “It has stopped our lives” to “Spending more time together has strengthened bonds”: The varied experiences of Australian families during COVID-19. Frontiers in psychology, 2906. Oke, A., Butler, J. E., & O’Neill, C. (2021). Identifying Barriers and Solutions to Increase Parent-Practitioner Communication in Early Childhood Care and Educational Services: The Develop ment of an Online Communication Application. Early Childhood Education Journal, 49(2), 283–293. Goldfeld, S., O’Connor, E., Sung, V., Roberts, G., Wake, M., West, S., & Hiscock, H. (2022). Potential indirect impacts of the COVID- 19 pandemic on children: a narrative review using a commu nity child health lens. Medical Journal Of Australia. https://doi. org/10.5694/mja2.51368 Schmeer, K. K., Singletary, B., Purtell, K. M., & Justice, L. M. (2021). Declaration Family Disruption and Parenting During the COVID-19 Pan demic. Journal of Family Issues, 0192513X211042852 Guy-Evans, O. (2020). Bronfenbrenner’s ecological systems theory. https://www.simplypsychology.org/Bronfenbrenner Shaik, N. (2022). Transforming Teaching and Learning in Early Child hood Care and Education During COVID-19 in a Poor Commu nity of the Cape Flats, South Africa.Early Childhood Education Journal,1–10.i Hay, D. F. (2005). Early peer relations and their impact on children’s development. In: Tremblay R.E., Barr R.G., Peters R.DeV., (Eds.), Encyclopedia on Early Childhood Development (pp. 1–6). Montreal, Quebec: Centre of Excellence for Early Childhood Development; Available at: http://www.childencyclopedia.com/ documents/HayANGxp-Peers.pdf. Accessed [21/04/2022] Sinha, I. P., Harwood, R., Semple, M. G., Hawcutt, D. B., Thursfield, R., Narayan, O. … Southern, K. W. (2020). COVID-19 infection in children. The Lancet Respiratory Medicine, 8(5), 446–447. https://doi.org/10.1016/S2213-2600(20)30152-1 Hu, X., Chiu, M. M., Leung, W. M. V., & Yelland, N. (2021). Technol ogy integration for young children during COVID-19: Towards future online teaching. British Journal of Educational Technol ogy, 52(4), 1513–1537.i Stites, M. L., Sonneschein, S., & Galczyk, S. H. (2021). Preschool parents’ views of distance learning during COVID-19. Early Edu cation and Development, 32(7), 923–939. Jeon, S., Kwon, K. A., Guss, S., & Horm, D. (2020). Profiles of fam ily engagement in home-and center-based Early Head Start pro grams: Associations with child outcomes and parenting skills. Early Childhood Research Quarterly, 53, 108–123. Vuorinen, T. (2021). ‘It’s in my interest to collaborate… parents’ views of the process of interacting and building relationships with pre school practitioners in Sweden. Early Child Development and Care, 191(16), 2532–2544. Kim, E. M., & Sheridan, S. M. (2015). Foundational aspects of fam ily–school connections: Definitions, conceptual frameworks, and research needs. In E. M. Kim (Ed.), Foundational aspects of family-school partnership research (pp. 1–14). Cham: Springer International Publishing. Perrin, E.C., Leslie, L.K. & Boat T. (2016). Parenting as primary pre vention. JAMA Pediatrics, 170, 637–38. Christenson, S. L. (2003). The Family-School Partnership: An Oppor tunity to Promote the Learning Competence of All Students. School psychology quarterly, 18(4), 454. Lehrl, S., Evangelou, M., & Sammons, P. (2020). The home learn ing environment and its role in shaping children’s educational Publisher’s Note Springer Nature remains neutral with regard to juris dictional claims in published maps and institutional affiliations. 1 3 1 3
https://openalex.org/W3132519749	https://ubp.uni-bamberg.de/jfr/index.php/jfr/article/download/554/560	English	null	Changing roles of religiosity and patriarchy in women's employment in different religions in Europe between 2004 and 2016	Journal of family research	2,021	cc-by	12,527	JFR – Journal of Family Research, 2021, Vol. 33, No. 2, 405–438 doi: 10.20377/jfr-554 Ayse Guveli1 and Niels Spierings2 1 University of Essex. 2 Radboud University Nijmegen Address correspondence to: Ayse Guveli, Wivenhoe Park, Colchester, CO4 3SQ, United Kingdom. Email: aguveli@essex.ac.uk JFR – Journal of Family Research, 2021, Vol. 33, No. 2, 405–438 doi: 10.20377/jfr-554 1. Introduction Compared to other regions of the world, Europe arguably has the highest levels of gender- egalitarian views, the lowest levels of religious observance (religiosity)1, and the highest rates of women’s employment in the world, albeit with considerable variation between European countries (Bussemaker et al. 2017; Guveli & Platt 2020; Pfau-Effinger 2017; Spierings 2018). The research traditions for each of these outcomes are extensive, and are interlinked by a core recurring argument that over the course of industrialisation and educational expansion, religious observance and its importance in shaping people’s lives will fade, and traditional gender role patterns will dissolve, with a concomitant effect on women’s labour market participation (Inglehart 1997; Inglehart & Norris 2003; Guveli & Platt 2011, 2020; Guveli et al. 2016, 2017). Indeed, research has shown that religiosity is decreasing and gender egalitarianism is increasing for the majority populations in most European countries (Ruiter & Van Tubergen 2009; Guveli & Platt 2020; Spierings 2018). p ( g p g ) At the same time, however, a growing number of studies on the labour market outcomes of Muslim and migrant women in Europe have called into question the clear- cut linkages between religiosity, gender role attitudes, and women’s employment (Brynin & Guveli 2012; Guveli at al. 2016; Van Klingeren & Spierings 2020; Koenig, Maliepaard & Guveli 2016; Diehl, Koenig & Ruckdeschel 2009; Khoudja & Fleischmann 2015; Röder 2014). The labour market participation rates of migrant women, and especially of Muslim women, are disproportionately low. However, studies that have compared the roles of religiosity and patriarchal gender attitudes in accounting for this gap have generated conflicting results (Blommaert & Spierings 2019; Khattab & Hussein 2018; Khoudja & Fleischmann 2015; Guveli at al. 2016; Koopmans 2016; Pastore & Tenaglia 2013). For instance, Koopmans (2016) concluded that religiosity and patriarchal gender attitudes fully explain the Muslim employment gap, while others have found that these factors only partly explain the differences (Khoudja & Fleischmann 2015), or that religiosity has no additional explanatory power (Khoudja & Platt 2018). y ( j ) Connecting these two literatures may help to shed light on the question of whether decreases in religiosity and patriarchal gender attitudes develop simultaneously with increases in labour market participation, and on the question of whether developments in women’s labour market participation rates can be explained by changes in religiosity and gender attitudes. 1 We use the concept of “religiosity” or “religious observance” to denote the extent to which people follow and practice religious duties or prescriptions. We use “religion” or “religious group” to denote the religion with which people are affiliated. Finally, we use “no religion” to refer to people who are not affiliated with any religion. Abstract Objective: This paper seeks to understand the changing roles of religiosity and gender attitudes in the employment of women in Europe between 2004 and 2016. Background: Religiosity and gender traditionalism are both considered to decrease the likelihood of women’s employment. This study argues that this relationship needs to be decoupled, as religiosity and gender traditionalism have different underlying mechanisms. Method: We analysed rounds 2 (2004), 4 (2008), 8 (2010), and 10 (2016) of the European Social Survey (ESS), which include, among other data, information on employment, religious affiliation, religiosity, and gender role attitudes in 16 countries (N=39,233). Results: We show that taking religiosity into account further increases the already increased likelihood of employment for Catholic, Protestant, and Jewish women compared to women with no religion. We also find, however, that religiosity decreases the employment gap between Muslim and Orthodox women on the one hand and secular women on the other. Including gender role attitudes in the model only marginally explains the employment gap. Conclusion: Our findings support the idea that the mechanisms that underlie the relationships religiosity and traditional gender role attitudes have with women’s employment differ. Over time, the likelihood of employment increases for women of all religions, except for Muslim women, among whom it drops. Key words: women’s employment, religious affiliation, religiosity, gender attitudes, Europe 406 2.1 Religiosity and gender attitudes as drivers of women’s employment Research on how religiosity and patriarchal gender attitudes have affected the participation of women in the labour market is fairly extensive, and has a long tradition (Lehrer 1995; Brinkerhoff & MacKie 1988). Both factors are generally treated as part of the cultural domain of explanations (e.g., Moghadam 2013; Pettit & Hook 2005; Spierings 2016; Steiber & Haas 2012; Van der Lippe & Van Dijk 2002). For Europe, for instance, Fortin (2005) has shown that perceptions of women as primarily being homemakers, and the belief that being a housewife is just as fulfilling as working for pay, are closely associated with lower rates of women’s employment. Pastore and Tenaglia (2013) examined the influence of religion on women’s labour market participation in 47 European countries, and concluded that Eastern Orthodox and Muslim women are more likely to be jobless than agnostic women. Similarly, Guetto, Luijkx, and Scherer (2015) and Dildar (2015) found that both religiosity and traditional gender views are (indirectly) negatively associated with female labour force participation. g y p p Theoretically, such cultural explanations share a common logic: both piety and patriarchal attitudes reduce women’s labour market participation because of their interpretations of the role of women in the family and elsewhere in society. First, people’s attitudes, perspectives, and opinions inform their behaviour; thus, more conservative views on gender roles and sex segregation are related to lower levels of female employment. Second, the dominant interpretation of all major world religions is gender complementarity, whereby men are considered primarily responsible for and well-suited to earning an income, while women are mainly seen as mothers and homemakers (see also Schnabel 2016). Therefore, the more religious a group of people are, the more likely they are to have traditional gender attitudes; and, consequently, the less likely the women in the group are to enter the labour market. Third, these mechanisms are not restricted to individual-level attitudes and religiosity, as they are institutionalised through policies, norms, and views at the community or household level (Amin & Alam 2008; Pettit & Hook 2005; Spierings 2014a, 2016; Van der Lippe & Van Dijk 2002). 1. Introduction In this study, we seek to improve our understanding of these dynamics by assessing whether religiosity and patriarchal gender attitudes can (partly) explain employment changes over time and between religious groups. We theorise that the role of religiosity and gender attitudes in shaping employment decisions might have changed. This argument draws in particular from the literature on integration, Islam, and gender, which posits that in the context of secularised or European Islam, there has been a “decoupling” of gender and religion (Röder 2014; Spierings 2016; Van Klingeren & Spierings 2020). Building on this literature, we examine the question of whether the 407 impact of religiosity and gender attitudes on women’s employment has developed differently for different religions. Taking a temporal perspective, we asked the following question: How have religiosity and gender attitudes affected the likelihood of employment of women from different religious groups in Europe over time? 2. Theoretical background 2.1 Religiosity and gender attitudes as drivers of women’s employment 2.1 Religiosity and gender attitudes as drivers of women’s employment While the literature specifies and operationalises these cultural dimensions in different ways, the predominant view is that conservative religions, higher levels of religiosity, and patriarchal gender attitudes feed into lower female labour market participation rates (Amin & Alam 2008; Bozzano 2017; Camussi 2013; Davis & Gao 2020; Dildar, 2005; Guetto, Luijkx & Scherer 2015; Guveli 2011; Lisaliner & Bhatti 2005; Spierings 2014a, 2014b). Implicit in this logic is the path from religion to gender attitudes to employment. Bozzano (2017), for instance, showed that both the religious culture and 408 the degree of religiosity in a society are strongly linked to a more unequal representation of women in politics and in top managerial positions in that country. Figure 1: Direct and mediating relationships between religion, religiosity, and gender attitudes in shaping women`s employment + + (H1) (H2b) (H2a) Religious affiliation (H3c) (H3a) Religiosity Traditional gender attitudes Women’s employment + - - Religious affiliation Women’s employment Traditional gender attitudes While it is not inherent to the logic presented above, the literature has tended to discuss the linkages between religiosity, gender attitudes, and women’s employment in the context of larger modernisation processes, suggesting that the trend towards higher levels of employment among women is due to secularisation and the adoption of more progressive attitudes (e.g., Inglehart 1997; Lerner 1958; Richards & Gelleny 2007). However, some scholars have argued that these developments are recursive and nonlinear (Moghadam 1996; Pampel & Tanaka 1986; Spierings 2015, 2018; Walby 2009), drawing attention to processes and changes that occur over a long period of time. In other words, changes in the role of religiosity and gender attitudes in women’s likelihood of employment should be partly or fully due to developments (such as modernisation, industrialisation, and secularisation) over time. ) The above discussion leads to the visualisation of the mechanism (1) underlying the relationships between religion, religiosity, and gender attitudes, including the formulation of three core hypotheses: We expect to find that women who belong to a religion are less likely to be employed than women who are not affiliated with a religion (H1). Furthermore, we expect to observe that the more religious women are, the less likely they are to be employed (H2a), and that this association partly mediates the negative impact of belonging to a religion (H2b). 2.2 Changing roles of religiosity and gender attitudes in different religions 2.2 Changing roles of religiosity and gender attitudes in different religions Whereas some scholars see religiosity and gender attitudes as linked, others have suggested that the underlying mechanisms that link religiosity and patriarchy to women’ socio-economic participation and paid employment differ (e.g., Essers & Benschop 2009; Spierings 2014a). Indeed, a closer look at the studies described above suggests that these linkages are not universal. For instance, Guveli (2011) showed that in Turkey, which has a majority Muslim population, local governments with more progressive policies and platforms do not necessarily provide more labour market opportunities for women. Moreover, in their work on Indonesia, Malaysia, and Nigeria, Amin and Alam (2008) and Spierings (2014b) found no clear evidence that the labour force participation of Muslim women is lower than that of women from other religions. Finally, Davis and Gao (2020) estimated the happiness gain associated with being employed for men and women belonging to six world religions and for the non-religious, and found that this happiness gain helps to explain the gender employment gap among Buddhists, Orthodox Christians, and the non-religious; but not among Hindus, Muslims, Catholics, and Protestants. g g In short, the impact of religiosity on patriarchal gender attitudes and low levels of female employment is less clear than is often assumed, and might differ across religions. While early research on Christian denominations and Judaism has shown a fairly strong association between religiosity and gender attitudes (Lehrer 1995; Brinkerhoff & MacKie 1988), scholars of Islam and gender (Glas & Alexander 2020; Glas, Spierings & Scheepers 2018; Glas & Spierings 2019) and of the Muslim immigrant population in Europe (e.g., Diehl, Koenig & Ruckdeschel 2009; Röder 2014; Van Klingeren & Spierings 2020) arguably provide the clearest theoretical explanation for why the impact of religion varies. The patriarchal ideologies and gender traditionalism that emerged in male-dominated societies and religious traditions demand that people demonstrate their commitment and adherence to certain social norms, values, attitudes, and behaviours (Schwartz & Sagie 2000). Individuals with these attitudes assign greater importance to traditional family values and norms, which implies that men and women have distinct roles and primary responsibilities in society. These sharp divisions originate from patriarchal traditions and from the teachings of all of the Abrahamic religions, including Roman Catholicism, Protestantism, Judaism, Orthodox Christianity, and Islam (Inglehart & Norris 2003; Kandiyoti 1988; Spierings 2016; Walby 2009; Epstein 2007). 2.1 Religiosity and gender attitudes as drivers of women’s employment Likewise, we expect to find that the more traditional women’s gender attitudes are, the less likely they are to be employed (H3a), and that this association partly mediates the negative impact of belonging to a religion (H3b) and religiosity (H3c). 409 2.2 Changing roles of religiosity and gender attitudes in different religions The assumption that women should look after the home and family and prioritise motherhood, and should respect the division of labour between men and women (with the man being the primary breadwinner), are among the most salient features of patriarchal cultures. However, different interpretations of these role patterns exist within these religions. Recently, extensive research has shown that interpretations of the role of women in Muslim societies differ. Among the interpretations that have been reported are that women should be restricted to homemaking; that women and men cannot interact, but women can work in certain settings, such as in women-only factories or services; that women are primarily mothers, but can work outside the home if they do not have children; and that mothers should have an education (Badran 2015; Blommaert & Spierings 2019; Lisaniler & Bhatti 2005; Miyata & Yamada 2016; Price 2015; Spierings, Smits & Verloo 2010; Spierings 2014a). Each of these interpretations has different 410 employment implications, and can shift. Moreover, it is far from exceptional for people to combine a strongly religious orientation with progressive gender attitudes, including progressive views on women’s economic roles. Such combinations are on the rise, especially in societies where feminism and Islam are constructed as being less contradictory (Glas & Alexander 2020; Glas & Spierings 2019; Glas, Spierings, Lubbers & Scheepers 2020; Glas, Spierings & Scheepers 2018). Thus, based on these studies, we theorise that the gender role implications of being religious are not immutable, and are informed by the larger context of the society in which they are embedded. Changes in views on the religion-gender linkage reflect more recent developments in the literature that analyses the impact of religiosity and gender traditionalism on the economic participation of women with a migration background. A number of studies have shown that women with a migration background, and especially those from Muslim countries, are far less likely than native European women to be employed (Fleischmann & Hohne 2013; Khoudja & Platt 2018; Guveli at al. 2016; Koopmans 2016; Naseema & Adnan 2019; Zuccotti et al. 2017). However, these studies reached different conclusions when they looked at the extent to which religiosity or patriarchal gender attitudes explain this difference (e.g., Blommaert & Spierings 2019; Khoudja & Fleischmann 2015; Khoudja & Platt 2018; Koopmans 2016). 2.2 Changing roles of religiosity and gender attitudes in different religions Furthermore, some scholars have argued that religiosity does not explain the differences found between religious groups, attributing them instead to a “Muslim penalty” (Blommaert & Spierings 2019; Daoud & Khattab 2020; Khattab & Hussein 2018; Khattab, Johnston & Manley 2017; Abdelhadi 2017). Religion’s impact on women’s employment is informed by the larger societal context. The relationship between religion and employment among migrant women is ambiguous, and there are signs that attitudes about women’s economic activities are becoming more favourable in Europe (Spierings 2018). The question of whether the link between religiosity and employment has weakened over time has not been tested or theorised for employment, but the literature has suggested this might occur through “decoupling” (Röder 2014; Van Klingeren & Spierings 2020). That is, as women become more liberal in their gender role attitudes in increasingly diversifying and egalitarian societies, those who adhere to traditional gender roles will be gradually marginalised, and will be less likely to seek paid employment. p p y At the same time, as people’s religious affiliations and religiosity are becoming increasingly personal, and are disconnecting from dominant religious interpretations, responses to the challenges and inequalities women face are changing (Guveli 2015; Norris & Inglehart 2011). This process that has been theorised in the extensive literature on the supply side of religion (Finke & Iannaccone 1993). These scholars have argued that while demand for religion is constant over time, the supply of it is changing to meet the newly emerging needs in various societies and at different times. Therefore, religions are continuously reformulated, reinterpreted, and adapted to align with socioeconomic needs and individual lifestyles (Finke & Iannaccone 1993). The importance of the most dominant and institutionalised religions is fading, especially in Western Europe, but also increasingly in developing countries (Inglehart & Norris 2003). However, individualised religions and religiosities might provide a sense of belonging and identity, and often represent a source of spiritual and personal fulfilment. Although patriarchal gender attitudes inherently constrain women’s social, political, and labour market participation, 411 religiosity has various dimensions and manifestations that are open to multiple interpretations and experiences. Accordingly, decoupling the role of gender attitudes from religiosity in shaping women’s employment patterns takes time, and requires a relatively progressive context. This fits the setting of our study, as we are focusing on the relatively progressive context of European countries, while acknowledging the considerable degree of variation within these societies. 2.2 Changing roles of religiosity and gender attitudes in different religions When we apply the decoupling logic to our study, we expect to observe that the association between religiosity and employment gradually becomes weaker over time. At the same time, we expect to find that the connection between gender attitudes and women’s likelihood of employment increases over time because progressive gender role attitudes conflict less with the new interpretations of religiosity. Moreover, we expect to observe that the relationship between gender traditionalism and women’s labour market participation becomes stronger over time; or at least remains the same, because those individuals who continue to have conservative gender role attitudes tend to become marginalised in increasingly gender liberal societies. Consequently, we expect to find that the negative relationship between religiosity and women’s likelihood of employment weakens over time (H4), and that the negative relationship between traditional gender attitudes and women’s employment strengthens over time (H5). 2.3 Variations in changes in religiosity and gender attitudes in different religions 2 To produce the descriptive findings, we weighted the data for the population size, which generates more representative descriptive statistics. We have not used weighting for the multiple regression results because it increases the standard errors. Moreover, weighting by population size would drive the results towards the countries with the largest population sizes. Without weighting the regression results by population size, respondents count more or less equally across countries, leading to valid results across contexts. Therefore, 3. Data and methodology The European Social Survey (ESS) datasets are the only large-scale European datasets that allow researchers to make comparisons over time. The data have been used for studies on religiosity, migrant integration, and labour market participation within Europe, including on traditionalism and women’s employment (e.g., Guveli 2015; Guveli & Platt 2020; Immerzeel & van Tubergen 2013; Spierings 2018; Van Tubergen & Sindradóttir 2011; Zuccotti et al. 2017). For our purposes, the ESS datasets have two novel strengths: 1) all questions are asked the same way in all countries; and 2) all questions are asked the same way in the ESS 2004, 2008, 2010, and 2016 rounds, which allows us to trace changes over time. The ESS is designed to collect high-quality data on the beliefs, attitudes, and behaviours of representative samples of the populations of the participating countries, and enables the analysis of continuity and change over time. We have chosen to use these ESS rounds because they include information on paid employment, on religiosity, and on gender role attitudes; as well as on education, marital status, children living at home, and migration status. g The countries in our sample participated in at least three of the four rounds, and we included only countries with more than 30 respondents for each religious group. As we were analysing women’s employment, we only selected women, and excluded respondents for whom we had missing data on other measures, including our key dependent measure of paid employment. Our final analytic sample comprised 39,233 respondents. 2.3 Variations in changes in religiosity and gender attitudes in different religions Finally, the decoupling logic presented above might take different forms among women in different religious groups. Religiosity and gender ideologies are more closely linked in traditional societies than in secular societies. Given the relatively progressive context of Europe, decoupling is more likely to occur in more conservative religions because the connection is strongest in these religions. g g Moreover, the decoupling of religiosity and gender traditionalism might already have taken place in highly secularised religions, such as in the Roman Catholic, Protestant, and Jewish faiths, whereas gender traditionalism and religiosity might be more interlinked in more traditional religions, such as in Islam (Inglehart & Norris 2003). Some scholars have argued that European societies are still in the process of developing a Euro-Islam or secularised version of Islam (Asad 2003; Cesari 2009, 2015); i.e., that Islam in Europe and in other developing countries is evolving in response to modernisation processes, with new interpretations and their everyday reflections engendering individualised Islams, and fuelling the decoupling of religiosity from patriarchy. Thus, we might expect the role of religiosity and gender traditionalism in shaping women’s labour market outcomes to be more pronounced for Muslim women because of the relatively strong connections between traditionalism and religion in Islam (Guveli & Platt 2020; Inglehart & Norris 2003). A similar argument could be made for the Eastern Orthodox religion, which is arguably the second-most traditional of the main religions in Europe (Inglehart & Norris 2003). ) The reasoning above implies that, particularly among Muslim and Orthodox women, the likelihood of being employed will increase more among the most religious, and less 412 among the most gender traditional. Consequently, we might expect to find that over time, the negative relationship between religiosity and women’s employment weakens more for Muslim and Orthodox women than for other women (H6). Likewise, we would expect to observe that over time, the negative relationship between traditional gender attitudes and women’s likelihood of employment strengthens more for Muslim and Orthodox women than for other women, or stays the same (H7). 3.1 Dependent and independent variables Our dependent variable was paid employment, measured with a question on whether respondents were in employment in the last seven days, coded as 1 = employed, and as zero otherwise. We dropped respondents still in education from the analysis, and only included respondents aged 18 to 65, because those are considered the working ages in most countries. Table 1 shows the mean employment rates of women in the countries and across the ESS rounds between 2004 and 2016.2 The table indicates that there was an 413 overall increase in the employment rates of women over this period, with some fluctuations in some countries – including declining or stagnating rates in 2010 in most countries, which were most likely due to the 2008 financial crisis (Spierings 2018). The highest female employment rates were in the Nordic countries of Denmark, Norway, and Sweden; while the lowest rate was in Greece. Religious affiliation was measured with the question: “Do you consider yourself as belonging to any particular religion or denomination?” We allocated those who answered “no” to the category of “no religion”. For those who answered “yes”, the options were: Roman Catholic; Protestant; Eastern Orthodox; other Christian denomination; Jewish; Islamic; Eastern religions; other non-Christian religions. We combined Eastern religions and other non-Christian religions into a single “Other” category. Because of the small size and heterogeneity of this category, we included it in analyses for completeness, but we do not discuss it in detail here. Table 1: Share of women’s employment in countries across ESS rounds ESS round 2 2004 ESS round 4 2008 ESS round 5 2010 ESS round 8 2016 Belgium 55.1 61.7 66.2 65.1 Switzerland 68.7 67.7 75.1 76.2 Czech Republic 55.9 61.2 60.9 72.2 Germany 59.4 68.5 66.1 75.4 Denmark 76.0 77.0 74.2 No survey Spain 58.2 58.5 59.5 63.7 France No survey 64.0 67.6 65.3 UK 63.3 62.4 63.7 72.2 Greece 41.2 55.3 43.8 No survey Ireland 57.5 49.0 45.7 60.0 Israel No survey 59.3 57.2 70.2 Netherlands 61.3 64.3 69.2 70.1 Norway 79.1 81.7 81.6 80.4 Russia No survey 65.4 65.2 67.4 Sweden 78.2 84.8 82.1 81.1 Slovenia 59.9 58.0 60.5 62.6 Total N 10,299 9,247 13,283 10,820 Source: European Social Surveys Notes: Findings are weighted for population size. we display the unweighted regression results here, but we include the weighted regression results for Table 4 in Appendix A3. They provide similar conclusions. 4 in Appendix A3. They provide similar conclusions. 3 There are two other measures for religiosity in the datasets. One is subjective religiosity, which was measured with the question: “Regardless of whether you belong to a particular religion, how religious would you say you are?” with responses on a scale from zero (not at all religious) to 10 (very religious). We ran the analyses by replacing praying with subjective religiosity, which provided similar conclusions (available upon request). Since subjective religiosity is more sensitive to contextual factors, such as discriminatory experiences or certain societal debates or events that can make identities more manifest (Guveli 2015), we decided to use information on the frequency of praying as our religiosity variable. This variable is less sensitive to contextual factors because it asks about the frequency of praying. We did not use the religiosity indicator attendance at religious meetings because of the high shares of missing values, especially for Muslim women, and because it measures exposure and socialisation, whereas our theoretical focus was on individual forms of religiosity. We did not construct a scale using all three measures because different mechanisms underlie each of these measures, and they are conceptually distinct across religions and between migrants and natives, with different effects on gender issues (Guveli 2015; Spierings 2019, Van Klingeren & Spierings 2020). 3.1 Dependent and independent variables Table 1: Share of women’s employment in countries across ESS rounds We included ESS rounds 2 (2004), 4 (2008), 5 (2010), and 8 (2016) in all models, which serve as proxies for the time trends. Table 2 shows the proportion of all religions and no religion in each ESS round (2, 4, 5, and 8). The largest category in the table is the “no religion” group; Roman Catholic women form the largest religious group, followed by Protestant women, and then by Eastern Orthodox women. Muslims constitute about two 414 per cent of the sample in 2004 and four per cent of the sample in 2016, and thus represent one of the smallest religious groups in the ESS data (and in Europe). Table 2: Share of religious groups across ESS rounds Round 2 2004 Round 4 2008 Round 5 2010 Round 8 2016 Total No religion 47.3 42.0 42.8 45.1 44.0 Roman Catholic 28.8 20.5 19.7 19.5 21.2 Protestant 15.7 12.4 10.8 10.2 11.8 Eastern Orthodox 5.1 19.4 19.7 17.7 16.9 Jewish 0.02 1.4 1.2 1.2 1.1 Islam 1.5 3.1 3.6 4.2 3.3 Other religions 1.6 1.2 2.0 2.1 1.8 Total N 9,093 9,018 11,610 9,512 39,233 Notes: Findings are weighted for population size. Table 2: Share of religious groups across ESS rounds Our two main independent variables were religiosity and gender role attitudes. To operationalise religiosity, we used information on how often respondents prayed,3 which was collected with the following question in the survey: “How often do you pray apart from at religious services?” The answer categories were: 0 = never; 1 = less often; 2 = only on special holidays; 3 = at least once a month; 4 = once a week; 5 = more than once a week; 6 = every day. The question on gender role attitudes was repeated in all selected ESS rounds (2004, 2008, 2010, and 2016) using the following statement: “Men should have more right to a job than women when jobs are scarce”. The answer categories ranged from 1 = disagree strongly to 5 = agree strongly. 4 We have operationalised time with “ESS-round” as a covariate in our regression models, which shows a fairly linear relationship with the dependant variable employment. However, there is a slight diversion from linearity in the year 2010 (round 5) across all religions and countries, as Table 1 and Figure 2 demonstrate. 3.2 Control variables We included a number of individual-level characteristics shown to be important in determining women’s labour market participation. We included age to control for whether 415 age might (partly) account for the relationship between religiosity and labour market participation. Age might also (partly) account for the relationship between gender role attitudes and women’s likelihood of employment. That is, it might be the case that gender attitudes and religiosity are more common among older women in some religions, while they are less dependent on age in other religions. Family structure and the presence of dependent children are other important factors in women’s labour market participation levels (Naseema & Adnan 2019; Spierings 2014a; Van der Lippe & Van Dijk 2002). Therefore, we included marital status as a categorical variable: 1 = married/cohabiting/in legal partnership; 2 = divorced/widowed/separated; 3 = never married. Finally, for having children, we added a dummy variable: 1= having children living at home, and zero otherwise. As the ESS provides information on each respondent’s country of birth, as well as on the birthplace of her/his father and mother, we could take into account the differences between majority Europeans (natives) and first- and second-generation migrants. We defined the categories as follows: natives – neither the respondent nor either of the respondent’s parents was born abroad; migrants (first generation) – the respondent and both parents were born abroad; second generation – at least one of the respondent’s parents was born abroad, but the respondent was born in the survey country. y y A key variable commonly linked to labour market participation is educational level. Women with higher qualifications are more often economically active, even if they are married and have children, although some may be more likely to experience unemployment due to discrimination in recruiting practices or to their choices or preferences based on their religion (Bayrakdar & Guveli 2020; Blommaert, Coender & Van Tubergen 2014a, 2014b; Daoud & Khattab 2020; Khattab & Hussein 2018; Naseema & Adnanb 2019; Zuccotti et al. 2017; Güveli 2006). We included education measured in years, as this was the only feasible way of proxying educational attainment across a diverse range of countries. A description of all of the variables is provided in Appendix A1. 3.3 Analytical approach Our dependent variable paid employment was dichotomous; therefore, we estimated a series of logistic regressions. In our base model (Model 1), along with the control variables in all models (age, age2, marital status, children living at home, education, migration status) for compositional influences on paid work, we included only the main effects for religions and the time trend (ESS rounds)4. Model 2 examined the relationship between religiosity (praying) and women’s paid employment. Model 3 looked at the relationship between gender role attitudes and paid employment without the variable religiosity. Model 4 added both religiosity and gender attitudes to determine whether they explained the differences in women’s likelihood of employment in different religious groups. These models allowed us to test Hypotheses 1 through 3. 416 We also estimated models with two-way interactions between time (survey years) and religiosity (Model 5, testing Hypothesis 4), and time and gender role attitudes (Model 6, testing Hypothesis 5). Our final model (Model 7) incorporated all two-way interactions and a three-way interaction between time, religiosity, and religions, as well as a three-way interaction between time, gender role attitudes, and religions, to reveal changes in the association between employment and gender role attitudes and religiosity in particular religions (Hypotheses 6 and 7). Models 5, 6, and 7 are presented in Appendix A2. g ( yp ) p pp In the following, we discuss our findings on the evolution of the effects of religiosity and gender role attitudes on women’s paid employment. The odds ratios of the logistic regression are presented in Appendix Table A2. For the purposes of visual presentation and ease of interpretation, particularly given the challenges of reading results from three- way interactions in tables, we present our main results for Model 7 (Table A2) in graphical form in Figures 1 and 2. We discuss the results from other models to the extent they are relevant to answering our main research question. For completeness, we provide the full sequence of models (Models 5, 6, and 7) in Appendix Table A2. Note that our aim was not to estimate contextual effects. Rather, we were analysing the evolution of the effects of religiosity and gender attitudes on women’s paid employment over time (across 16 European countries, or, more precisely, ESS countries). 3.3 Analytical approach There has been some discussion of the best models to apply in the analysis of cross- national surveys like the ESS (Bryan & Jenkins 2016; Te Grotenhuis et al. 2015). To account for all of the country-level factors that might be associated with the distribution of female employment rates, religiosity, and gender attitudes across countries, we included country fixed effects in all of our models (Clarke et al. 2015). This is a conservative approach (Bryan & Jenkins 2016), but it enabled us to identify the patterns we were most interested in. 4.1 Bivariate results Over time, women in all religions became increasingly likely to be in paid employment, as Figure 2 shows – albeit with declining or stagnating rates in 2010 across religious groups and countries, which were likely due to the 2008 financial crisis. This pattern is also clear in Table 1, which displays women’s employment rates per country. Protestant women had a higher percentage of employment than women in any other group, and a percentage that was very similar to that of women with no religion across all time points between 2004 and 2016. Muslim women had the lowest percentage of employment, but they had the largest increase in employment between 2004 and 2016. While Eastern Orthodox women had the second-lowest percentage of employment in 2004, their employment rates also increased in this period. Women with no religious affiliation did not have the highest employment share, which contradicts the assumption that belonging to a religion in general impedes women’s employment. 417 Figure 2: Religious groups and their employment rates across ESS rounds (times) Notes: Weighted results for population size .4 .6 .8 .4 .6 .8 .4 .6 .8 2004 2008 2010 2016 2004 2008 2010 2016 2004 2008 2010 2016 No Religion Roman Catholic Protestant Eastern Orthodox Judaism Islam Other Religions Probability having paid work Survey year No Religion Roman Catholic Probability having paid work Judaism Survey year Notes: Weighted results for population size Table 3 shows the rank correlation between gender attitudes and religiosity in each religious group over time. The overall correlation between them for all religions was moderate to weak; i.e., it was 0.18 in 2004, and had declined to 0.12 in 2016. The strongest correlations in 2004 were for Eastern Orthodox women, Muslim women, and women in other religions, but they decreased sharply over time. Our finding that these correlations were declining supports the claim that there was a decoupling of gender attitudes and religiosity in religions from industrialising countries (Eastern Orthodox, Islam, and Other religions). The correlation between religiosity and gender attitudes was weak but relatively stable over time for women in all other religious groups and for women with no religion, which indicates that the decoupling had already taken place in religions from post- industrial and highly secularised European countries. 418 Table 3: Spearman correlation coefficient between religiosity and gender attitudes, per religious groups and ESS rounds g g p Round 2 2004 Round 4 2008 Round 5 2010 Round 8 2016 Overall correlation 0.18 0.17 0.17 0.12 No religion 0.06 0.08 0.10 0.10 Roman Catholic 0.08 0.06 0.07 0.06 Protestant 0.17 0.18 0.19 0.17 Eastern Orthodox 0.25 0.14 0.07 0.03 Jewish 0.15 0.11 0.17 0.19 Islam 0.21 0.09 0.02 0.06 Other religions 0.27 0.08 0.12 0.15 Note: Weighted for population size 4.2 Multiple regression results However, adding traditional gender attitudes explained only a marginal part of the differences observed between religious groups, refuting Hypothesis 3b. Moreover, when both religiosity and gender role attitudes were added to Model 4, the results confirmed the previous results, and suggested that gender role attitudes did not really mediate religiosity’s impact. women when jobs are scarce were significantly less likely to be in paid employment than women who disagreed with this statement. However, adding traditional gender attitudes explained only a marginal part of the differences observed between religious groups, refuting Hypothesis 3b. Moreover, when both religiosity and gender role attitudes were added to Model 4, the results confirmed the previous results, and suggested that gender role attitudes did not really mediate religiosity’s impact. Furthermore, when religiosity and gender role attitudes were added together, their coefficients were not different from those in Model 2 and Model 3, which corresponded with the low correlations shown in Table 3. That is, the roles of religiosity and of gender attitudes in shaping women’s likelihood of employment were found to be independent of each other. This finding refuted our H3c, which stated that the negative relationship between religiosity and employment is mediated by gender role attitudes. Model 1 demonstrated that over time, the likelihood of employment increased significantly. In contrast to our findings regarding the differences between religious affiliations, adding religiosity hardly changed this positive trend, but adding gender role attitudes partly and marginally explained this trend. Religiosity and gender role attitudes thus seemed to explain the different patterns in women’s employment. Table 4: Odds ratios for the effects of religiosity and gender role attitudes on the likelihood of employment Notes: Standard errors in parentheses; all models include country fixed effects; all models are controlled for age, age2, education, marital status, having children at home, and migration status; Weighted results are demonstrated in Appendix A3; * p<0.001, p<0.01, * p<0.05, + p<0.1. 4.2 Multiple regression results 4.2 Multiple regression results To determine whether these differences were significant and to take into account other factors that might be driving them, we performed multiple regression analysis. Table 4 shows the results for the four models used to test our expectations. We accounted for country-level differences with fixed effects in all models, but to avoid overly long tables, these coefficients are not displayed in the table. The table shows the odds ratios: the closer the coefficient is to one, the smaller the effect; coefficients below one indicate negative effects, and coefficients above one indicate positive effects. p Model 1 shows that compared to women who were not affiliated with any religion, Eastern Orthodox and Muslim women and those from other religions were significantly less likely to be in paid employment. Jewish women were significantly more likely to have a paid job than women with no religion. There were no significant differences between Catholic and Protestant women. Thus, our first hypothesis was only partly confirmed (H1). With religiosity added, Model 2 shows that the women who prayed more often were significantly less likely to have a paid job. Religiosity fully explained the lower likelihood of employment for Eastern Orthodox women and for women in other religions than for non- religious women; while the gap between Muslim women and those with no religion also decreased, as the coefficient moved closer to zero. The results confirm Hypotheses 2a and 2b. That is, religiosity partly explained the lower employment rates of Muslim women, but it accounted for most of the differences in the employment rates of Muslim and secular women. Moreover, once religiosity was taken into account, the insignificant but negative relationship between Catholic women and the likelihood of employment (in Model 1) became significant and positive (Model 2). That is, Catholic and Protestant women were significantly more likely to be in paid employment than women with no religion when taking the religiosity of these women into account. Model 3 of Table 4 shows the relationship between gender attitudes and the likelihood of employment. As expected, and in line with previous research findings and Hypothesis 3a, women agreeing with the statement that men have more right to have a job than 419 women when jobs are scarce were significantly less likely to be in paid employment than women who disagreed with this statement. 4.3 Evolution of the roles of religiosity and gender attitudes 4.3 Evolution of the roles of religiosity and gender attitudes To assess the changing influences of religiosity and gender attitudes over time, we interacted the time variable (survey year/ESS round) with religiosity (Model 5, Appendix A2) and gender attitudes (Model 6), and added the three-way interactions with the religious affiliations (Model 7). To estimate the three-way interactions correctly, we also interacted time with the religious affiliations (Model 7). The results for Model 7 showed no stark differences in the changes over time in the likelihood of employment for women across religious groups, except for Muslim women. Controlling for other factors in the model, the likelihood of employment for Muslim women fell during this period (Model 7 of Appendix A2), which may indicate that they were hit hardest by the economic crisis, and were “last in line” (Blommaert & Spierings 2019). This downwards trend in the likelihood of employment for Muslim women found in the multiple regression model was contrary to the finding in the bivariate result (i.e., not controlled for education, parenthood, etc.) in Figure 2, which shows a substantial increase in the employment rate for Muslim women between 2004 and 2016. These differences might indicate that there were changes in Muslim women’s characteristics that led to an increase in their employment rate, such as an increasing level of education over time. This issue merits more detailed scrutiny in future research. y We had expected to find that over time, the effect of religiosity on the paid employment of women would weaken, while the association between gender role attitudes and employment would become stronger. The association between religiosity and paid employment was negative in 2004, but we saw no change in this association over time (in Model 5 of Table A2), thus refuting H4. The association between traditional gender role attitudes and women’s likelihood of employment was significant and negative in 2004, but this negative impact on employment weakened over time (Model 6 of Table A2). This result contradicted our expectations, and thus refuted H5. p Finally, in Model 7, we assessed whether these patterns differed by religious affiliation (Appendix A2), as shown in Figures 3 and 4. Figure 3 plots the predicted probabilities for women who never prayed, who prayed at least once a month, and who prayed every day – the two opposite and marginal categories and the middle category. 4.2 Multiple regression results Model 1 Model 2 Model 3 Model 4 Survey year (ESS round) 1.065* 1.063* 1.046* 1.045* (0.0066) (0.0066) (0.0066) (0.0066) Religions (ref: No religion) Roman Catholic 0.983 1.116 1.006 1.125 (0.0332) (0.0407) (0.0341) (0.0412) Protestant 1.066 1.194*** 1.089* 1.204*** (0.0457) (0.0532) (0.0468) (0.0537) Eastern Orthodox 0.855* 0.964 0.871* 0.969 (0.0599) (0.0687) (0.0612) (0.0693) Jewish 1.347* 1.429** 1.286+ 1.360* (0.1782) (0.1892) (0.1709) (0.1806) Muslim 0.463* 0.548* 0.502* 0.582* (0.0373) (0.0452) (0.0407) (0.0483) Other religions 0.765 0.916 0.793 0.930 (0.0676) (0.0828) (0.0703) (0.0843) Praying (1-5) 0.946* 0.951* (0.0056) (0.0057) Traditional gender attitudes (0-6) 0.841* 0.846* (0.0093) (0.0094) N 39,233 39,233 39,233 39,233 Table 4: Odds ratios for the effects of religiosity and gender role attitudes on the likelihood of employment N 420 4.3 Evolution of the roles of religiosity and gender attitudes Figure 3 shows that when other variables were taken into account (Model 7 in Appendix A2), religiosity was not a significant determinant of women’s likelihood of employment over time for secular, Roman Catholic, Protestant, and Eastern Orthodox women. This pattern is visualised by the lines in Figure 3 being close to each other instead of showing substantial deviation. However, there were some noteworthy patterns for Jewish and Muslim women, in line with Hypothesis 6. Over time, the gap between more and less religious women closed, which made religiosity a weaker determinant of employment for the women in these groups. For instance, in 2004, the likelihood of employment for Jewish women who never prayed was substantially higher than that for Jewish women who prayed every day; whereas in 2016, the likelihood of employment increased significantly for Jewish women overall, but the difference between these two groups was insignificant. As we derived our logic mostly from the literature on migration and Islam, it was interesting to observe that Muslim women who never prayed had a higher likelihood of being in paid employment in 421 2004 than their counterparts who prayed every day, but that these groups converged over time without increasing their overall likelihood of paid employment. 2004 than their counterparts who prayed every day, but that these groups converged over time without increasing their overall likelihood of paid employment. Figure 3: Average predicted probabilities from model results. Effects of religiosity (“How often do you pray apart from at religious services?”) on the employment of women belonging to different religions between 2004 and 2016 Figure 3: Average predicted probabilities from model results. Effects of religiosity (“How often do you pray apart from at religious services?”) on the employment of women belonging to different religions between 2004 and 2016 Note: Controlled for country, age, age2, education, marital status, having children at home, and migration status. .4 .6 .8 .4 .6 .8 .4 .6 .8 2004 2008 2010 2016 2004 2008 2010 2016 2004 2008 2010 2016 No Religion Roman Catholic Protestant Eastern Orthodox Judaism Islam Other Religions Never At least once month Every day Probability having paid work Survey year Survey year Note: Controlled for country, age, age2, education, marital status, having children at home, and migration status. 4.3 Evolution of the roles of religiosity and gender attitudes To analyse gender role attitudes, we distinguished between women who strongly disagreed, neither agreed nor disagreed, and strongly agreed with the statement: “Men should have more right to a job than women when jobs are scarce”. The results appear in Figure 4, which are based on Model 7 in Appendix A2. The association between gender role attitudes and women’s employment remained fairly constant for secular, Roman Catholic, and Protestant women (i.e., the lines are rather horizontal), but the association for Eastern Orthodox women showed some divergence over time, which was in line with our expectation in Hypothesis 7. That is, Eastern Orthodox women with more egalitarian gender attitudes were more likely to be in paid employment than their counterparts with conservative gender role attitudes in 2004, and the likelihood of employment of these two groups was becoming less similar over time. For Orthodox women, this finding was in line with Hypothesis 7. However, the overall expectation was not confirmed in the results for Jewish and Muslim women, and was only somewhat confirmed for Protestants. The 422 likelihood of employment for women with traditional gender role attitudes increased considerably over time for Jewish and Muslim women, and for women in other religions. The biggest differences in the likelihood of employment were between the progressive and the traditional Muslim women in 2004; the most progressive women had the highest likelihood of employment, and the most traditional had the lowest. Over time, the differences converged, with the likelihood of being in paid employment increasing for traditional Muslim women and slightly decreasing for the most progressive women; this pattern was similar to, but less pronounced than, the association between religiosity and the likelihood of employment over time for Muslim women. Figure 4: Average predicted probabilities from model results. Gender attitudes (“Men should have more right to a job than women when jobs are scarce”) in the employment of women belonging to different religions between 2004 and 2016 Figure 4: Average predicted probabilities from model results. Gender attitudes (“Men should have more right to a job than women when jobs are scarce”) in the employment of women belonging to different religions between 2004 and 2016 Note: Controlled for country, age, age2, education, marital status, having children at home, and migration status. 4.3 Evolution of the roles of religiosity and gender attitudes .2 .4 .6 .8 .2 .4 .6 .8 .2 .4 .6 .8 2004 2008 2010 2016 2004 2008 2010 2016 2004 2008 2010 2016 No Religion Roman Catholic Protestant Eastern Orthodox Judaism Islam Other Religions Agree strongly Neither agree nor disagree Disagree strongly Probability having paid work Survey year Probability having paid work Survey year Agree strongly Disagree strongly Note: Controlled for country, age, age2, education, marital status, having children at home, and migration status. 5. Conclusion At the outset, we argued that religiosity and traditional gender role attitudes may impede women’s likelihood of employment, and that these factors might explain the differences 423 in the employment rates of women in different religious groups (Inglehart & Norris 2003). After reviewing the literature on the decoupling of religiosity and gender attitudes, we expected to find that the role of religiosity in employment weakens, whereas the negative association between gender attitudes and employment becomes stronger over time. We examined these expectations using ESS data ranging from 2004 to 2016. in the employment rates of women in different religious groups (Inglehart & Norris 2003). After reviewing the literature on the decoupling of religiosity and gender attitudes, we expected to find that the role of religiosity in employment weakens, whereas the negative association between gender attitudes and employment becomes stronger over time. We examined these expectations using ESS data ranging from 2004 to 2016. p g g g If anything, our results underscore that it is far too simple to assume that there is a recursive process in which women’s likelihood of employment increases across the board, or to use religiosity and gender role attitudes to explain employment differences between religious groups, and particularly the Muslim employment gap (Khattab & Hussein 2018; Khattab et al. 2017). Both religiosity and traditional gender role attitudes were shown to impede employment, but only religiosity really explained the differences between different religious groups, and only traditional gender role attitudes slightly explained the developments over time. Moreover, after taking both factors into account, some group differences remained, mainly between Muslim women and other women. After controlling for gender role attitudes and religiosity, Catholic, Protestant, and Jewish women had the highest likelihood of employment of all the religious groups studied. When religiosity was considered, the higher employment rates for Catholic, Protestant, and Jewish women relative to women with no religion became more prominent, while the lower employment rates for Muslim and Eastern Orthodox women decreased. This finding of our multiple regression analysis was in line with our descriptive findings that women with no religion had the highest employment rates by far. Future research should examine the dynamics at play in the association between religiosity and women’s employment for women of different religions. Delving more deeply, we found that different religious groups were also affected differently by gender attitudes and religiosity. 5. Conclusion Notably, over time, the effects of religiosity decreased among Jewish and Muslim women; and the effects of traditional gender role attitudes decreased among Jewish, Muslim, and, to some extent, Protestant women. For all other combinations, the developments appeared to be parallel, with no strongly increasing effects. g While these results support our overarching expectation that women’s employment rates would not follow the same trajectories for all religious groups, our specific expectations were only partly supported. We found support for the idea that religiosity and traditional gender role attitudes have their own underlying mechanisms that explain their relationships to women’s employment. However, our findings contradicted the expectation that the negative association between gender role attitudes and likelihood of employment would become stronger over time. In fact, it became weaker. That is, the likelihood of employment for women with more traditional gender attitudes increased over time between 2004 and 2016 – although a somewhat different trend was observed for Eastern Orthodox women, which was in line with our expectations. One explanation for these findings might be that the role of gender role attitudes in entering paid employment became less important because these attitudes were overridden by economic pressures in societies where men and women were expected to earn their own living. However, future research should investigate this question more in detail. Our research represents a starting point in the analysis of gender traditionalism and religiosity and their associations with the employment of women belonging to different 424 religions. Future research should study these associations in longer time windows, and look at changes over the life course. Including contextual factors in the analysis might shed more light on the significantly lower and decreasing likelihood of employment found for Muslim women, even after religiosity and gender traditionalism are taken into account. To conduct a detailed analysis, more data and a longer time span are needed. Such an analysis might also provide us with a better understanding of our finding that the association between gender traditionalism and employment has been decreasing over time among Muslim women in Europe. Acknowledgments We thank the ESRC research Centre on Micro-Social Change (MiSoC) for funding this research (award number ES/S012486/1). References Abdelhadi, E. (2017). Religiosity and Muslim women’s employment in the United States. Socius: Sociological Research for a Dynamic World, 3, 1-17. Abdelhadi, E. (2017). Religiosity and Muslim women’s employment in the United States. Socius: Sociological Research for a Dynamic World, 3, 1-17. https://doi.org/10.1177/2378023117729969 https://doi.org/10.1177/2378023117729969 Amin, S. & Alam, I. (2008). Women’s employment decisions in Malaysia: Does religion matter? The Journal of Socio-Economics, 37, 6, 2368-2379. https://doi.org/10.1016/j.socec.2008.04.012 Asad, T. (2003). Formations of the secular: Christianity, Islam, modernity. Stanford: Stanford University Press. Badran, M. (2005). Between secular and Islamic feminism/s reflections on the Middle East and Beyond. Journal of Middle East Women’s Studies, 1, 1, 6-28. https://muse.jhu.edu/article/408727 Bayrakdar, S. & Guveli, A. (2020). Understanding the benefits of migration: Multigenerational transmission, gender and educational outcomes of Turks in Europe. Journal of Ethnic and Migration Studies, online first, 1-22. https://doi.org/10.1080/1369183X.2020.1736531 p // g/ / Blommaert, L., Coenders, M. & Van Tubergen, F. (2014a). Ethnic discrimination in recruitment and decision makers’ features: Evidence from laboratory experiment and survey data using a student sample. Social Indicators Research, 116, 3, 731-754. https://doi.org/10.1007/s11205-013-0329-4 p // g/ / Blommaert, L., Coenders, M. & Van Tubergen, F. (2014b). Discrimination of Arabic- named applicants in the Netherlands: An internet-based field experiment examining different phases in online recruitment procedures. Social Forces, 92, 3, 957-982. https://doi.org/10.1093/sf/sot124 425 Blommaert, L. & Spierings, N. (2019). Examining ethno-religious labor market inequalities among women in the Netherlands. Research in Social Stratification and Mobility, 61, 38-51. https://doi.org/10.1016/j.rssm.2019.01.005 Bozzano, M. (2017). On the historical roots of women’s empowerment across Italian provinces: Religion or family culture? European Journal of Political Economy, 49, 24-46. https://doi.org/10.1016/j.ejpoleco.2016.12.002 Brinkerhoff, M.B. & MacKie, M. (1988). Religious sources of gender traditionalism. In: Thomas, D.L. (ed.): The religion and family connection: Social Science Perspectives. Provo, UT: Religious Studies Center, Brigham Young University, 232-57. g g g y Bryan, M.L. & Jenkins, S.P. (2016). Multilevel modelling of country effects: A cautionary tale. European Sociological Review, 32, 1, 3-22. https://doi.org/10.1093/esr/jcv059 Brynin, M. & Guveli, A. (2012). Understanding the ethnic pay gap in Britain. Work, Employment and Society, 26, 4, 574-587. https://doi.org/10.1177/0950017012445095 Bussemakers, C., van Oosterhout, K., Kraaykamp, G. & Spierings, N. (2017). Women’s worldwide education–employment connection: a multilevel analysis of the moderating impact of economic, political, and cultural contexts. World Development, 99, 28-41. https://doi org/10 1016/j worlddev 2017 07 002 p // g/ /j Camussi, S. (2013). Female labour market participation and cultural variables. Occasional Paper No. 178. References Bank of Italy. http://dx.doi.org/10.2139/ssrn.2297874 Cesari, J. (2009). The securitisation of Islam in Europe (Vol. 15). CEPS. Cesari, J. (Ed.). (2015). The Oxford handbook of European Islam. Oxford: Oxford UP. Clarke, P., Crawford, C., Steele, S. & Vignoles, A. (2015). Revisiting fixed- and random- effects models: some considerations for policy-relevant education research. Education Economics, 23, 3, 259-277. https://doi.org/10.1080/09645292.2013.855705 Daoud, Y. & Khattab, N. (2020). Women, labour market outcomes and religion: Evidence from the British labour market. Review of Social Economy, (online). https://doi.org/10.1080/00346764.2020.1739321 Davis, L. & Gao, J. (2020). Preferences or patriarchy: Why do religious women work less? Social Indicators Research, 147, 287-310. https://doi.org/10.1007/s11205-019-02152-6 Diehl, C., Koenig, M. & Ruckdeschel, K. (2009). Religiosity and gender equality: Comparing natives and Muslim migrants in Germany. Ethnic and Racial Studies, 32, 2, 278-301. https://doi.org/10.1080/01419870802298454 Dildar, Y. (2015). Patriarchal norms, religion, and female labour supply: Evidence from Turkey. World Development, 76, 40-61. https://doi.org/10.1016/j.worlddev.2015.06.010 p // g/ /j Epstein, C.F. (2007). Great divides: The cultural, cognitive, and social bases of the global subordination of women. American Sociological Review, 72, 1, 1-22. https://doi.org/10.1177/000312240707200101 426 Essers, C. & Benschop, Y. (2009). Muslim businesswomen doing boundary work: The negotiation of Islam, gender and ethnicity within entrepreneurial contexts. Human Relations, 62, 3, 403-423. https://doi.org/10.1177/0018726708101042 Finke, R. & Iannaccone, L.R. (1993). Supply-side explanations for religious change. The Annals of the American Academy of Political and Social Science, 527, 1, 27-39. https://doi.org/10.1177/0002716293527001003 p // g/ / Fleischmann, F. & Höhne, J. (2013). Gender and migration on the labour market: Additive or interacting disadvantages in Germany? Social Science Research, 42, 5, 1325-1345. https://doi.org/10.1016/j.ssresearch.2013.05.006 Fortin, N.M. (2005). Gender role attitudes and the labour-market outcomes of women across OECD countries. Oxford review of Economic Policy, 21, 3, 416-438. https://doi.org/10.1093/oxrep/gri024 g g Glas, S. & Alexander, A. (2020). Explaining support for Muslim feminism in the Arab Middle East and North Africa. Gender & Society, 34, 3, 437-466. https://doi.org/10.1177/0891243220915494 Glas, S. & Spierings, N. (2019). Support for feminism among highly religious Muslim citizens in the Arab region. European Journal of Politics and Gender, 2, 2, 283-310. https://doi.org/10.1332/251510819X15538590890492 Glas, S., Spierings, N., Lubbers, M. & Scheepers, P. (2019). How polities shape support for gender equality and religiosity’s impact in Arab countries. European Sociological Review, 35, 3, 299-315. https://doi.org/10.1093/esr/jcz004 Glas, S., Spierings, N. & Scheepers, P. (2018). Re-understanding religion and support for gender equality in Arab countries. Gender & Society, 32, 5, 686-712. https://doi:10.1177/0891243218783670 Guetto, R., Luijkx, R. References & Scherer, S. (2015). Religiosity, gender attitudes and women’s labour market participation and fertility decisions in Europe. Acta Sociologica, 58, 2, 155-172. https://doi.org/10.1177/0001699315573335 Güveli, A. (2006). New social classes within the service class in the Netherlands and Britain. Adjusting the EGP class schema for the technocrats and the social and cultural specialists. Nijmegen: Dissertation Radboud University Nijmegen. Guveli, A. (2011). Social and economic impact of the headscarf ban on women in Turkey. European Societies, 13, 2, 171-189. https://doi.org/10.1080/14616696.2010.547942 Guveli, A. (2015). Are Movers More Religious than Stayers? Religiosity of European Majority, Turks in Europe and Turkey. Review of Religious Research, 57, 1, 43-62. http://www.jstor.org/stable/43920079 j g Guveli, A. & Platt, L. (2011). Understanding the religious behaviour of Muslims in the Netherlands and the UK. Sociology, 45, 6, 1008-1027. https://doi.org/10.1177/0038038511416165 Guveli, A. & Platt, L. (2020). Recent trends in religiosity of majority and minority European populations, Social Policy Working Paper 02-20, London: LSE Department of Social Policy. 427 Guveli, A., Ganzeboom, H.B.G., Baykara-Krumme, H., Platt, L., Eroglu, S., Spierings, N., Bayrakdar, S., Nauck, B. & Sozeri, E. (2017). 2,000 Families: Identifying the research potential of an origins-of-migration study. Ethnic and Racial Studies, 40, 14, 2558-2576. https://doi.org/10.1080/01419870.2016.1234628 p // g/ / Guveli, A., Ganzeboom, H.B.G., Platt, L., Nauck, B., Baykara-Krumme, H., Eroğlu, Ş., Bayrakdar, S., Sözeri, E.K. & Spierings, N. (2016). Intergenerational consequences of migration: Socio-economic, family and cultural patterns of stability and change in Turkey and Europe. London: Palgrave Macmillan. Immerzeel, T. & Van Tubergen, F. (2013). Religion as reassurance? Testing the insecurity theory in 26 European countries. European Sociological Review, 29, 2, 359-372. https://doi.org/10.1093/esr/jcr072 Inglehart, R. (1997). Modernization and postmodernization: Cultural, economic, and political change in 43 societies. Princeton: Princeton University Press. Inglehart, R. & Norris, P. (2003). Rising tide: Gender equality and cultural change around the world. Cambridge: Cambridge University Press. Kandiyoti, D. (1988). Bargaining with patriarchy. Gender and Society, 2, 3, 274-89. https://doi.org/10.1177/089124388002003004 Kandiyoti, D. (1988). Bargaining with patriarchy. Gender an https://doi.org/10.1177/089124388002003004 Khattab, N. & Hussein, S. (2018). Can religious affiliation explain the disadvantage of Muslim women in the British labour market? Work, Employment and Society, 32, 6, 1011-1028. https://doi.org/10.1177/0950017017711099 Khattab, N., Johnston, R. & Manley, D. (2018). Human capital, family structure and religiosity shaping British Muslim women’s labour market participation. Journal of Ethnic and Migration Studies, 44, 9, 1541-1559. https://doi.org/10.1080/1369183X.2017.1334541 Khoudja, Y. & Fleischmann, F. (2015). References Ethnic differences in female labour force participation in the Netherlands: Adding gender role attitudes and religiosity to the explanation. European Sociological Review, 31, 1, 91-102. https://doi.org/10.1093/esr/jcu084 Khoudja, Y. & Platt, L. (2018). Labour market entries and exits of women from different origin countries in the UK. Social Science Research, 69, 1-18. https://doi.org/10.1016/j.ssresearch.2017.10.003 Koenig, M., Maliepaard, M. & Guveli, A. (2016). Religion and new immigrants’ labor market entry in Western Europe. Ethnicities, 16, 2, 213-235. https://doi.org/10.1177/1468796815616159 Koopmans, R. (2016). Does assimilation work? Sociocultural determinants of labour market participation of European Muslims. Journal of Ethnic and Migration Studies, 42, 2, 197-216. https://doi.org/10.1080/1369183X.2015.1082903 Lehrer, E.L. (1995). The effects of religion on the labor supply of married women. Social Science Research, 24, 3, 281-301. https://doi.org/10.1006/ssre.1995.1011 Lerner, D. (1958). The passing of traditional society: Modernizing the Middle East. Glencoe (Illinois): Free Press Lerner, D. (1958). The passing of traditional society: Modernizing the Middle East. Glencoe (Illinois): Free Press 428 Lisaliner, F.G. & Bhatti, F. (2005). Determinants of female labour force participation: A study of North Cyprus. Review of Social, Economic & Business Studies, 5, 6, 209-226. https://www.researchgate.net/profile/Fatma_Guven_Lisaniler/publication/262005349 _Determinant_of_Labor_Force_Participation_A_Case_of_Cyprus/links/02e7e5363ee6 d89956000000.pdf Miyata, S. & Hiroyuki, Y. (2016). Do female gender role attitudes affect labour market participation in Egypt? The Journal of Development Studies, 52, 6, 876-894. https://doi.org/10.1080/00220388.2015.1113262 Moghadam, V.M. (1996). Patriarchy and economic development. Women’s positions at the end of the twentieth century. Oxford: Clarendon Press. Moghadam, V.M. (2013). What is democracy? Promises and perils of the Arab Spring. Current Sociology, 61, 4, 393-408. https://doi.org/10.1177/0011392113479739. p // g/ / Naseema, J. & Adnan, W. (2019). Being a second-generation Muslim woman in the French labour market. Understanding the dynamics of (visibility of) religion and gender in labour market access, outcomes and experiences. Research in Social Stratification and Mobility, 61, 79-93. https://doi.org/10.1016/j.rssm.2019.02.003 p // g/ /j Norris, P. & Inglehart, R. (2011). Sacred and secular: Religion and politics worldwide. Cambridge: Cambridge University Press. g g y Pampel, F.C. & Tanaka, K. (1986). Economic development and female labor force participation: A reconsideration. Social Forces, 64, 3, 599-619. https://doi.org/10.1093/sf/64.3.599 Pastore, F. & Tenaglia S. (2013). Ora et no labora? A test of the impact of religion on female labour supply. IZA Discussion Paper Series No. 7356. https://ssrn.com/abstract=2260671 Pettit, B. & Hook, J. (2005). The structure of women's employment in comparative perspective. Social Forces, 84, 2, 779-801. https://doi.org/10.1353/sof.2006.0029 Pfau-Effinger, B. (2017). Development of culture, welfare states and women’s employment in Europe. Lisaliner, F.G. & Bhatti, F. (2005). Determinants of female labour force participation: A study of North Cyprus. Review of Social, Economic & Business Studies, 5, 6, 209-226. https://www.researchgate.net/profile/Fatma_Guven_Lisaniler/publication/262005349 _Determinant_of_Labor_Force_Participation_A_Case_of_Cyprus/links/02e7e5363ee6 d89956000000.pdf References London & New York: Routledge. Price, A. (2015). How national structures shape attitudes toward women’s right to employment in the Middle East. International Journal of Comparative Sociology, 56, 6, 408-32. https://doi.org/10.1177/0020715215625494 Richards, D.L. & Gelleny, R. (2007). Women’s status and economic globalization. International Studies Quarterly, 51, 4, 855-876. https://doi.org/10.1111/j.1468-2478.2007.00480.x Röder, A. (2014). Explaining religious differences in immigrants’ gender role attitudes: The changing impact of origin country and individual religiosity. Ethnic and Racial Studies 37, 14, 2615-2635. https://doi.org/10.1080/01419870.2013.854919 Ruiter, S. & Van Tubergen, F. (2009). Religious attendance in cross-national perspective: A multilevel analysis of 60 countries. American Journal of Sociology, 115, 3, 863-95. https://doi.org/10.1086/603536 429 Schnabel, L. (2016). Religion and gender equality worldwide: A country-level analysis. Social Indicators Research, 129, 2, 893-907. https://doi.org/10.1007/s11205-015-1147-7 p // g/ / Schwartz, S.H. & Galit, S. (2000). Value consensus and importance: A cross-national study. Journal of Cross-Cultural Psychology, 31, 4, 465-97. https://10.1177/0022022100031004003 Schwartz, S.H. & Galit, S. (2000). Value consensus and importance: A cross-national study. Journal of Cross-Cultural Psychology, 31, 4, 465-97. https://10.1177/0022022100031004003 p // / Spierings, N. (2014a). The influence of patriarchal norms, institutions, and household composition on women's employment in twenty-eight Muslim-majority countries. Feminist Economics, 20, 4, 87-112. https://doi.org/10.1080/13545701.2014.963136 Spierings, N. (2014a). The influence of patriarchal norms, institutions, and household composition on women's employment in twenty-eight Muslim-majority countries. Feminist Economics, 20, 4, 87-112. https://doi.org/10.1080/13545701.2014.963136 Spierings, N. (2014b). How Islam influences women’s paid non-farm employment: Evidence from 26 Indonesian and 37 Nigerian provinces. Review of Religious Research, 56, 3, 399-431. https://www.jstor.org/stable/43186236 Spierings, N. (2015). Gender equality attitudes among Turks in Western Europe and Turkey: The interrelated impact of migration and parents’ attitudes. Journal of Ethnic and Migration Studies, 41, 5, 749-771. https://doi.org/10.1080/1369183X.2014.948394 Spierings, N. (2015). Gender equality attitudes among Turks in Western Europe and Turkey: The interrelated impact of migration and parents’ attitudes. Journal of Ethnic and Migration Studies, 41, 5, 749-771. https://doi.org/10.1080/1369183X.2014.948394 p // g/ / Spierings, N. (2016). Women’s employment in Muslim countries: Patterns of diversity. London: Palgrave Macmillan. g Spierings, N. (2018). Popular opposition to economic gender equality and homosexual lifestyles. In: Verloo, M. (Eds.), Varieties of opposition to gender equality in Europe, London & New York: Routledge, pp. 172-194. Spierings, N. (2019). The multidimensional impact of Islamic religiosity on ethno- religious social tolerance in the Middle East and North Africa. Social Forces, 97, 4, 1693-1730. https://doi.org/10.1093/sf/soy092 Spierings, N., Smits, J. & Verloo, M. (2010). References Micro‐and macrolevel determinants of women's employment in six Arab countries. Journal of Marriage and Family, 72, 5, 1391-1407. https://doi:10.1111/j.1741-3737.2010.00772.x https://doi:10.1111/j.1741-3737.2010.00772.x Steiber, N. & Haas, B. (2012). Advances in explaining women’s employment patterns. Socio-Economic Review, 10, 2, 343-367. https://doi.org/10.1093/ser/mwr039 Te Grotenhuis, M., Scholte, M., de Graaf, N.D. & Pelzer, B. (2015). The between and within effects of social security on church attendance in Europe 1980–1998: The danger of testing hypotheses cross-nationally. European Sociological Review, 31, 5, 643- 654. https://doi.org/10.1093/esr/jcv069 p // g/ / /j Van der Lippe, T. & Van Dijk, L. (2002). Comparative research on women’s employment. Annual Review of Sociology, 28, 1, 221-241. https://doi.org/10.1146/annurev.soc.28.110601.140833 430 Van Klingeren, M., & Spierings, N. (2020). Acculturation, decoupling, or both? Migration’s impact on the linkage between religiosity and gender equality attitudes. Journal of Ethnic and Migration Studies, 46, 15, 3079-3100. https://doi.org/10.1080/1369183X.2020.1733947 g Van Tubergen, F. & Sindradottir, J.I. (2011). The religiosity of immigrants in Europe: A cross‐national study. Journal for the Scientific Study of Religion, 50, 2, 272-288. https://doi.org/10.1111/j.1468-5906.2011.01567 g j Walby, S. (2009). Globalization & inequalities. Complexities and contested modernities. London, Thousand Oaks, CA: Sage. Zuccotti, CV., Ganzeboom, H.B.G. & Guveli, A. (2017). Has migration been beneficial for migrants and their children? International Migration Review, 51, 1, 97-126. https://doi.org/10.1111/imre.12219 431 Appendix Table A.1: Unweighted descriptive statistics (N = 43649) Mean/Proportion Freq. SD Min. Max. Years of full-time education 13.18 3.69 0.00 48.0 Age 44.61 12.2 18.00 65.0 Doing paid work last 7 days 0.64 Country BE Belgium 0.06 2390 CH Switzerland 0.06 2386 CZ Chech 0.08 3255 DE Germany 0.09 3635 DK Denmark 0.04 1472 ES Spain 0.06 2526 FR France 0.05 1890 GB UK 0.07 2670 GR Greece 0.07 2692 IE Ireland 0.09 3344 IL Israel 0.06 2427 NL Netherlands 0.07 2567 NO Norway 0.05 1922 RU Russia 0.06 2520 SE Sweden 0.05 1995 SI Slovenia 0.04 1542 Children living at home 0.54 Marital status Married 0.62 24218 Divorced/widowed 0.23 8911 Never married 0.16 6104 Migration status First-generation migrants 0.11 4278 Second generation 0.09 3567 Natives 0.80 31388 Religion No religion 0.41 15945 Roman catholic 0.25 9970 Protestant 0.14 5390 Eastern orthodox 0.10 4035 Jewish 0.05 2015 Islam 0.03 1159 Other religions 0.02 719 Survey year (ESS round) 2 – 2004 0.23 9093 4 – 2008 0.23 9018 5 – 2010 0.30 11610 8 - 2016 0.24 9512 432 Table A.1: Unweighted descriptive statistics (N = 43649) (continued) Mean/Proportion Freq. References SD Min. Max. References Pray (How often pray apart from at religious services) 0 never 0.37 14537 1 less often 0.16 6370 2 only on special holidays 0.05 1781 3 at least once month 0.06 2386 4 once a week 0.06 2472 5 more than once a week 0.09 3436 6 every day 0.21 8251 Gender attitudes (Men should have more right to job than women when jobs are scarce) 1 disagree strongly 0.40 15864 2 disagree 0.32 12724 3 neither agree nor disagree 0.12 4661 4 agree 0.11 4366 5 agree strongly 0.04 1618 433 Table A.2: Odd ratios role of religiosity and gender role attitudes on employment likelihood of women in different religion over time (Model 1, 2, 3 and 4 are in Table 4) Model 5 Model 6 Model 7 Country (Ref: Austria) CH Switzerland 2.156* 2.155* 2.155*** (0.1506) (0.1506) (0.1513) CZ Czech 1.085 1.082 1.071 (0.0669) (0.0667) (0.0664) DE Germany 1.152* 1.148* 1.146* (0.0708) (0.0706) (0.0707) DK Denmark 1.773* 1.752* 1.765* (0.1531) (0.1516) (0.1557) ES Spain 0.901 0.900 0.895+ (0.0600) (0.0599) (0.0597) FR France 1.132+ 1.132+ 1.129+ (0.0811) (0.0811) (0.0811) GB UK 1.048 1.045 1.036 (0.0688) (0.0686) (0.0682) GR Greece 0.728* 0.730* 0.701* (0.0659) (0.0660) (0.0674) IE Ireland 0.587* 0.587* 0.571*** (0.0367) (0.0368) (0.0363) IL Israel 1.060 1.046 1.057 (0.1337) (0.1321) (0.1367) NL Netherlands 1.149* 1.147* 1.150* (0.0759) (0.0758) (0.0762) NO Norway 2.207* 2.200* 2.223* (0.1786) (0.1781) (0.1810) RU Russia 1.376* 1.360* 1.360* (0.1063) (0.1054) (0.1072) SE Sweden 2.755* 2.743* 2.765* (0.2183) (0.2175) (0.2201) SI Slovenia 1.066 1.062 1.072 (0.0795) (0.0793) (0.0802) Children living at home 0.658* 0.659* 0.658* (0.0188) (0.0188) (0.0189) Marital status (Ref: Married) Divorced/widowed 1.095 1.093 1.097 (0.0327) (0.0326) (0.0329) Never married 1.011 1.014 1.018 (0.0401) (0.0402) (0.0405) Years of full-time education completed 1.113* 1.113* 1.114* (0.0041) (0.0041) (0.0042) Age 1.420* 1.420* 1.422* (0.0112) (0.0112) (0.0112) Age2 0.996* 0.996* 0.996* (0.0001) (0.0001) (0.0001) Table A.2: Odd ratios role of religiosity and gender role attitudes on employment likelihood of women in different religion over time (Model 1, 2, 3 and 4 are in Table 4) 434 Table A.2: Odd ratios role of religiosity and gender role attitudes on employment likelihood of women in different religion over time (Model 1, 2, 3 and 4 are in Table 4) (continued) Model 5 Model 6 Model 7 Migration status (Ref: First generation migrant) Second-generation 1.099+ 1.100+ 1.100+ (0.0588) (0.0588) (0.0592) Natives 1.297* 1.298* 1.298* (0.0522) (0.0522) (0.0527) Religion (Ref: No religion) Roman Catholic 1.125 1.126 1.130 (0.0412) (0.0412) (0.2022) Protestant 1.203* 1.204*** 1.432 (0.0537) (0.0537) (0.3267) Eastern Orthodox 0.971 0.971 1.436 (0.0694) (0.0694) (0.4623) Jewish 1.357* 1.364* 1.306 (0.1803) (0.1811) (0.5639) Islam 0.583* 0.581* 2.908+ (0.0485) (0.0482) (1.7305) Other religions 0.931 0.930 1.026 (0.0844) (0.0843) (0.6163) Praying (1-5) 0.959* 0.951* 0.943* (0.0121) (0.0057) (0.0246) Survey year (ESS round) 1.050*** 1.025* 1.036+ (0.0089) (0.0124) (0.0194) PraySurvey year 0.998 0.999 (0.0023) (0.0050) Roman CatholicPray 1.068+ (0.0381) ProtestantPray 1.035 (0.0446) Eastern OrthodoxPray 0.933 (0.0535) JewishPray 0.860 (0.0661) IslamPray 0.912 (0.0828) Other religionsPray 1.004 (0.1020) Roman CatholicSurvey year 1.017 (0.0336) ProtestantSurvey year 0.975 (0.0433) Eastern OrthodoxSurvey year 0.943 (0.0584) JewishSurvey year 1.004 (0.0697) IslamSurvey year 0.808 (0.0774) Other religionsSurvey year 0.948 (0.1031) Odd ratios role of religiosity and gender role attitudes on employment likelihood of women in different religion over time (Model 1, 2, 3 and 4 are in Table 4) (continued) 435 Table A.2: Odd ratios role of religiosity and gender role attitudes on employment likelihood of women in different religion over time (Model 1, 2, 3 and 4 are in Table 4) (continued) Model 5 Model 6 Model 7 Roman CatholicPraySurvey year 0.992 (0.0068) ProtestantPraySurvey year 0.992 (0.0086) Eastern OrthodoxPraySurvey year 1.021+ (0.0117) JewishpraySurvey year 1.026+ (0.0137) IslampraySurvey year 1.013 (0.0155) Other religionspraySurvey year 1.008 (0.0188) Traditional gender attitudes (0-6) 0.846 0.809* 0.842*** (0.0094) (0.0208) (0.0346) Traditional gender attitudesSurvey year 1.010+ 1.008 (0.0051) (0.0083) Roman CatholicTraditional gender attitudes 0.967 (0.0601) ProtestantTraditional gender attitudes 0.865 (0.0780) Eastern OrthodoxTraditional gender attitudes 1.027 (0.0918) JewishTraditional gender attitudes 0.965 (0.1464) IslamTraditional gender attitudes 0.640** (0.1062) Other religionsTraditional gender attitudes 0.807 (0.1660) Roman CatholicTraditional gender attitudesSurvey year 0.991 (0.0126) ProtestantTraditional gender attitudesSurvey year 1.031 (0.0195) Eastern OrthodoxTraditional gender attitudesSurvey year 0.979 (0.0178) JewishTraditional gender attitudesSurvey year 1.009 (0.0270) IslamTraditional gender attitudesSurvey year 1.061 (0.0294) Other religionsTraditional gender attitudesSurvey year 1.047 (0.0419) N 39,233 39,233 39,233 Notes: Robust standard error in parentheses; * p<0.001, p<0.01, * p<0.05, + p<0.1 436 Table A.3: Odd ratios for role of religiosity and gender role attitudes on employment likelihood of women belonging to different religions (weighted for population size) 1 2 3 4 Survey years (ESS round) 1.069* 1.068* 1.052* 1.052* (0.0098) (0.0098) (0.0098) (0.0098) Religions (ref: no Religion) Roman Catholic 1.070 1.180 1.089+ 1.185* (0.0496) (0.0601) (0.0508) (0.0605) Protestant 1.026 1.120+ 1.045 1.128* (0.0581) (0.0667) (0.0594) (0.0675) Eastern Orthodox 0.903 0.988 0.905 0.979 (0.0754) (0.0851) (0.0757) (0.0845) Jewish 1.121 1.184 1.081 1.136 (0.2736) (0.2887) (0.2607) (0.2738) Islam 0.547* 0.615* 0.582* 0.644* (0.0581) (0.0674) (0.0618) (0.0705) Other religions 0.894 1.020 0.907 1.018 (0.1131) (0.1336) (0.1151) (0.1334) Praying (1-5) 0.960* 0.965* (0.0086) (0.0086) Traditional gender attitudes (0-6) 0.852* 0.855* (0.0140) (0.0141) Observations 39,233 39,233 39,233 39,233 Notes: All models are controlled for age, age2, education, marital status, having children at home and migration status. References All models include country fixed effects. * p<0.001, p<0.01, * p<0.05, + p<0.1. Robust s.e. in parentheses Notes: All models are controlled for age, age2, education, marital status, having children at home and migration status. All models include country fixed effects. * p<0.001, p<0.01, * p<0.05, + p<0.1. Robust s.e. in parentheses 437 Deutscher Titel Veränderungen des Einflusses von Religiosität und Geschlechterrollenvorstellungen auf das Erwerbsverhalten von Frauen nach Religionszugehörigkeit in Europa zwischen 2004 und 2016 Schlagwörter: Erwerbstätigkeit von Frauen, Religionszugehörigkeit, Geschlechterrollenvorstellungen, Europa Zusammenfassung Fragestellung: Dieser Beitrag untersucht den Zusammenhang von Religiosität, Geschlechterrollenvorstellungen und Erwerbsverhalten von Frauen in Europa zwischen 2004 und 2016. Hintergrund: Religiosität und traditionelle Geschlechterrollenvorstellungen gelten als wesentliche Determinanten, die das Erwerbsverhalten von Frauen bestimmen. Dieser Beitrag argumentiert, dass diese beiden Faktoren differenziert betrachtet werden müssen, da der Einfluss, den Religiosität und Geschlechterrollenvorstellungen auf das Erwerbsverhalten ausübt, auf andere Mechanismen zurückgeführt werden muss. Methode: Es werden die Daten des European Social Survey (ESS) aus den Jahren 2004, 2008, 2010 und 2016 (Welle 2, 4, 8, 10) verwendet, in denen u.a. Informationen zur Religiosität, Religionszugehörigkeit und zu den Geschlechterrollenvorstellungen für 16 Länder vorliegen (N=39.233). Ergebnisse: Nach Kontrolle von Religiosität vergrößern sich die Unterschiede zwischen Personen ohne Religionszugehörigkeit und Personen mit römisch-katholischer, protestantischer oder jüdischer Religionszugehörigkeit. Hingegen reduzieren sich die Unterschiede zwischen Frauen mit islamischen und orthodoxem Glaubens in Vergleich zu säkularisierten Frauen. Die Hinzunahme von Geschlechterrollenvorstellungen erklärt die Erwerbsunterschiede nur marginal. Schlussfolgerung: Dieser Beitrag unterstützt die Idee, dass Religiosität und traditionelle Geschlechterrollenvorstellungen durch verschiedene Mechanismen mit dem Erwerbsverhalten von Frauen verbunden sind. Über die Zeit ist die Erwerbsneigung für alle Frauen, unabhängig von der konfessionellen Zugehörigkeit gestiegen, abgesehen von islamischen Frauen, wo wir einen Rückgang beobachten können. Schlagwörter: Erwerbstätigkeit von Frauen, Religionszugehörigkeit, Geschlechterrollenvorstellungen Europa Schlagwörter: Erwerbstätigkeit von Frauen, Religionszugehörigkeit, Geschlechterrollenvorstellungen, Europa Geschlechterrollenvorstellungen, Europa 438 JFR – Journal of Family Research, 2021, vol. 33, no. 2, pp. 405–438. doi: 10.20377/jfr-554 Submitted: August 23, 2020 Accepted: February 3, 2021 Published online: February 19, 2021 Ayse Guveli: https://orcid.org/0000-0003-3679-0238 Niels Spierings: https://orcid.org/0000-0002-3116-3262 This work is licensed under a Creative Commons Attribution 4.0 International License.
https://openalex.org/W2921832314	https://www.epj-conferences.org/articles/epjconf/pdf/2019/07/epjconf_charm2018_02001.pdf	Latin	null	Status of charmed meson spectroscopy	EPJ web of conferences	2,019	cc-by	5,625	EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 https://doi.org/10.1051/epjconf/201920202001 © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). ∗e-mail: fkguo@itp.ac.cn Status of charmed meson spectroscopy Feng-Kun Guo1,2,∗ 1CAS Key Laboratory of Theoretical Physics, Institute of Theoretical Physics, Chinese Academy of Sciences, Beijing 100190, China 2School of Physical Sciences University of Chinese Academy of Sciences Beijing 100049 China Feng-Kun Guo 1CAS Key Laboratory of Theoretical Physics, Institute of Theoretical Physics, Chinese Academy of Sciences, Beijing 100190, China 2 2School of Physical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China Abstract. The discovery of the ground state positive-parity charm-strange and charm-nonstrange mesons D∗ s0(2317), Ds1(2460), D∗ 0(2400) and D1(2430) in 2003 and 2004 brought up several mysteries related to their masses. Here I brieﬂy review recent progresses from lattice calculations and analysis of the precise LHCb measurements of the B−→D+π−π−in the framework of uni- tarized chiral perturbation theory. It turns out that all the mysteries can be understood in a picture consistent with both lattice results and the LHCb mea- surements. In this picture, the main components of D∗ s0(2317) and Ds1(2460) are DK and D∗K hadronic molecules, respectively. Furthermore, the resonance parameters of the ground state 0+ and 1+ charm-nonstrange mesons take values very diﬀerent from the known ones of the D∗ 0(2400) and D1(2430), which were obtained by using an improper resonance parameterization. It is pointed out that there should be two D∗ 0 and two D1 broad states in region relevant to the D∗ 0(2400) and D1(2430). Suggestions towards identifying the higher nonstrange resonances are given. 1 Introduction 1800 2000 2200 2400 2600 2800 3000 Charm-strange mesons Mass (MeV) 0- 1- 0+ 1+ 2+ 3- ?? Ds1(2460) Ds0 * (2317) DK threshold DK threshold Godfrey, Isgur, PRD32(1985)189 mesons discovered before 2003 mesons discovered after 2003 1800 2000 2200 2400 2600 2800 3000 Charm-nonstrange mesons Mass (MeV) 0- 1- 0+ 1+ 2+ 3- ?? D0 (2400) D1(2430) Godfrey, Isgur, PRD32(1985)189 mesons discovered before 2003 mesons discovered after 2003 Figure 1. The charm-strange and charm-nonstrange meson mass spectra. The predictions in the GI quark model [14] are shown as in black lines, while the mesons observed before and since 2003 are shown as blue and red lines, respectively. For the observed mesons, the masses given in the RPP [13] are taken, and the bands correspond to their uncertainties. 1800 2000 2200 2400 2600 2800 3000 Charm-strange mesons Mass (MeV) 0- 1- 0+ 1+ 2+ 3- ?? Ds1(2460) Ds0 * (2317) DK threshold DK threshold Godfrey, Isgur, PRD32(1985)189 mesons discovered before 2003 mesons discovered after 2003 1800 2000 2200 2400 2600 2800 3000 Charm-nonstrange mesons Mass (MeV) 0- 1- 0+ 1+ 2+ 3- ?? D0 (2400) D1(2430) Godfrey, Isgur, PRD32(1985)189 mesons discovered before 2003 mesons discovered after 2003 Figure 1. The charm-strange and charm-nonstrange meson mass spectra. The predictions in the GI quark model [14] are shown as in black lines, while the mesons observed before and since 2003 are shown as blue and red lines, respectively. For the observed mesons, the masses given in the RPP [13] are taken, and the bands correspond to their uncertainties. Shortly after these discoveries, the lowest scalar and axial-vector charm-nonstrange mesons D∗ 0(2400)0 and D1(2430)0 were reported by the Belle Collaboration [32]. They show up as broad bumps in the D(∗)π invariant mass distributions, and each of them was ﬁtted with a single Breit-Wigner resonance. The D∗ 0(2400)0 was later on also reported by the FOCUS, BaBar and LHCb Collaborations. While all experiments reported a large width, with values consistent with each another, the central values of the mass diﬀer largely: (2308 ± 36) MeV (Belle) [32], (2407 ± 41) MeV (FOCUS) [33], and (2297 ± 22) MeV (BaBar) [34]. The av- eraged values is (2318 ± 27) MeV in the Review of Particle Physics (RPP) [13]. The charged partner D∗ 0(2400)± was reported by the FOCUS [33] and LHCb [35, 36] Collaborations, and the averaged mass is (2351 ± 7) MeV [13]. 1 Introduction The discovery of the D∗ s0(2317) with JP = 0+ by the BaBar Collaboration in 2003 [1] marked the start of a new era in the study of hadron spectroscopy. After that, tens of hadronic resonant structures were reported in experiments with properties at odds with the quark model which describes mesons as bound states of a quark and an antiquark. Many of these new hadrons are then natural candidates of exotic hadrons. For recent reviews of the experimental observations and the eﬀorts devoted to understanding them, see Refs. [2–11]. Here I will discuss the positive-parity charmed mesons, i.e., D∗ s0(2317), Ds1(2460), D∗ 0(2400) and D1(2430). s0 0 The discovery of the D∗ s0(2317) was closely followed by that of the axial-vector meson Ds1(2460) by the CLEO Collaboration [12]. The masses of these two charm-strange mesons, (2317.7 ± 0.6) MeV and (2459.5 ± 0.6) MeV, respectively [13], are much lower than the values predicted in potential quark models such as the Godfrey-Isgur (GI) quark model [14] (the updated results for charmed mesons can be found in [15]), which was very successful in describing the observed hadron spectroscopy till then. No isospin partners were found. Their widths are tiny, and so far only upper limits were reported [13]. So these two charm-strange mesons are isospin-scalar states. As they are located below the DK and D∗K thresholds, respectively, the only allowed strong decays are into Dsπ and D∗ sπ for the D∗ s0(2317) and EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 https://doi.org/10.1051/epjconf/201920202001 Ds1(2460), respectively, breaking isospin symmetry. These two states spurred lots of interest because of the unexpected low masses. Attempts were made trying to explain them as c¯s mesons [16–20], chiral partners of the ground state Ds and D∗ s mesons [21, 22], compact [cq][¯s¯q] tetraquark states [23], mixing of the c¯s and tetraquarks [24], the Dπ atom for the D∗ s0(2317) [25], and the D(∗)K hadronic molecules [26–31]. Ds1(2460), respectively, breaking isospin symmetry. These two states spurred lots of interest because of the unexpected low masses. Attempts were made trying to explain them as c¯s mesons [16–20], chiral partners of the ground state Ds and D∗ s mesons [21, 22], compact [cq][¯s¯q] tetraquark states [23], mixing of the c¯s and tetraquarks [24], the Dπ atom for the D∗ s0(2317) [25], and the D(∗)K hadronic molecules [26–31]. 2 Lattice QCD studies of the positive-parity charmed mesons The calculation of the positive-parity charmed meson spectrum on lattice is diﬃcult as their masses are close to the strongly-coupled S -wave D(∗)K thresholds, and thus it is not easy to get the correct masses without taking the D(∗)K into account. The inclusion of the latter, how- ever, requires the calculation of quark-antiquark annihilation-type Wick contractions, which remained a challanging problem for a long time. Attempts of calculating the ground state scalar charm-strange meson were made shortly after the D∗ s0(2317) discovery. Interpolating operators of the c¯s type were used, and the obtained masses [40, 41] were larger than the observed value for the D∗ s0(2317). The ﬁrst calculation using interpolating operators of both the c¯s and DK (constructed with two quark ﬁelds and two antiquark ﬁelds) types was made in [42, 43]. The obtained value for the mass diﬀerence MD∗ s0 −(MDs + 3MD∗s)/3, (266 ± 16) MeV, was consistent with the experimental value (241.5 ± 0.8) MeV. A more complete lattice calculation on the charm-strange meson spectrum was made in Ref. [44], where the masses of the 0+ and 1+ charm-strange mesons were 2348(4) MeV and 2451(4) MeV, respectively, at a pion mass of 150 MeV. These values are close to the measured masses of the D∗ s0(2317) and Ds1(2460). The masses of the ground state 0−, 1−, 0+ and 1+ charm-strange mesons obtained in Ref. [44] using two diﬀerent pion masses are shown in Table 1. It is clear that the pion-mass dependence of the 0+ and 1+ states is much stronger than that of the 0−and 1−ones, indicating the important role of up and down quarks in the charm-strange mesons. The charmed-meson–light-meson scattering lengths in a few channels (Dπ with I = 3/2, D ¯K with I = 0 and 1, DsK and Dsπ) free of quark-annihilation contractions were ﬁrst calcu- lated on lattice in Ref. [45] (updated in Ref. [46]). These results provide valuable information on the low-energy interaction between a charmed meson and a light pseudoscalar meson, and thus leads to important insights into the nature of the scalar charmed mesons, to be discussed in the next section. As for the non-strange sector, the Dπ I = 1/2 S -wave cattering length and the lowest scalar D∗ 0 meson were calculated using the Lüscher’s method in Ref. [47] using the c¯s + Dπ- type interpolating operators. 1 Introduction Mπ = 150 MeV Mπ = 290 MeV RPP MDs 1977 ± 1 1980 ± 1 1968.34 ± 0.07 MD∗s 2094 ± 1 2101 ± 1 2112.2 ± 0.4 MD∗ s0 2348 ± 4 2384 ± 3 2317.7 ± 0.6 MDs1 2351 ± 4 2397 ± 4 2459.5 ± 0.6 Although their bottom cousins have not been discovered so far, it is natural to ask whether such mysteries will repeat there. Next I will discuss the progress on understanding these mysteries using lattice quantum chromodynamics (QCD) and eﬀective ﬁeld theory calculations. 1 Introduction One should not regard the diﬀerence between the neutral and charged RPP masses as a large isospin splitting. The mass spectra of the charm- strange and charm-nonstrange mesons are shown in Fig. 1, where the predictions in the GI quark model [14] are also shown. These observations led to three mysteries regarding the masses of the positive-parity charmed mesons: (1) Mass problem: Why are the D∗ s0(2317) and Ds1(2460) much lighter than the quark model expectations for the lowest scalar and axial-vector c¯s mesons? (2) Fine-tuning problem: Why is MDs1(2460) −MD∗ s0(2317) ≃MD∗± −MD± within 2 MeV? Notice that the positive-parity states and the negative-parity ones are in diﬀerent spin multiplets. If they are not connected by any symmetry, it would be mysterious why the hyperﬁne splittings should be ﬁne tuned to be almost equal. In Refs. [21, 22], this equality is obtained as a consequence of the chiral-partner assumption. From Fig. 1, one sees that the splittings are very diﬀerent in the GI quark model. (3) Hierachy problem: Why are MD∗ 0(2400) ≳MD∗ s0(2317) and MD1(2430) ∼MDs1(2460)? Such relationships were used as inputs in many works, see e.g. [19, 37–39]. But the natural mass hierachy within a given SU(3) ﬂavor multiplet should be that the hadron with a strange quark is at least 100 MeV heavier than the nonstrange ones, as a consequence of the strange quark being much heavier than the up and down quarks. 2 EPJ Web of Conferences 202, 02001 (2019) https://doi.org/10.1051/epjconf/201920202001 EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 Table 1. Masses of the ground state 0+ and 0−charm-strange mesons in the lattice calculation using two diﬀerent pion masses in Ref. [44]. The experimental values [13] are given in the last column. All values are in units of MeV. Table 1. Masses of the ground state 0+ and 0−charm-strange mesons in the lattice calculation using two diﬀerent pion masses in Ref. [44]. The experimental values [13] are given in the last column. All values are in units of MeV. 2 Lattice QCD studies of the positive-parity charmed mesons Here M and Γ are the mass and the total decay width, respectively. For comparison, the RPP values [13] and latest lattice QCD results are also shown. All values are in units of MeV Prediction RPP Lattice QCD D∗ s0 2315+18 −28 2317.7 ± 0.6 2348+7 −4 [44] Ds1 2456+15 −21 2459.5 ± 0.6 2451 ± 4 [44] B∗ s0 5720+16 −23 − 5711 ± 23 [64] Bs1 5772+15 −21 − 5750 ± 25 [64] D∗ 0 2105+6 −8, 102+10 −11 , 2451+35 −26, 134+7 −8 (2318 ± 29, 134 ± 20) − D1 2247+5 −6, 107+11 −10 , 2555+47 −30, 203+8 −9 2427 ± 40, 192+65 −55 − B∗ 0 5535+9 −11, 113+15 −17 , 5852+16 −19, 36 ± 5 − − B1 5584+9 −11, 119+14 −17 , 5912+15 −18, 42+5 −4 − − interpolating ﬁelds are employed, and the computed energy levels cover the energy region up to above the Ds ¯K threshold. The obtained energy levels were analyzed using a coupled- channel K-matrix formalism. Many K-matrix parameterizations were used to ﬁt the energy levels. The obtained K-matrix were then substituted into the inﬁnite-volume T-matrix, which has poles corresponding to hadronic resonances. In this way, a pole at (2275.9 ± 0.9) MeV, slightly below the Dπ threshold at (2276.4 ± 0.9) MeV, were found for Mπ ≃391 MeV. It is worthwhile to notice that this mass is lower than the D∗ 0(2400) in RPP [13], even though the pion mass takes a value much higher than its physical value. 2 Lattice QCD studies of the positive-parity charmed mesons With a pion mass of about 266 MeV, the authors got MD∗ 0 − (MD + 3MD∗)/3 = (351 ± 21) MeV and MD1 −(MD + 3MD∗)/3 = (380 ± 21) MeV, which may be confronted with (347 ± 29) MeV and (456 ± 40) MeV, respectively, using the masses of the D∗ 0(2400) and D1(2430) in the RPP [13]. 0 The most updated and sophisticated lattice calculation of the I = 1/2 D∗ 0 is presented in Ref. [48] using the coupled-channel Lüscher’s method. The c¯s + Dπ + Dη + Ds ¯K-type 3 EPJ Web of Conferences 202, 02001 (2019) https://doi.org/10.1051/epjconf/201920202001 EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 Table 2. Predicted masses of the 0+ and 1+ heavy-strange mesons, and poles given as (M, Γ/2) for the 0+ and 1+ heavy-nonstrange mesons. Here M and Γ are the mass and the total decay width, respectively. For comparison, the RPP values [13] and latest lattice QCD results are also shown. All values are in units of MeV. Prediction RPP Lattice QCD D∗ s0 2315+18 −28 2317.7 ± 0.6 2348+7 −4 [44] Ds1 2456+15 −21 2459.5 ± 0.6 2451 ± 4 [44] B∗ s0 5720+16 −23 − 5711 ± 23 [64] Bs1 5772+15 −21 − 5750 ± 25 [64] D∗ 0 2105+6 −8, 102+10 −11 , 2451+35 −26, 134+7 −8 (2318 ± 29, 134 ± 20) − D1 2247+5 −6, 107+11 −10 , 2555+47 −30, 203+8 −9 2427 ± 40, 192+65 −55 − B∗ 0 5535+9 −11, 113+15 −17 , 5852+16 −19, 36 ± 5 − − B1 5584+9 −11, 119+14 −17 , 5912+15 −18, 42+5 −4 − − Table 2. Predicted masses of the 0+ and 1+ heavy-strange mesons, and poles given as (M, Γ/2) for the 0+ and 1+ heavy-nonstrange mesons. Here M and Γ are the mass and the total decay width, respectively. For comparison, the RPP values [13] and latest lattice QCD results are also shown. All values are in units of MeV. Table 2. Predicted masses of the 0+ and 1+ heavy-strange mesons, and poles given as (M, Γ/2) for Table 2. Predicted masses of the 0+ and 1+ heavy-strange mesons, and poles given as (M, Γ/2) for the 0+ and 1+ heavy-nonstrange mesons. 4 Confronting two D∗ 0 states with the LHCb data 4 Confronting two D∗ 0 states with the LHCb data ×106 ●●●●●●●●●●●●●●●●●●●●● ● ●● ● ● ● 2.0 2.1 2.2 2.3 2.4 2.5 0.0 0.5 1.0 1.5 MD+ π- [GeV] 〈P0〉/(20 MeV) ×106 ● ●● ● ●● ●●● ●●● ●● ●● ●●●● ● ●● ● ●●● 2.0 2.1 2.2 2.3 2.4 2.5 -0.05 0.00 0.05 0.10 0.15 MD+ π- [GeV] (〈P1〉-14〈P3〉/9) /(20 MeV) ×106 ●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ●● 2.0 2.1 2.2 2.3 2.4 2.5 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 MD+ π- [GeV] 〈P2〉/(20 MeV) Figure 2. Fit to the LHCb data for the angular moments ⟨P0⟩, ⟨P1⟩−14⟨P3⟩/9 and ⟨P2⟩for the B−→D+π−π−reaction [69]. The solid lines are the best ﬁt results, with error bands the one-sigma uncertainties propagated from the input scattering amplitudes. The dashed lines represent the LHCb ﬁt [69]. For more details, see Ref. [57]. ×106 ●●●●●●●●●●●●●●●●●●●●● ● ●● ● ● ● 2.0 2.1 2.2 2.3 2.4 2.5 0.0 0.5 1.0 1.5 MD+ π- [GeV] 〈P0〉/(20 MeV) ●●●●●●●●●●●●●●● ● ●● ● ● ● 2.2 2.3 2.4 2.5 MD+ π- [GeV] ×106 ● ●● ● ●● ●●● ●●● ●● ●● ●●●● ● ●● ● ●●● 2.0 2.1 2.2 2.3 2.4 2.5 -0.05 0.00 0.05 0.10 0.15 MD+ π- [GeV] (〈P1〉-14〈P3〉/9) /(20 MeV) ×106 ● ●● ● ●● ●●● ●●● ●● ●● ●●●● ● ●● ● ●●● 2.0 2.1 2.2 2.3 2.4 2.5 -0.05 0.00 0.05 0.10 0.15 MD+ π- [GeV] (〈P1〉-14〈P3〉/9) /(20 MeV) ×106 ●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ●● 2.0 2.1 2.2 2.3 2.4 2.5 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 MD+ π- [GeV] 〈P2〉/(20 MeV) Figure 2. Fit to the LHCb data for the angular moments ⟨P0⟩, ⟨P1⟩−14⟨P3⟩/9 and ⟨P2⟩for the B−→D+π−π−reaction [69]. The solid lines are the best ﬁt results, with error bands the one-sigma uncertainties propagated from the input scattering amplitudes. The dashed lines represent the LHCb ﬁt [69]. For more details, see Ref. [57]. The best data that can be used to constrain the D∗ 0 spectrum are from three-body B decays. In particular, the B−→D+π−π−has an advantage that there are no cross-channel π−π− resonances. In fact, the ﬁrst observations of the D∗ 0(2400) and D1(2430) were made in the B−→D(∗)+π−π−processes by Belle [32]. Recently, the LHCb experiment measured the angular moments for this process in Ref. [69]. Diﬀerent angular moments are sensitive to diﬀerent combinations of the partial-wave phases. 3 Insights from SU(3) and chiral symmetry The scalar (axial-vector) charmed mesons can be studied via the S -wave scattering between the ground state pseudoscalar (vector) charmed mesons and the light pseudoscalar mesons, and they appear as poles in the pertinent scattering amplitudes. This is the method taken in Refs. [29–31, 46, 49–58], where various versions of unitarized chiral perturbation theory (UCHPT) [59–61] were used. In short, the charmed mesons can be treated as matter ﬁelds in chiral eﬀective Lagrangians, and the corresponding scattering amplitudes are then resummed in the s-channel to satisfy unitarity and to generate poles. The SU(3) ﬂavor symmetry is incorporated into the Lagrangians, which gives the method a predictive power connecting diﬀerent isospin-strangeness channels. It is instructive to notice that the ﬂavor symmetry structure for mesonic resonances generated from such meson-meson interactions is diﬀerent from that in quark models, see, e.g., [62]. As for the charmed mesons, while c¯q mesons form an SU(3) antitriplet (¯3), the charmed-meson and light-meson pairs form an ¯3, a 6 and a 15. The leading order chiral interactions in both the ¯3 and 6 are attractive, with the strength of the former being larger. It turns out that the D∗ s0(2317) can be dynamically generated and is in the ¯3. Its nonstrange partners are the lowest D∗ 0 isospin-doublet, and their masses are about 2.1 GeV [29, 30, 63], much lower than the measured value of the D∗ 0(2400). 0 The D∗ 0(2400) (and D1(2430)) resonance parameters in the RPP [13] were obtained using the Breit-Wigner parameterization in ﬁtting to the D(∗)π invariant mass distributions. They should be questioned for two reasons: the chiral symmetry of QCD is not incorporated at the lower end of the resonance; coupled channels (D(∗)η, D(∗) s ¯K) are not considered at the higher end. Both factors can be taken into account in the unitarized chiral approach. The lattice results of the scattering lengths in a few channels at several pion masses mentioned above [45, 46] were used to ﬁx the parameters in the next-to-leading order 4 EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 https://doi.org/10.1051/epjconf/201920202001 UCHPT [46]. Using heavy quark spin and ﬂavor symmetries [65], this leads to predictions of a family of states listed in Table 2. One sees that the masses of the D∗ s0 and Ds1 are in a remarkable agreement with the measured D∗ s0(2317) and Ds1(2460) masses [13]. 1The description of the lattice energy levels in [44] relevant for the D∗ s0(2317) and Ds1(2460) using the same parameters can be found in Ref. [68]. 3 Insights from SU(3) and chiral symmetry Using the relation between the scattering length and the compositeness for shallow S -wave bound states [66], it is found that the DK molecular component of the D∗ s0(2317) wave function is about 70%. Qualitatively similar result, i.e., the D∗ s0(2317) as mainly a DK molecule, was also obtained in Ref. [67] by analyzing the lattice data of Ref. [43], as well as in the lattice calculation of Ref. [44]. A very intriguing feature of the predictions in Table 2 is that there are two D∗ 0 states, and the lower one has a mass much lower than 2.4 GeV as mentioned above. It turns out that the lower one corresponds to the SU(3) nonstrange partners of the D∗ s0(2317), while the higher one arises from the ﬂavor sextet, apparently beyond the conventional quark model for mesons. One may wonder whether the predictions are reliable. It was demonstrated [63] that the I = 1/2 energy levels in the center-of-mass frame computed by the Hadron Spec- trum Collaboration with Mπ ≃391 MeV [48] can be successfully described using the same parameters ﬁxed in [46].1 This provides a support. Recently, using updated parameters (the moving-frame data in [48] were also included in the ﬁt), it is found that the two-pole scenario persists [58]. The question is whether the two-D∗ 0 scenario is consistent with experimental measurements which were well ﬁtted using a single D∗ 0(2400). This is the focus of the next section. 5 Summary and outlook To summarize, there have been signiﬁcant progress in understanding the positive-parity charmed mesons, thanking recent developments in lattice calculations, UCHPT analyses, and precise experimental measurements. The lattice results and the relevant UCHPT anal- yses on the charm-strange mesons suggest that the main components of the D∗ s0(2317) and Ds1(2460) are DK and D∗K, respectively. This solves the ﬁrst mystery. In this hadronic molecule scenario, the nearly equal mass splittings in the second mystery is a natural con- sequence of the heavy quark spin symmetry. The resolution of the last mystery comes from that the RPP masses do not represent the lowest D∗ 0 and D1. There are two broad D∗ 0 states in that region, and the lowest one has a mass smaller than that of the D∗ s0(2317). Yet, the smoking gun for the D∗ s0(2317) to be a DK molecule is a width of O(100 keV) [46, 70], while other scenarios expect a much smaller value. This will be measured by the PANDA experi- ment. Very recently, BESIII measured the branching fraction of the D∗ s0(2317)−→D− s π0 to be 1.00+0.00 −0.14 ± 0.14 [71]. Further suggestions for Belle-II and LHCb include: • To measure precisely the angular moments, in particular ⟨P1⟩−14⟨P3⟩/9, for the B ( ) • To measure precisely the angular moments, in particular ⟨P1⟩−14⟨P3⟩/9, for the B → D(∗)ππ and B →D(∗) s ¯Kπ reactions. • To measure precisely the angular moments, in particular ⟨P1⟩−14⟨P3⟩/9, for the B → D(∗)ππ and B →D(∗) s ¯Kπ reactions. ( ) ( ) ( ) • To measure the D(∗) s ¯K spectrum in processes such as B−→D(∗)+ s K−π−, B−→D(∗)+ s K−K−, ¯B0 s →D(∗)+ s ¯K0π−and so on. This is because the higher D∗ 0 and D1 states couples strongly to the D(∗) s ¯K and could show up as near-threshold enhancements. There have been hints in the BaBar and Belle data of B−→D+ s K−π+ [72, 73], B0 →D− s KS π+ and B+ →D− s K+K+ [74]. • To search for the bottom partners of the D∗ s0(2317) (M ≃5.72 GeV) in B∗ sγ and Ds1(2460) (M ≃5.77 GeV) in B(∗) s γ. Notice that the latter is not the known Bs1(5830). Moreover, the existence of the ﬂavor-sextet state can be searched for on lattice using large SU(3) symmetric quark masses, and the bottom-nonstrange 0+ and 1+ mesons can also be studied on lattice. 4 Confronting two D∗ 0 states with the LHCb data The LHCb data for the lowest few moments are ﬁtted in Ref. [57] using a decay amplitude taking into account unitarity. The coupled-channel (Dπ, Dη, Ds ¯K) scattering amplitudes enter the decay through ﬁnal state interactions. When partial waves with L ≥3 are neglected, 5 5 EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 https://doi.org/10.1051/epjconf/201920202001 which is a good approximation for MDπ ≲2.55 GeV as indicated by the LHCb data, one can construct a special linear combination ⟨P13⟩≡⟨P1⟩−14 9 ⟨P3⟩where the D-wave contribution is cancelled out. It is particularly sensitive to the S -wave phase motion, and thus sensitive to the D∗ 0 spectrum in that energy region. As can be seen from Fig. 2, the data can be well described with the amplitudes having two D∗ 0 states. 0 From Fig. 2, the data of ⟨P13⟩show a drastic variation between 2.4 and 2.5 GeV. Such a variation naturally appears as there must be two cusps at the D0η and D+ s K−thresholds at 2.413 and 2.462 GeV, respectively. The higher cusp gets ampliﬁed by the higher pole which couples most strongly to the Ds ¯K channel. [1] B. Aubert et al. (BaBar), Phys. Rev. Lett. 90, 242001 (2003), hep-ex/0304021 [2] H.X. Chen, W. Chen, X. Liu, S.L. Zhu, Phys. Rept. 639, 1 (2016), 1601.02092 5 Summary and outlook It is foreseen that the nature of the positive-parity heavy mesons will be revealed with the joint eﬀorts from experimental, lattice QCD and theoretical sides. We are grateful to the supports received from the National Natural Science Foundation of China under Grant No. 11621131001, No. 11747601 and No. 11835015, from the Chinese Acedamy of Sciences (CAS) under Grunt No. QYZDB-SSW-SYS013 and No. XDPB09, by the CAS Center for Excellence in Particle Physics (CCEPP). References 6 https://doi.org/10.1051/epjconf/201920202001 EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 [3] H.X. Chen, W. Chen, X. Liu, Y.R. Liu, S.L. Zhu, Rept. Prog. Phys. 80, 076201 (2017), 1609.08928 [4] R.F. Lebed, R.E. Mitchell, E.S. Swanson, Prog. Part. Nucl. Phys. 93, 143 (2017), 1610.04528 [5] A. Esposito, A. Pilloni, A.D. Polosa, Phys. Rept. 668, 1 (2016), 1611.07920 [6] F.K. Guo, C. Hanhart, U.G. Meißner, Q. Wang, Q. Zhao, B.S. Zou, Rev. Mod. Phys. 90, 015004 (2018), 1705.00141 ] A. Ali, J.S. Lange, S. Stone, Prog. Part. Nucl. Phys. 97, 123 (2017), 1706.00610 [8] S.L. Olsen, T. Skwarnicki, D. Zieminska, Rev. Mod. Phys. 90, 015003 (2018), 1708.04012 [9] M. Karliner, J.L. Rosner, T. Skwarnicki, Ann. Rev. Nucl. Part. Sci. 68, 17 (2018), 1711.10626 [10] W. Altmannshofer et al. (Belle II) (2018), 1808.10567 [11] A. Cerri et al. (2018), 1812.07638 [12] D. Besson et al. (CLEO), Phys. Rev. D68, 032002 (2003), [Erratum: Phys. Rev.D75,119908(2007)], hep-ex/0305100 [13] M. Tanabashi et al. (Particle Data Group), Phys. Rev. D98, 030001 (2018) [14] S. Godfrey, N. Isgur, Phys. Rev. D32, 189 (1985) [15] S. Godfrey, K. Moats, Phys. Rev. D93, 034035 (2016), 1510.08305 [16] R.N. Cahn, J.D. Jackson, Phys. Rev. D68, 037502 (2003), hep-ph/0305012 [17] S. Godfrey, Phys. Lett. B568, 254 (2003), hep-ph/0305122 [18] P. Colangelo, F. De Fazio, Phys. Lett. B570, 180 (2003), hep-ph/0305140 [19] T. Mehen, R.P. Springer, Phys. Rev. D72, 034006 (2005), hep-ph/0503134 [19] T. Mehen, R.P. Springer, Phys. Rev. D72, 034006 (2005), hep ph/0503134 [20] O. Lakhina, E.S. Swanson, Phys. Lett. B650, 159 (2007), hep-ph/0608011 [20] O. Lakhina, E.S. Swanson, Phys. Lett. B650, 159 (2007), hep-ph/0608 [21] W.A. Bardeen, E.J. Eichten, C.T. Hill, Phys. Rev. D68, 054024 (2003), hep-ph/0305049 [22] M.A. Nowak, M. Rho, I. Zahed, Acta Phys. Polon. B35, 2377 (2004), hep-ph/0307102 [23] L. Maiani, F. Piccinini, A.D. Polosa, V. Riquer, Phys. Rev. D71, 014028 (2005), hep-ph/0412098 y ( ) p p [23] L. Maiani, F. Piccinini, A.D. Polosa, V. Riquer, Phys. Rev. D71, 014028 (2005), hep-ph/0412098 4] T.E. Browder, S. Pakvasa, A.A. Petrov, Phys. Lett. B578, 365 (2004), hep-ph/0307054 [25] A.P. Szczepaniak, Phys. Lett. B567, 23 (2003), hep-ph/0305060 6] T. Barnes, F.E. Close, H.J. Lipkin, Phys. Rev. D68, 054006 (2003), hep-ph/0305025 [26] T. Barnes, F.E. Close, H.J. Lipkin, Phys. Rev. D68, 054006 (2003), hep-ph/0305025 [27] E. van Beveren, G. Rupp, Phys. Rev. Lett. References 91, 012003 (2003), hep-ph/0305035 [28] YQ Chen X Q Li Phys Rev Lett 93 232001 (2004) hep ph/0407062 [27] E. van Beveren, G. Rupp, Phys. Rev. Lett. 91, 012003 (2003), hep-ph/0305035 ] Y.Q. Chen, X.Q. Li, Phys. Rev. Lett. 93, 232001 (2004), hep-ph/0407062 [29] E.E. Kolomeitsev, M.F.M. Lutz, Phys. Lett. B582, 39 (2004), hep-ph/0307133 [30] F.K. Guo, P.N. Shen, H.C. Chiang, R.G. Ping, B.S. Zou, Phys. Lett. B641, 278 (2006), hep-ph/0603072 hep-ph/0603072 [31] F.K. Guo, P.N. Shen, H.C. Chiang, Phys. Lett. B647, 133 (2007), hep-ph/0610008 [32] K. Abe et al. (Belle), Phys. Rev. D69, 112002 (2004), hep-ex/0307021 [33] J.M. Link et al. (FOCUS), Phys. Lett. B586, 11 (2004), hep-ex/0312060 [34] B. Aubert et al. (BaBar), Phys. Rev. D79, 112004 (2009), 0901.1291 [35] R. Aaij et al. (LHCb), Phys. Rev. D92, 012012 (2015), 1505.01505 [36] R. Aaij et al. (LHCb), Phys. Rev. D92, 032002 (2015), 1505.01710 [37] P. Colangelo, F. De Fazio, F. Giannuzzi, S. Nicotri, Phys. Rev. D86, 054024 (2012), [31] F.K. Guo, P.N. Shen, H.C. Chiang, Phys. Lett. B647, 133 (2007), hep-ph/0610008 [32] K. Abe et al. (Belle), Phys. Rev. D69, 112002 (2004), hep-ex/0307021 [33] J.M. Link et al. (FOCUS), Phys. Lett. B586, 11 (2004), hep-ex/0312060 [33] J.M. Link et al. (FOCUS), Phys. Lett. B586, 11 (2004), hep-ex/0312060 [34] B. Aubert et al. (BaBar), Phys. Rev. D79, 112004 (2009), 0901.1291 [34] B. Aubert et al. (BaBar), Phys. Rev. D79, 112004 (2009), 0901.1291 [35] R. Aaij et al. (LHCb), Phys. Rev. D92, 012012 (2015), 1505.01505 [36] R. Aaij et al. (LHCb), Phys. Rev. D92, 032002 (2015), 1505.01710 [37] P. Colangelo, F. De Fazio, F. Giannuzzi, S. Nicotri, Phys. Rev. D86, 054024 (2012), 1207.6940 [38] M.H. Alhakami, Phys. Rev. D93, 094007 (2016), 1603.08848 [39] H.Y. Cheng, F.S. Yu, Eur. Phys. J. C77, 668 (2017), 1704.01208 7 7 https://doi.org/10.1051/epjconf/201920202001 EPJ Web of Conferences 202, 02001 (2019) CHARM 2018 [40] G.S. Bali, Phys. Rev. D68, 071501 (2003), hep-ph/0305209 [41] A. Dougall, R.D. Kenway, C.M. Maynard, C. McNeile (UKQCD), Phys. Lett. B569, (2003), hep-lat/0307001 [42] D. Mohler, C.B. Lang, L. Leskovec, S. Prelovsek, R.M. Woloshyn, Phys. Rev. Lett. 111, 222001 (2013), 1308.3175 [43] C.B. Lang, L. Leskovec, D. Mohler, S. Prelovsek, R.M. Woloshyn, Phys. Rev. D90, 034510 (2014), 1403.8103 [44] G.S. Bali, S. Collins, A. Cox, A. Schäfer, Phys. Rev. D96, 074501 (2017), 1706.012 [44] G.S. Bali, S. Collins, A. Cox, A. Schäfer, Phys. Rev. D96, 074501 (2017), 1706.01247 [45] L. Liu, H.W. References Lin, K. Orginos, PoS LATTICE2008, 112 (2008), 0810.5412 [ ] , , , , y , ( ), [45] L. Liu, H.W. Lin, K. Orginos, PoS LATTICE2008, 112 (2008), 0810.5412 [46] L. Liu, K. Orginos, F.K. Guo, C. Hanhart, U.G. Meißner, Phys. Rev. D87, 014508 (2013), 1208.4535 [47] D. Mohler, S. Prelovsek, R.M. Woloshyn, Phys. Rev. D87, 034501 (2013), 1208.4059 [48] G. Moir, M. Peardon, S.M. Ryan, C.E. Thomas, D.J. Wilson, JHEP 10, 011 (2016), 1607.07093 [ ] , , y , y , ( ), [48] G. Moir, M. Peardon, S.M. Ryan, C.E. Thomas, D.J. Wilson, JHEP 10, 011 (2016), 1607.07093 [49] J. Hofmann, M.F.M. Lutz, Nucl. Phys. A733, 142 (2004), hep-ph/0308263 [49] J. Hofmann, M.F.M. Lutz, Nucl. Phys. A733, 142 (2004), hep-ph/0308263 [50] D. Gamermann, E. Oset, D. Strottman, M.J. Vicente Vacas, Phys. Rev. D76, 074016 (2007), hep-ph/0612179 [50] D. Gamermann, E. Oset, D. Strottman, M.J. Vicente Vacas, Phys. Rev. D76, 074016 (2007), hep-ph/0612179 [51] J.M. Flynn, J. Nieves, Phys. Rev. D75, 074024 (2007), hep-ph/0703047 [52] F.K. Guo, C. Hanhart, U.G. Meißner, Eur. Phys. J. A40, 171 (2009), 0901.1597 [53] P. Wang, X.G. Wang, Phys. Rev. D86, 014030 (2012), 1204.5553 [54] M. Altenbuchinger, L.S. Geng, W. Weise, Phys. Rev. D89, 014026 (2014), 1309.47 [55] Z.H. Guo, U.G. Meißner, D.L. Yao, Phys. Rev. D92, 094008 (2015), 1507.03123 [56] M.L. Du, F.K. Guo, U.G. Meißner, D.L. Yao, Eur. Phys. J. C77, 728 (2017), 1703.10836 [57] M.L. Du, M. Albaladejo, P. Fernández-Soler, F.K. Guo, C. Hanhart, U.G. Meißner, J. Nieves, D.L. Yao, Phys. Rev. D98, 094018 (2018), 1712.07957 [58] Z.H. Guo, L. Liu, U.G. Meißner, J.A. Oller, A. Rusetsky, Eur. Phys. J. C79, 13 (2019), 1811.05585 [59] T.N. Truong, Phys. Rev. Lett. 61, 2526 (1988) [60] J.A. Oller, E. Oset, Phys. Rev. D60, 074023 (1999), hep-ph/9809337 [61] J.R. Pelaez, Phys. Rept. 658, 1 (2016), 1510.00653 [62] T. Hyodo, D. Jido, A. Hosaka, Phys. Rev. D75, 034002 (2007), hep-ph/0611004 [63] M. Albaladejo, P. Fernandez-Soler, F.K. Guo, J. Nieves, Phys. Lett. B767, 465 (2017), 1610.06727 [64] C.B. Lang, D. Mohler, S. Prelovsek, R.M. Woloshyn, Phys. Lett. B750, 17 (2015), 1501.01646 [65] M. Neubert, Phys. Rept. 245, 259 (1994), hep-ph/9306320 [66] S. Weinberg, Phys. Rev. 137, B672 (1965) [67] A. Martínez Torres, E. Oset, S. Prelovsek, A. Ramos, JHEP 05, 153 (2015), 1412.1706 [68] M. Albaladejo, P. Fernandez-Soler, J. Nieves, P.G. Ortega, Eur. Phys. J. C78, 722 (2018), 1805.07104 [69] R. Aaij et al. References (LHCb), Phys. Rev. D94, 072001 (2016), 1608.01289 [70] M.F.M. Lutz, M. Soyeur, Nucl. Phys. A813, 14 (2008), 0710.1545 [71] M. Ablikim et al. (BESIII), Phys. Rev. D97, 051103 (2018), 1711.08293 [72] B. Aubert et al. (BaBar), Phys. Rev. Lett. 100, 171803 (2008), 0707.1043 [73] J. Wiechczynski et al. (Belle), Phys. Rev. D80, 052005 (2009), 0903.4956 [74] J. Wiechczynski et al. (Belle), Phys. Rev. D91, 032008 (2015), 1411.2035 8
https://openalex.org/W2949926905	https://hal.inria.fr/hal-02365733/document	English	null	Identifying Information Security Risks in a Social Network Using Self-organising Maps	IFIP advances in information and communication technology	2,019	cc-by	6,306	To cite this version: Rudi Serfontein, Hennie Kruger, Lynette Drevin. Identifying Information Security Risks in a Social Network Using Self-organising Maps. 12th IFIP World Conference on Information Security Education (WISE), Jun 2019, Lisbon, Portugal. pp.114-126, 10.1007/978-3-030-23451-5_9. hal-02365733 Identifying Information Security Risks in a Social Network Using Self-organising Maps Rudi Serfontein, Hennie Kruger, Lynette Drevin To cite this version: Rudi Serfontein, Hennie Kruger, Lynette Drevin. Identifying Information Security Risks in a Social Network Using Self-organising Maps. 12th IFIP World Conference on Information Security Education (WISE), Jun 2019, Lisbon, Portugal. pp.114-126, 10.1007/978-3-030-23451-5_9. hal-02365733 Distributed under a Creative Commons Attribution 4.0 International License HAL Id: hal-02365733 https://inria.hal.science/hal-02365733v1 Submitted on 15 Nov 2019 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License Identifying information security risks in a social network using self-organising maps Rudi Serfontein1[0000-0002-0428-6494], Hennie Kruger2[0000-0001-8514-4422], Lynette Drevin3[0000-0001-9370-8216] North-West University, Potchefstroom, South Africa 1 rudi.serfontein@nwu.ac.za 2 hennie.kruger@nwu.ac.za 3 lynette.drevin@nwu.ac.za Abstract. Managing information security risks in an organisation is one of the most important tasks an organisation has. Unfortunately, due to the complexity of most organisational systems, identifying information security risks can be dif- ficult. One way to identify possible risks in an organisation is to make use of Social Network Analysis (SNA). While they can be used to identify risks, the metrics calculated using SNA are often numerous and daunting to managers un- familiar with SNA. Furthermore, as the data in this form tend to be uncomfortable to process, educating managers about risks in their organisation can be quite dif- ficult. Also, as these metrics often require quantitative processing in order to be useful, SNA on its own is not always an attractive method to use to identify risks in an organisation. In this paper the use of self-organising maps to identify pos- sible information security risks in an organisation is investigated. Risk data were obtained from an organisation that deals in risk management, which were used to build a social network. A number of metrics associated with risk were calculated from the network, and these metrics were used to cluster the various entities using a self-organising map. Certain entities that pose a possible information security risk were identified. The results suggest that it may be viable to use self-organis- ing maps, in concord with SNA, to more easily identify risks in an organisation using visual methods. Keywords: Self-Organising Maps · Social Network Analysis · Information Se- curity. Keywords: Self-Organising Maps · Social Network Analysis · Information Se- curity. 1 Introduction 1 Information security risk management is one of the most crucial parts of information security and should be one of the most important actions taken by organisations [1]. Unfortunately, due to the relative complexity of most organisational systems, identify- ing information security risks that are inherent to people using the systems, and making managers aware of them, is often quite difficult. One of the methods proposed in recent years to address such risks involve the use of Social Network Analysis (SNA) [2-4]. SNA is a method that can be used to evaluate an organisation, for instance a community or business, in such a way that social interactions can be studied quantitatively, rather than qualitatively [5]. It does however have a significant drawback in that large net- works, when visualised, may have so many nodes and arcs that the network is visually incomprehensible. In order to address this drawback, a number of studies have em- ployed techniques that alter significant nodes and edges of a visualised network in order to draw attention to certain aspects. Some of these techniques include differentiating the colour of nodes and edges [6], using differing sized nodes to correspond to certain metrics [7], and using labels of various sizes [3]. A somewhat more novel technique makes use of Self Organising Maps (SOMs) to directly visualise network data [8]. A SOM is an effective technique that can be used not only to visualise high-dimensional data, but to visualise it in such a way that the result can act as both a similarity graph and a clustering diagram [9]. The SOM technique can be used to identify similar nodes within a social network, even in the presence of seemingly contradicting attributes, and present this data in a way that managers are quickly informed of risks in the network. SOMs have also been used in information security research to propose improvements to intrusion detection methods [10], and as a method for analysing information security behavioural data [11, 12]. While the approach suggested by Boulet, Jouve, Rossi and Villa [8] does allow for social networks to be visualised as SOMs, it has a shortcoming in that the SOMs generated can not necessarily be used in a way that is relevant to the process of identifying possible information security risks. 2 Background The primary theme of this paper is the use of Social Network Analysis (SNA) metrics as inputs for a Self-Organising Map (SOM), which should aid in evaluating risk in an organisation. In order to demonstrate how this can be done, five SNA metrics will be discussed briefly. While there are dozens of SNA metrics that can potentially be used, the five discussed here were chosen based on their established relationships with risk in the literature. The section will start with a description of SOMs, followed by the evaluation of the selected SNA metrics in relation to risk. 1 Introduction This is mainly as a result of the fact that, in order to identify risks within a social network, a number of metrics calculated from the network data is used rather than the raw data itself. In this paper the feasibility of using existing SOM techniques to inform managers of possible risks in an organisation is discussed. The value of such an application is two- fold; firstly, by using a visualisation method that reduces the amount of data that is visualised, the often confusing graphs produced by traditional SNA visualisation tech- niques can be replaced with SOMs that are easier to process graphically. Furthermore, as SOM algorithms produce maps that naturally display data of interest, analysis and evaluation of the data should no longer require the visualisation results to be adapted (node enlargement, colouration, etc.) in order to be meaningful. Secondly, as SOMs organise similar data into clusters, their application should make it easier to inform a manager of groups of similar at-risk entities. This is thanks to the clustering done by the SOM algorithms, as, due to the relationship between certain SNA metrics and the CIA triad (Confidentiality, Integrity and Availability), as will be discussed in Section 2, clustering the nodes according to these metrics allows an evaluator to quickly identify similar problematic nodes. Furthermore, because the SOM algorithm uses the calcu- lated SNA metrics as attributes to determine the clusters, the clusters themselves can be used to infer similarities that may not be readily transparent from the available data. Another advantage is that managers can be informed and educated of possible risks in an organisation early on, which may aid in developing effective awareness, education, and training programs. The graphical nature of SOM may also make it a useful tool for training inexperienced risk managers, and can potentially aid in identifying standard trends and patterns. In the remainder of the article the background, research methodology, and results, will be discussed respectively. The background discussion will focus on SNA in the context of information security, as well as SOMs. The discussion of the method will focus on both the application of the techniques, and the data collection phase. The paper will then conclude with a discussion of the results and implications. 2.1 Self-Organising Maps (SOM) The self-organising map (SOM) is a neural network technique that can be used to vis- ualise and evaluate high-dimensional data [13]. The SOM technique uses given data to produce a self-organising neural network wherein the data points are clustered into topographical regions [14]. This visualisation technique has a wide range of known applications, from evaluating comparable biological adaptations [15] and improving optimisation algorithms [16, 17], to clustering data for problem-solving purposes [14, 18]. One of the greatest advantages SOM has over other high-dimension visualisation techniques is that it produces a two-dimensional topographical map that can be evalu- ated and interpreted without any special knowledge or skills. In addition to clustering known data points, depending on the data, the technique can also be used in vector quantisation, and as a regression modelling technique [13]. All these methods can ar- guably be used to obtain valuable information about data, but in the context of this paper only the clustering function of SOM will be considered. The algorithm for devel- oping a SOM [19] is shown below. Input: Dataset N Output: A topographical map M containing the data from N, sorted into topographical areas Variables: 𝑤𝑖𝑗 – Weight vector describing topographical area 𝑖𝑗; either randomised or defined at start 𝑥 – An input vector contained in N 𝛼 – Learning rate that is a slowly decreasing function of time 1. while Stop condition is false 2. For each 𝑥 in N 3. For each vector 𝑗 4. \| Compute 𝐷(𝑗) = ∑(𝑤𝑖𝑗− 𝑥𝑖)2 𝑖 5. end 𝛼 – Learning rate that is a slowly decreasing function of time 6. Find index 𝐽 such that 𝐷(𝐽) is a minimum 7. For all units 𝑗 in a topographical area 𝐽, and for all 𝑖: 8. \| Compute 𝑤𝑖𝑗(𝑛𝑒𝑤) = 𝑤𝑖𝑗(𝑜𝑙𝑑) + 𝛼[𝑥𝑖− 𝑤𝑖𝑗(𝑜𝑙𝑑)] 9. end 10. end 11. Update 𝛼 12. Reduce radius of topographical area at specified times 13. Test Stop condition 14. end 15. return M This algorithm produces one map with all of the entities sorted into clusters. Certain software suites, such as Viscovery SOMine [20], provide additional information by colouring the same map using values from different attributes. As stated, SOMs can be used to cluster high-dimensional data on a two-dimensional map, producing a result that can be interpreted easily without training. 2.1 Self-Organising Maps (SOM) This makes the technique especially valuable to those in managerial positions, as these individuals may not have the time to study large reports and data sets in detail, and may also hold true for the outcomes of SNA based studies – especially if the resulting network is particu- larly large or complex. By calculating the SNA metrics, as discussed in the next section, and applying SOM to the resulting data set, the risks posed by certain individuals or groups can be determined and presented visually in a way that is easy to process and interpret. Additionally, a number of at-risk individuals may be identified that would not necessarily have been evident through the use of more traditional visualisation tech- niques such as bar-graphs. It is possible for an individual to have all the traits of a high- risk individual and not be an obvious risk from the data itself. In these instances, a clustering technique such as SOM can be used to identify individuals that have similar, possibly hidden, attributes. This makes it significantly easier to address certain infor- mation security risks, as larger scale programmes can be developed to target groups consisting of similar individuals. In summary, SOM is a valuable technique to use in addition to SNA, as the clustering function of SOM can be used to infer invaluable information about information security risks if the correct and relevant SNA metrics are used. 2.2 Social Network Analysis (SNA) in the context of Information Security One of the most well-known frameworks for information security is the CIA triad [21], which references Confidentiality, Integrity, and Availability. Confidentiality describes the access rights that users have to a piece of information. For example, a manager having confidential access to certain business data that his employees do not, or should not, have. One possible SNA metric that can be used to evaluate a risk to confidentiality is closeness centrality. Closeness centrality is calculated by determining all the shortest distances to all other nodes within the network [22], so a node with a high closeness centrality has a large number of close relationships to other nodes in the network. Such a node may therefore have access to information that it should not have access to. Al- ternatively, if the node is an object or a resource, such as a shared computer or a pho- tocopier, it could become a significant confidentiality risk if malware or untrustworthy maintenance personnel are involved. Integrity describes not only how accurate any piece of information is but, by exten- sion, how trustworthy it is. One of the SNA metrics that can be used to evaluate the risk a node poses to the integrity of the information in the network is total degree centrality. The total degree centrality measure is concerned with an individual node’s position within the network [22, 23], and is determined by using the number of nodes leading into and out of a node. A node with a high total degree centrality is well connected within the network, and may have enough influence over other nodes to impact the integrity of the information passing through them. Consider, as an example, an office worker with a high total degree centrality that has to capture data for a corporate data- base. If this worker were to make a mistake in capturing the data, the integrity of the data that a large number of nodes rely on may be compromised. The betweenness centrality measure can be used to identify nodes that are risks to both integrity and confidentiality. This measure is a representation of the number of times that a particular node is part of the shortest path between any two nodes in the network [24]. It is reflective of the number of indirect nodes that are connected to that node. 2.2 Social Network Analysis (SNA) in the context of Information Security To demonstrate the rationale behind using betweenness centrality as an indicator of risk to both integrity and confidentiality, consider a department with a “go-to” indi- vidual. This individual will likely have access to greater amounts of information than is ideal, and would be in a position to alter the information flowing through the network. Furthermore, as nodes with high betweenness measures tend to act as brokers, this in- dividual may be seen as a trustworthy shortcut to obtain information in the network, which places it in a position to obtain greater amounts of information, as well as ma- nipulate information as it flows through the network. The final member of the triad, availability, deals with the ability to access the data in a timely manner. Availability is often at odds with both confidentiality and integrity, as systems meant to protect availability and confidentiality often impact on the availa- bility of the data. With regards to SNA, one of the metrics that may identify a high risk node in terms of availability is the one that identifies a node as a boundary spanner. A boundary spanner is a node that, if removed, will cause one part of the network to be- come completely isolated [25], thereby negatively impacting the availability of infor- mation in certain parts of the network. The final SNA metric to be mentioned here is eigenvector centrality. Eigenvector centrality measures the extent to which a particular node is connected to highly con- nected nodes [22]. Nodes that have a high eigenvector measure are considered to pos- sess emergent leadership properties [26] and may be considered a potential risk to con- fidentiality, integrity, and availability. For example, consider the impact an informal leader can have on information in a network. Confidential information may be shared with such a node as a result of the connections with highly connected nodes, whereas the integrity of the data in the network may be impacted by the additional knowledge the node obtains. Availability may also be impacted negatively if the emergent leader convinces other nodes to delay the flow of information, or if the information is redi- rected through the network along suboptimal routes. A summary of the SNA metrics discussed, and how they relate to the CIA triad, is given in Table 1. Table 1. 2.2 Social Network Analysis (SNA) in the context of Information Security SNA metrics in the context of Information Security SNA Metric CIA Triad Rationale C I A Total degree centrality X Nodes with a high total degree centrality have influence in the net- work and are connected to a significant portion of the network. Closeness cen- trality X Nodes with a high closeness centrality have access to a significant amount of information in the network. Betweenness centrality X X Nodes with a high betweenness centrality may be prone to infor- mation brokering and tampering. Boundary spanner X If a boundary spanner node is removed, an entire section of the net- work becomes isolated Eigenvector centrality X X X Nodes with a high eigenvector centrality are considered emergent leaders and, depending on their influence and attitude, may be a gen- eral risk, depending on the circumstances Table 1. SNA metrics in the context of Information Security It should be emphasised that, while this brief discussion focussed on each member of the CIA triad individually, confidentiality, integrity, and availability are all intercon- nected. It is possible, for instance, for a significant enough increase in confidentiality to result in a significant reduction in availability. This is also true for the relationship between confidentiality and integrity, and the relationship between integrity and avail- ability. The goal is typically to find a balance between these three aspects that is appro- priate to the particular situation. This interrelatedness should be kept in mind when evaluating risks using SNA, as well as when selecting controls to address these risks. From this short discussion it is clear that SNA metrics can be used to identify and possibly evaluate risks in terms of the CIA triad. The use of SNA metrics as input data for a SOM is therefore appropriate, and its use in this manner may help to identify and visualise risks in an organisation. This should help to improve overall awareness of risks in the organisation, aid in training managers, and may even help to determine overall preventative measures. The application in the rest of the paper uses the five SNA metrics discussed in this section as input for a SOM. 3 Method The study was conducted using data provided by a manager from a large company that deals with risk evaluation. The data are confidential, and were subsequently anony- mised prior to publication. Using this data, a network was built that describes the rela- tionship between various entities. This network is shown in Fig. 1. The entities, or nodes, of the network include 26 real-world risks, 612 controls, 6 risk owners, 26 control owners, 13 risk coordinators and 12 governing bodies. The risks are those risks that the organisation has to manage, whilst the controls are those controls used to manage the risks. The risk- and control owners are ultimately responsible for the risks and controls respectively, whereas the risk coordinators ensure that the correct risks are managed using the appropriate controls. The governing bodies are responsible for determining which control is used with which risk. These bodies also determine what the probability of a risk occurring is, as well as the severity of such an occurrence. The network is undirected, as unidirectional relationships between entities such as risks and risk owners do not seem realistic. The network data were processed using the software suite ORA-Lite [24], while the SOMs were generated using the Viscovery SOMine Suite [20]. The data for the SOM consists of the 5 SNA metrics for each node. In total, 695 nodes are contained in the network, and a total number of 1738 links exist between them. The focus of the network is on managing real-world risks, and it was subsequently processed in a risk centric way, i.e. the relationships between the nodes are based on similar relationships to par- ticular risks. This means that the relationship between a risk coordinator and a control owner, for example, is described only in terms of their shared relationship to the same risk. Fig. 1. Social network showing relationships between risks (blue), risk controls (pur- ple), risk coordinators (yellow), risk owners (orange), control owners (pink) and the governing bodies (green) responsible for appointing the various role-players. Fig. 1. Social network showing relationships between risks (blue), risk controls (pur- ple), risk coordinators (yellow), risk owners (orange), control owners (pink) and the governing bodies (green) responsible for appointing the various role-players. With networks of this size, the large number of metric values that are produced can be quite complex. 3 Method In order to help evaluate such complex data in a simpler, graphical way, the SOM algorithm is used. The SOMs can be used to quickly identify problem areas, which should make it easier to evaluate the data. Additionally, as SOMs are graphical in nature, they can be applied iteratively to investigate how the risk in a network changes over time, or as certain controls are introduced that aim to manage those risks. When applying this technique to training risk managers, one of the aims is to high- light certain trends or groups that may pose a natural risk in the network. In doing so, the risk manager should be better informed of the nuances of the network and may be able to introduce more effective risk mitigation measures than would have been possi- ble otherwise. Consider, for example, a network, such as the one shown in Fig. 1, with risks, control coordinators, and controls. If a SOM is developed for the network, the manager should be able to readily identify groups of nodes that have similar risk pro- files based on their clustering. If a certain grouping of controls and control coordinators are found in the same cluster, for instance, it may indicate that there is a problem with the way in which the controls are managed. Alternatively, if all of the control coordi- nators are grouped into one high risk cluster, it may be appropriate to introduce measures, such as policies, to address the risks posed by these nodes. Another way in which the technique can be used in training is to monitor how the risk profile of certain clusters change when new controls are implemented to address the identified risks. As a SOM is graphical in nature, and the geographical structure of the map changes as the risk values for the nodes change, it should be possible to identify the changes in the network graphically. This is especially true for clusters that lose nodes, as the area of the map that the cluster occupies should be reduced. By using the SOMs as a graphical aid, the manager should be able to identify which approaches work best, and under which circumstances. 4 Results and discussion The SOM algorithm produced a map with three regions, or clusters, when applied to the network data. This map is shown in Fig. 2 to Fig. 6. Each figure shows the same map, but with different colourations. The colourations are used to show how the values of the five measures differ for various nodes. The clusters are the same in each image, as one map, built using all five SNA metrics as node attributes, was obtained. In Fig. 2, where the boundary spanner measure is used to colour the map, the red colouration that covers most of Cluster C2 indicates that the nodes in C2 pose a signif- icant possible risk, as the nodes in this cluster have a much higher boundary spanner value than the nodes in other clusters. A cursory evaluation of the cluster’s data shows that C2 exclusively contains all of the nodes that represent the risks. It should be noted that, while the risks are all found in the same cluster for this network, this will not necessarily be the case for all networks. As the boundary spanner metric indicates that a node’s removal will completely isolate a part of the network, this suggests that the current structure in the network includes nodes that will be isolated if any of the risks are resolved. The network data itself shows that these nodes are primarily controls. If, for example, the risk “Corporate Brand” is completely resolved, i.e. if the company finds itself in a situation where there is no risk at all to the company brand, then the controls that exist to manage that risk, such as “Social Media strategy and protocols” and “Expert communications resources”, are no longer needed. In order to ensure that these controls are not kept in place unnecessarily, additional measures need to be im- plemented. for example, the risk “Corporate Brand” is completely resolved, i.e. if the company finds itself in a situation where there is no risk at all to the company brand, then the controls that exist to manage that risk, such as “Social Media strategy and protocols” and “Expert communications resources”, are no longer needed. In order to ensure that these controls are not kept in place unnecessarily, additional measures need to be im- plemented. Fig. 2. SOM (boundary spanner) Fig. 3. SOM (total degree centrality) Fig. 4. SOM (closeness centrality) Fig. 5. SOM (eigenvector centrality) Fig. 6. 4 Results and discussion This area, which is situated on the left hand side of cluster C3 where it borders Cluster C2, is coloured green in Fig. 4 and green, orange, and red in Fig. 5. Additionally, a hotspot is present on the border between C1 and C2, as shown in Fig. 4. While the nodes in the hotspot area between C1 and C2 certainly pose a risk to confidentiality, the section of C3, where the nodes have much higher values for closeness centrality and eigenvector centrality than the rest of the nodes in the cluster, warrants further investigation. The higher closeness centrality measure of these nodes, shown in Fig. 4, suggest that they are a risk to confidentiality, whereas the very high eigenvector centrality measure shown in Fig. 5 indicate that they are an over- all risk. There are two nodes in particular that fall into this region, one being a risk owner that is responsible for 11 out of the 26 risks, and the other being a risk coordinator that is responsible for 6 risks. The remainder of C3 is low risk and, as both of these nodes are in the cluster, it intimates that the risk posed by this risk owner and risk co- ordinator could be alleviated by reducing the number of risks that they are responsible for. Some of the risks could be transferred to other risk owners and coordinators. Alter- natively, the risks can be co-owned and co-coordinated with owners and coordinators that are responsible for a smaller number of risks. With the exception of the small area of low risk in C3, Fig. 4 shows that most of the nodes in the network have a high measure of closeness centrality. This suggests that controls should be in place to protect the confidentiality of the information in the net- work in general, as almost any one of the nodes could be responsible for compromising confidentiality. Fig. 6 highlights the existence of a single hotspot that exists in cluster C1 with re- gards to betweenness centrality. The hotspot, which is situated on the left hand side of C1 and has a red colour, contains a single risk coordinator, which poses a risk to both the integrity and the confidentiality of information in the network. In order to resolve this risk, the dependence on the specific risk coordinator should be reduced. 4 Results and discussion SOM (betweenness centrality) The colouration, based on the total degree centrality, used on the SOM in Fig. 3 shows that there is an area in cluster C2 where the nodes have unusually high total degree Fig. 2. SOM (boundary spanner) Fig. 3. SOM (total degree centrality) Fig. 4. SOM (closeness centrality) Fig. 5. SOM (eigenvector centrality) Fig. 3. SOM (total degree centrality) Fig. 2. SOM (boundary spanner) Fig. 3. SOM (total degree centrality) Fig. 2. SOM (boundary spanner) Fig. 3. SOM (total degree centrality) Fig. 2. SOM (boundary spanner) Fig. 5. SOM (eigenvector centrality) Fig. 4. SOM (closeness centrality) Fig. 5. SOM (eigenvector centrality) Fig. 5. SOM (eigenvector centrality) Fig. 4. SOM (closeness centrality) Fig. 4. SOM (closeness centrality) Fig. 6. SOM (betweenness centrality) Fig. 6. SOM (betweenness centrality) The colouration, based on the total degree centrality, used on the SOM in Fig. 3 shows that there is an area in cluster C2 where the nodes have unusually high total degree centrality values, which is associated with a higher level of risk to integrity. Of the nodes in C2, there are six nodes, indicated by the red area in C2, that have a significantly higher amount of total degree centrality than the rest of the nodes. These nodes are the risks “Forest fires”, “Environmental impacts”, “Interruption to supply networks”, “Waste Treatment Capacity”, “Urban Resource Capacity”, and “Rural Resource Ca- pacity”. All of these risks can have a significant impact if realised, which is why the influence they have is so substantial. This also means that any errors with regards to these risks, such as the risk of forest fires being over- or underemphasised, can have a significant impact on the information that is ultimately used in the network to manage other risks. If, for example, the integrity of the information with regards to the chances of a forest fire occurring is compromised, then there may not be enough water available to address the fire. If the fire affects any industrial assets, this may have environmental impacts, which in turn could negatively affect the company’s corporate brand image. To protect the integrity of the information of these nodes, additional controls should be implemented. An area with values much higher than the surrounding areas is shown to be present in cluster C3 in both Fig. 4 and Fig. 5. 4 Results and discussion The de- pendence of the risk coordinator can be reduced in one of at least two ways. The first method involves transferring some of the coordinator’s responsibilities, such as risks, to another coordinator. This coordinator should preferably be located in cluster C1, as such a coordinator is likely to have a similar amount of power and influence. The sec- ond method that could be used is to employ an additional risk coordinator to take over some of the duties. This coordinator could also assume some of the duties of the risk coordinator in C3, thereby reducing the risk of two nodes simultaneously. y g y Based on all the SOMS shown in Fig. 2 to Fig. 6, there are six risks, two risk coor- dinators, and a risk owner that pose potential risks to the overall security of information in this network. From this discussion, the advantage of using the SOM method to visu- alise SNA metric data is clear: by using SOMs to visualise SNA metric data, a relatively simple process can be followed in order to evaluate the risks in a network. The ad- vantages are especially clear when compared to the process that would be needed in order to evaluate the risks in a network, such as the one shown in Fig. 1, or when using only raw data and statistical analysis. The use of SOMs in this manner gives managers the opportunity to evaluate risks graphically, as well as to compile risk discussion re- ports that do not require any prior knowledge of SOMs or SNA. This, in turn, could help improve the nature and quality of risk management, as a greater number of options and plans could develop as a result. Furthermore, SOMs allow for a way to systemati- cally identify risks, and can also be used to monitor the progress and impact of risk mitigation strategies. Depending on the situation, it may be possible to identify positive or negative changes in the network almost instantaneously using this method. The cen- tral premise, i.e. that SOMs can be used to visualise SNA risk data, and in turn help educate managers of risks in their organisation in a quick and simple way, is therefore feasible. References 1. Wangen, G.: Information security risk assessment: A method comparison. Computer, vol. 50, no. 4, pp. 52-61 (2017). 2. Armstrong, H., Armstrong, C. and McCulloh, I.: A course applying network analysis to organizational risk in information security, In: South African Information Security Multi-Conference, pp. 204-214 (2010). 3. Dang-Pham, D., Pittayachawan, S. and Bruno, V.: Investigation into the formation of information security influence: Network analysis of an emerging organisation. Computers & Security, vol. 70, pp. 111-123 (2017). 4. Serfontein, R., Drevin, L. and Kruger, H.A.: The feasibility of raising information security awareness in an academic environment using SNA, In: IFIP World Conference on Information Security Education, pp. 69-80. Springer (2018). 5. Scott, J. and Carrington, P.J.: The SAGE handbook of social network analysis. SAGE publications (2011). 6. Tsui, E. and Liebowitz, J.: Linking social network analysis with the analytic hierarchy process for knowledge mapping in organizations. Journal of Knowledge Management, vol. 9, no. 1, pp. 76-86 (2005). 7. Dang-Pham, D., Pittayachawan, S. and Bruno, V.: Applying network analysis to investigate interpersonal influence of information security behaviours in the workplace. Information & Management, vol. 54, no. 5, pp. 625-637 (2017). 8. Boulet, R., Jouve, B., Rossi, F. and Villa, N.: Batch kernel SOM and related laplacian methods for social network analysis. Neurocomputing, vol. 71, no. 7, pp. 1257-1273 (2008). 9. Kohonen, T.: The self-organizing map. Neurocomputing, vol. 21, no. 1-3, pp. 1-6 (1998). 10. De la Hoz, E., De la Hoz, E., Ortiz, A., Ortega, J. and Prieto, B.: PCA filtering and probabilistic SOM for network intrusion detection. Neurocomputing, vol. 164, no. Supplement C, pp. 71-81 (2015). 11. Hunt, R. and Hill, S.: Using security logs to identify and manage user behaviour to enhance information security, In: 14th European Conference on Cyber Warfare and Security, pp. 111. Academic Conferences Limited (2015). 12. López, A.U., Mateo, F., Navío-Marco, J., Martínez-Martínez, J.M., Gómez-Sanchís, J., Vila-Francés, J. and Serrano-López, A.J.: Analysis of computer user behavior, security incidents and fraud using self-organizing maps. Computers & Security, vol. 83, pp. 38-51 (2019). 13. Bäck, T., Kok, J.N. and Rozenberg, G.: Handbook of natural computing. Springer (2012). 14. Pal, C., Hirayama, S., Narahari, S., Jeyabharath, M., Prakash, G. and Kulothungan, V.: An insight of world health organization (WHO) accident database by cluster analysis with self-organizing map (SOM). Traffic Injury Prevention, vol. 19, no. sup1, pp. S15- S20 (2018). 15. 5 Conclusion Information security risk management is one of the most crucial parts of information security, but it is often complicated by the complexity of most organisational systems. In order to simplify the task of identifying risks in an organisation, SOM can be used to identify possible risks in an organisation by visualising SNA metric data of the or- ganisation. A SOM, which clusters similar entities into geographically separate regions, is relatively simple to evaluate due to its graphical nature. When compared to other risk identification techniques that employ SNA metrics, which may require risk managers to process and evaluate large tables of numbers, the use of SOM may reduce the amount of work needed, as entities that pose a threat to the organisation can be identified with relative ease. Additionally, as a SOM is easier to evaluate, inexperienced risk managers may find the use of SOM less daunting than to use numerical data and statistical anal- ysis. Finally, as a SOM provides an additional level of information that may not be readily apparent from the data, it could aid in educating risk managers of dangers in the organisation that may not be known, or obvious. References Nakayama, H., Sakamoto, T., Okegawa, Y., Kaminoyama, K., Fujie, M., Ichihashi, Y., Kurata, T., Motohashi, K., Al-Shehbaz, I., Sinha, N. and Kimura, S.: Comparative transcriptomics with self-organizing map reveals cryptic photosynthetic differences between two accessions of north american lake cress. Scientific Reports, vol. 8, no. 1, pp. 3302 (2018). 16. 16. Gu, F. and Cheung, Y.-M.: Self-organizing map-based weight design for decomposition-based many-objective evolutionary algorithm. IEEE Transactions on Evolutionary Computation, vol. 22, no. 2, pp. 211-225 (2018). 17. Kuo, R.J., Rizki, M., Zulvia, F.E. and Khasanah, A.U.: Integration of growing self- organizing map and bee colony optimization algorithm for part clustering. Computers & Industrial Engineering, vol. 120, pp. 251-265 (2018). 18. Lee, Y.: Using self-organizing map and clustering to investigate problem-solving patterns in the massive open online course: An exploratory study. Journal of Educational Computing Research, pp. 0735633117753364 (2018). 19. Fausett, L.V.: Fundamentals of neural networks: Architectures, algorithms, and applications. Prentice-Hall Englewood Cliffs (1994). 20. Viscovey SOMine. www.viscovery.net/somine, last accessed 10/02/2019. 21. Au, C.H., Fung, W.S. and Tses, A.: An investigation on the relationship between control self-assessment, cloud security, and cloud-related business performance-using partial least squares, In: Industrial Engineering and Engineering Management (IEEM), pp. 1879-1883. IEEE (2016). 22. Armstrong, H. and McCulloh, I.: Organizational risk using network analysis, In: South African Information Security Multi-Conference, pp. 132-141 (2010). 23. Hanneman, R.A. and Riddle, M.: Introduction to social network methods. University of California (2005). Hanneman, R.A. and Riddle, M.: Introduction to social network methods. University of California (2005). . ORA-Lite. www.casos.cs.cmu.edu/projects/ora, last accessed 24/04/2018. 25. Cormen, T.H., Leiserson, C.E., Rivest, R.L. and Stein, C.: Introduction to algorithms second edition. The MIT Press (2001). 26. Borgatti, S.P.: Centrality and network flow. Social networks, no. 27, pp. 55-71 (2005).
https://openalex.org/W4367317802	https://www.frontiersin.org/articles/10.3389/fimmu.2023.1156470/pdf	English	null	Efficacy and safety of rituximab for primary membranous nephropathy with different clinical presentations: a retrospective study	Frontiers in immunology	2,023	cc-by	8,050	OPEN ACCESS OPEN ACCESS EDITED BY Qi Wang, Huazhong University of Science and Technology, China REVIEWED BY Marco Allinovi, Careggi University Hospital, Italy Marco Fiorentino, University of Bari Aldo Moro, Italy Marios Papasotiriou, University of Patras, Greece Jin Shang, First Afﬁliated Hospital of Zhengzhou University, China CORRESPONDENCE Bing Chen chenbing3668@163.com RECEIVED 01 February 2023 ACCEPTED 06 April 2023 PUBLISHED 28 April 2023 CITATION Zhang S, Huang J, Dong J, Li Z, Sun M, Sun Y and Chen B (2023) Efﬁcacy and safety of rituximab for primary membranous nephropathy with different clinical presentations: a retrospective study. Front. Immunol. 14:1156470. doi: 10.3389/fimmu.2023.1156470 COPYRIGHT EDITED BY Qi Wang, Huazhong University of Science and Technology, China Shasha Zhang 1, Jing Huang 2, Jianwei Dong 3, Zhuo Li 1, Mengyao Sun 1, Yujiao Sun 1 and Bing Chen 1 1Department of Nephrology, Shandong Provincial Hospital Afﬁliated to Shandong First Medical University, Jinan, Shandong, China, 2Department of Nephrology, Jinan Shizhong People’s Hospital, Jinan, China, 3Department of Thoracic Surgery, The People’s Hospital of Rongcheng, Rongcheng, Shandong, China Background: Rituximab (RTX) is gaining increasing clinical acceptance in the treatment of primary membranous nephropathy (PMN), with demonstrated efﬁcacy and safety. However, there are few clinical studies on RTX for PMN in Asian populations, especially in China. Methods: To observe and analyse the efﬁcacy and safety of RTX treatment, 81 patients with PMN suffering from nephrotic syndrome (NS) were enrolled and divided into an initial therapy group, a conventional immunosuppressive therapy relapse group, and a conventional immunosuppressive therapy ineffective group according to their pre-RTX treatment background. Patients in each group were followed up for 12 months. The primary outcome was clinical remission at 12 months, and the secondary outcomes were safety and the occurrence of adverse events. COPYRIGHT © 2023 Zhang, Huang, Dong, Li, Sun, Sun and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Results: At 12 months, 65 of 81 (80.2%) patients achieved complete (n=21, 25.9%) or partial (n=44, 54.3%) remission after rituximab treatment. TYPE Original Research PUBLISHED 28 April 2023 DOI 10.3389/fimmu.2023.1156470 COPYRIGHT © 2023 Zhang, Huang, Dong, Li, Sun, Sun and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. TYPE Original Research PUBLISHED 28 April 2023 DOI 10.3389/fimmu.2023.1156470 1 Introduction targeted by podocytes. The pathogenic role of autoantibody- producing B cells in PMN is gradually being understood, providing powerful evidence for B-cell-targeted therapy for the PMN. Rituximab (RTX) is a selective B-cell depleting agent that depletes CD20-positive B cells by speciﬁcally binding to the B-cell surface antigen CD20 (14), reducing circulating antibody production and thereby preventing the formation of subepithelial immune deposits in the glomerulus and attenuating glomerular Primary membranous nephropathy (PMN) is one of the most common pathological types of adult nephrotic syndrome (NS), in which subepithelial immune complex deposition (mainly IgG and C3) and complement activation are responsible for impaired podocyte function. The course of the disease is highly variable, ranging from spontaneous remission to persistent proteinuria or end-stage kidney disease (ESKD) (1, 2). Spontaneous partial remission occurs in approximately 1/3 of patients (3); however, remission is unlikely to occur in intermediate- and high-risk patients (4, 5). Approximately 40%-50% of patients with untreated persistent NS will develop ESKD and therefore require prompt clinical intervention and treatment. ﬁltration barrier damage, thus leading to PMN remission (15). Since 2002, various clinical studies, represented by the MENTOR study (16), have conﬁrmed the clinical efﬁcacy and safety of RTX through comparative tracking of RTX and conventional regimens for the treatment of PMN, and the 2021 KDIGO guidelines (5) have also recommended the clinical use of RTX. However, its treatment has been less studied in Asian populations, especially in Chinese populations. This study selected domestic rituximab injection, manufactured by Shanghai Fosun Pharmaceutical Company in China, as the primary investigational agent. This anti-CD20 monoclonal antibody has been studied equivalently with imported rituximab from Roche Diagnostics Gmbh (17, 18), conﬁrming that there is no signiﬁcant difference in the efﬁcacy of these two products in the treatment of lymphoma. However, there is still a lack of adequate research support for the therapeutic efﬁcacy of this product in membranous nephropathy, especially in the Chinese population. The study was performed to further evaluate the efﬁcacy and safety of RTX in PMN by retrospectively analysing the outcomes of this product when applied alone to PMN in three different treatment settings. OPEN ACCESS Thirty-two of 36 (88.9%) patients in the initial therapy group, 11 of 12 (91.7%) patients in the relapse group and 22 of 33 (66.7%) patients in the ineffective group achieved clinical remission. All 59 patients with positive anti-PLA2R antibodies showed a decreasing trend in antibody levels after RTX treatment, and 55 (93.2%) of them achieved antibody clearance (<20 U/mL). Logistic regression analysis showed that a high anti-PLA2R antibody titer (OR=0.993, P=0.032) was an independent risk factor for nonremission. Adverse events occurred in 18 (22.2%) patients, of which 5 (6.2%) were serious adverse events, and none were malignant or otherwise fatal. Conclusion: RTX alone can effectively induce remission PMN and maintain stable renal function. It is recommended as the ﬁrst choice of treatment and is also Frontiers in Immunology 01 frontiersin.org Zhang et al. 10.3389/ﬁmmu.2023.1156470 10.3389/ﬁmmu.2023.1156470 effective in patients who relapse and have poor responses to conventional immunosuppressive therapy. Anti-PLA2R antibodies can be used as a marker for RTX treatment monitoring, and antibody clearance is necessary to achieve and improve the rates of clinical remission. primary membranous nephropathy, rituximab, clinical remission rate, anti-PLA2R antibody, safety Frontiers in Immunology 1 Introduction The Kidney Disease: Improving Global Outcomes (KDIGO) 2012 Clinical Practice Guideline for the Management of Glomerular Diseases recommended glucocorticoids combined with alkylating agents [cyclophosphamide (CTX) or azelaic acid phenylbutyrate] for immunosuppressive therapy (6), but in clinical practice, alkylating agents have demonstrated signiﬁcant toxic side effects, including myelosuppression, infection, gonadal suppression and an increased risk of tumour formation (7). Other preferred treatment options include calcineurin phosphatase inhibitors (CNIs), such as cyclosporine A or tacrolimus (8). However, recurrence rates of up to 40%-50% have been commonly reported for patients treated with this regimen (9), and patients may become treatment dependent; the risk of chronic nephrotoxicity should not be underestimated. Better alternatives (lower recurrence rate and higher safety) to conventional treatment regimens are urgently needed for PMN and should be further explored and pursued. Approximately 70-80% of PMN cases are mediated by autoantibodies against the M-type phospholipase A2 receptor (PLA2R) of glomerular podocytes, and an additional 3%-5% of cases are mediated by antibodies against thrombospondin type- 1 domain-containing 7A (THSD7A) (10–12). The pathogenesis of PMN mainly involves T cells secreting various cytokines, such as interleukins, that stimulate B-cell proliferation and activation, and effector B cells secreting speciﬁc autoantibodies that bind to PLA2R and THSD7A on the surface of pedunculated cells, forming immune complexes that deposit under the glomerular epithelium, damaging the ﬁltration barrier and triggering proteinuria (13). These signiﬁcant breakthroughs in the understanding of the disease suggest that PMN is an autoimmune disease that is 2.1 Study population Ninety-eight adult (>18 years old) patients treated at the Department of Nephrology at Shandong Provincial Hospital between 04/2020 and 12/2021 were selected. Seven patients with irregular doses of RTX and 10 patients with less than 12 months of follow-up were excluded; 81 PMN patients were ﬁnally enrolled in the retrospective study. Inclusion criteria: (1) IMN diagnosed by renal biopsy. (2) RTX alone chosen as the initial or alternative treatment. (3) Clinical manifestations of NS with proteinuria >3.5 g/24 h and Frontiers in Immunology 02 frontiersin.org Zhang et al. 10.3389/ﬁmmu.2023.1156470 serum albumin <30 g/L prior to RTX treatment. (4) Patients have a well-documented clinical and laboratory examination and have been assessed for potential malignancies, reviewed for pathogenic drugs, screened for hepatitis B/C virus, HIV, autoimmune diseases, etc. by history, physical examination and laboratory tests (serology, imaging, etc.) to exclude factors that may contribute to secondary membranous nephropathy. (19, 20). RTX was dissolved in 9% saline to a concentration of 1 mg/ mL and infused at an initial rate of 40 mL/h, which was then gradually increased to 200 mL/h according to the tolerance of each patient. To reduce the infusion response to RTX, patients received methylprednisolone 40 mg, dexamethasone 5 mg, and isoproterenol 25 mg prior to injection. (19, 20). RTX was dissolved in 9% saline to a concentration of 1 mg/ mL and infused at an initial rate of 40 mL/h, which was then gradually increased to 200 mL/h according to the tolerance of each patient. To reduce the infusion response to RTX, patients received methylprednisolone 40 mg, dexamethasone 5 mg, and isoproterenol 25 mg prior to injection. Follow-up was performed every 3 months, i.e., before RTX treatment and at months 3, 6, 9, and 12 after treatment for study follow-up. Each monitoring index included routine blood, routine urine, liver and kidney function, blood lipid glucose, 24-h urine protein, anti-PLA2R antibody level, and circulating B-cell quantity. We used a standard commercial ELISA (Euroimmune, Lubeck, Germany) to determine anti-PLA2R antibody titers, which were deﬁned as antibody positive when the titer was >20 U/mL. CD19+ B lymphocyte depletion was deﬁned as a concentration of <5 cells/mL. Adverse events associated with rituximab were documented during drug infusion and throughout the follow-up period. Subsequent follow-ups were conducted at 6-month intervals to record the patients’ remission and recurrence. 2.1 Study population p p y The patients were divided into three groups according to their treatment background prior to RTX treatment. The ﬁrst group was the initial therapy group, in which patients were not given any immunosuppressive therapy before receiving RTX. However, all had received symptomatic supportive treatment such as blood pressure control, reduction of urinary protein levels, angiotensin converting enzyme inhibitor or angiotensin receptor blocker therapy for ≥3 months and remained in persistent NS status. During this treatment, the blood pressure was maintained below 140/90 mmHg, the glomerular ﬁltration rate (eGFR) ≥40 ml/min/ 1.73 m2, or the 24-hour endogenous creatinine clearance > 40 ml/ min/1.73 m2. The second group was the conventional immunosuppressive therapy relapse group (relapse group), in which patients had achieved complete response (CR) or partial response (PR) after conventional glucocorticoid combined with immunosuppressive (CTX or CNI) regimens and had reoccurrence of proteinuria >3.5 g/24 h during drug reduction. The third group was the conventional immunosuppressive therapy ineffective group (ineffective group), in which patients for whom conventional glucocorticoid combined with immunosuppressive (CTX or CNI) regimens were ineffective (not achieving CR or PR) and who received RTX after discontinuation were enrolled. In this group of patients, one subgroup of patients remained in persistent NS status after ≥6 months of treatment with conventional induction remission regimens, and then suspended of immunosuppressive drugs already in use and administration of RTX. The other subgroup of patients achieved CR or PR after receiving conventional induction remission regimens, these patients later relapsed with NS and were given conventional induction remission therapy again for ≥6 months without achieving CR or PR, and then suspended of immunosuppressive drugs already in use and administration of RTX. For both dosing regimens, the decision to reinject with 375 mg/ m2 × 1-2doses or 1 g × 1-2doses was made at 6 months, depended on the extent of B-cell rebound, anti-PLA2R antibody levels and clinical remission, etc. Subsequent evaluations were repeated every six months or so to see if another injection was needed. 2.3 Treatment responses and renal outcomes The primary outcome was clinical remission at 12 months, and the secondary outcomes were safety and the occurrence of adverse events after medication. To assess treatment response, CR was deﬁned as proteinuria <0.3 g/24h on the premise of stable renal function (eGFR ≥45 ml/min/1.73 m2). PR was deﬁned as proteinuria <3.5 g/24h on the premise of stable renal function (eGFR ≥45 ml/min/1.73 m2), or a decrease in 24-h urinary protein quantiﬁcation ≥50% from the pretreatment, a ≥30% increase or normalization of serum albumin concentration, and stable or <30% increase in serum creatinine. Patients who did not meet these deﬁnitions were considered nonresponders, i.e., they did not achieve clinical remission. Relapse was deﬁned as the reoccurrence of 24-h urine protein quantiﬁcation >50% of baseline value or >3.5 g in patients who achieved CR or PR. The primary observed endpoint for renal outcomes was a deterioration in renal status or the occurrence of ESKD. Deterioration in renal status was deﬁned as a posttreatment increase in serum creatinine >133 mmol/L or a doubling of baseline serum creatinine levels lasting more than 3 months. ESKD was deﬁned as a creatinine clearance below 15 ml/min at the last follow-up, initiation of dialysis or renal transplantation. A serious adverse event was deﬁned as the occurrence of clinical death or the emergence of a serious pulmonary infection, pulmonary embolism, cerebral infarction, myocardial infarction, or the hospitalization of a patient as a result of an adverse event. This study was carried out in accordance with the Helsinki Declaration, and the study protocol was approved by the Ethics Review Committee of Shandong Provincial Hospital in China (JNKJ: NO. 2020-3028). 3.1 General baseline parameters Eighty-one patients with PMN were enrolled with a median age of 53.0 (35.5, 60.0) years, of whom 60 were male and 21 were female. At the time of enrolment, the median protein level was 6.3 (4.4, 10.8) g/24 h, serum albumin was 24.5 ± 5.4 g/L, serum creatinine was 73.6 (62.1, 87.9) mmol/L, and eGFR was 102.3 (74.8, 122.4) mL/ min/1.73 m2. All 81 patients were tested for anti-PLA2R antibodies in blood; the median level of anti-PLA2R antibodies was 61.0 (17.5, 147.8) U/ml, and 59 patients (72.8%) were tested positive for antibodies (>20 U/mL) (Table 1). Thirty-six patients were enrolled in the initial therapy group, 12 patients were enrolled in the relapse group, and 33 patients were enrolled in the ineffective group (Figure 1). Patients in the ineffective treatment group, compared with those in the initial therapy and relapse groups, had low levels of hemoglobin [115.0 (102.0, 137.0) vs. 135.0 (126.5, 145.0) vs. 138.5 (121.0, 152.0) g/L, P=0.001], globulin [22.0 ± 3.6 vs. 24.0 ± 3.2 vs. 24.2 ± 3.3 g/L, P=0.040] and absolute CD19 values [156.0 (103.8, 279.2) vs. 320.5 (198.0, 587.4) vs. 211.7 (143.2, 379.5)/ml, P=0.022], and had high proportion of eGFR <60 mL/min/1.73 m2 [6/33 (18.2%) vs. 2/36 (5.6%) vs. 2/12 (16.7%), P=0.249, difference not statistically signiﬁcant] (Table 1). The eGFR was >60 mL/min/1.73 m2 in 71 patients and 40-60 mL/min/1.73 m2 in 10 patients. The clinical remission rate was higher in patients with high levels of eGFR than in those with low levels (83.1% vs. 60.0%, P=0.086, difference not statistically signiﬁcant). There were 3 low-risk patients, 50 intermediate-risk patients and 28 high-risk patients. Fifty-seven (70.4%) patients received repeat RTX injections during follow-up after completion of full-dose (375 mg/m2×4 doses or 1 g×2 doses) RTX treatment regimens. All 59 patients with positive anti-PLA2R antibodies showed a decrease in antibody levels after RTX treatment, and 55 patients (93.2%) achieved antibody conversion to antibody negativity (<20 U/ mL), including 10 with levels <4 U/ml and 35 with levels <2 U/ml. Forty-four of 55 (80.0%) patients achieved clinical remission. Among the 4 patients with nonconverted antibody negativity (all B-cell levels <5/mL), 1 patient (25.0%) achieved clinical remission. The remission rates were signiﬁcantly different between antibody conversion and nonconversion (P=0.013). Of the 81 patients, the patients with high antibody titers (>150 U/mL) had signiﬁcantly lower clinical remission rates than patients with low titers (50% vs. 90.16%, P < 0.001). 3.1 General baseline parameters Seventy-one (87.7%) patients were positive for histological anti- PLA2R antibodies, and their clinical remission rate was 78.9% (15/ 71); the remission rate in antibody-negative patients was 90% (9/10) (P=0.408, no statistically signiﬁcant difference). 2.2 Treatment options and follow-up There were two dosing regimens in this study. In the ﬁrst, RTX was administered intravenously at 375 mg/m2 once a week for 4 weeks as a course of treatment. In the second dosing regimen, RTX was administered intravenously at 1 g/dose used at 2-week intervals for a total of 2 doses as a course of treatment. The 2021 KDIGO guidelines recommend both treatment regimens for use in patients with PMN (5). Previous studies have demonstrated no signiﬁcant difference in the proportion of CR or PR using the two regimens Frontiers in Immunology frontiersin.org 03 Zhang et al. 10.3389/ﬁmmu.2023.1156470 2.4 Statistical analysis patients in the initial therapy group achieved clinical remission, of whom 12 (33.3%) patients achieved CR and 20 (55.6%) patients achieved PR. Eleven of 12 (91.7%) patients in the conventional immunotherapy relapse group achieved clinical remission, of whom 3 (25.0%) patients achieved CR and 8 (66.8%) patients achieved PR. Twenty-two of 33 (66.7%) patients in the conventional immunotherapy ineffective group achieved clinical remission, of whom 6 (18.2%) patients achieved CR and 16 (48.5%) patients achieved PR (Table 2). Statistical analysis was performed using the statistical software SPSS 22.0. Normally distributed data are described as the means ± SD and were compared by independent t tests or one-way analysis of variance. Nonnormally distributed data are described as the median (interquartile range [IQR]) and were compared by the Mann−Whitney U test or Kruskal−Wallis test. Categorical variables are described as percentages, and Pearson chi-square tests were performed. All probabilities were two-tailed, and the level of signiﬁcance was set at 0.05. Logistic regression analysis was performed to conﬁrm the potential risk or protective factors for treatment response. The Kaplan−Meier method was used to compare the clinical remission of patients in different background groups after treatment with RTX. During the follow-up period, the clinical remission rate of the patients gradually increased with longer follow-up. The ineffective group had a lower remission rate than the initial therapy group (66.7% vs. 88.9%, P=0.025). The difference in clinical remission rates between the two groups was already signiﬁcant at 3 months (P=0.047) and was maintained throughout the study period (Figure 2). 3 Results During the 12-month monitoring period, all PMN patients showed an overall decreasing trend in anti-PLA2R antibodies, 24-h urine protein quantiﬁcation, and absolute CD19 values and an upward trend in serum albumin (Figure 3). The results at month 12 showed a decrease in anti-PLA2R antibodies from 61.0 (17.5, 147.8) U/mL to 2.0 (2.0, 3.7) U/mL in all PMN patients, with overall antibody levels becoming negative. The 24-hour urine protein levels dropped from 6.3 (4.4, 10.8) g/24 h to 1.2 (0.4, 2.5) g/24 h. Serum albumin gradually increased from 24.5 ± 5.4 g/L to 38.0 (33.0, 40.7) g/L, and the difference between the three groups was statistically signiﬁcant (P=0.033). The relapse group had a higher increase in albumin level than the initial therapy group (P=0.047) and the ineffective group (P=0.009) (Table 3). One of the 68 (1.5%) patients relapsed after achieving clinical remission, and two patients developed worsening renal function, both in the ineffective group. No patients progressed to ESKD. Frontiers in Immunology 3.4 Analysis of risk factors All patients completed at least 12 months of follow-up. At 12 months, 65 of 81 (80.2%) patients achieved CR (n=21, 25.9%) or PR (n=44, 54.3%) with rituximab treatment. Thirty-two of 36 (88.9%) Compared with patients who achieved clinical remission, nonresponders had higher levels of anti-PLA2R antibodies [204.5 Frontiers in Immunology frontiersin.org 04 frontiersin.org Zhang et al. 10.3389/ﬁmmu.2023.1156470 TABLE 1 Baseline characteristics of patients with PMN included in this study. Characteristic Total(n=81) Initial therapy (n=36) Relapse (n=12) Ineffective(n=33) P Male sex, n (%) 60(74.1) 25(69.4) 9(75.0) 26(78.8) 0.674 Age (years) 53.0(35.5, 60.0) 53.0(34.5, 63.0) 55.0(38.0, 60.8) 53.0(36.5, 56.0) 0.922 Urine RBC/ml 20.2(8.8, 38.5) 20.2(9.5, 35.4) 16.5(5.1, 29.4) 25.1(8.9, 66.1) 0.365 Proteinuria (g/24 h) 6.3(4.4, 10.8) 7.1(5.3, 10.9) 5.6(4.3, 7.4) 6.2(4.0, 12.3) 0.348 WBC (×109/L) 7.3(6.2, 9.9) 7.1(6.0, 8.7) 6.9(5.1, 9.7) 8.3(6.8, 11.2) 0.115 Hemoglobin (g/L) 133.0(112.0, 145.0) 135.0(126.5, 145.0) 138.5(121.0, 152.0) 115.0(102.0, 137.0) 0.001 Platelet (×109/L) 262.3 ± 62.2 261.4 ± 56.8 258.4 ± 63.3 264.6 ± 69.3 0.953 AST (u/L) 20.5(17.0, 28.0) 22.0(17.0, 28.0) 21.5(17.0, 27.0) 19.0(17.0, 29.0) 0.537 ALT (u/L) 20.0914.0, 25.23) 20.0(15.0, 28.0) 20.5(15.3, 26.3) 19.0(13.0, 25.0) 0.654 Total protein (g/L) 47.9(42.0, 53.2) 48.0(42.2, 53.8) 52.5(46.3, 57.4) 45.1(39.9, 53.3) 0.513 Albumin (g/L) 24.5 ± 5.4 23.9 ± 5.6 27.9 ± 5.6 23.9 ± 4.9 0.062 Globulin (g/L) 23.2 ± 3.5 24.0 ± 3.2 24.2 ± 3.3 22.0 ± 3.6 0.040 BUN (mmol/L) 6.5(5.0, 9.4) 6.0(5.0, 9.5) 5.6(4.9, 7.9) 8.0(6.1, 10.5) 0.484 Serum creatinine(mmol/L) 73.6(62.1, 87.9) 69.3(56.2, 81.4) 70.6(63.0, 104.4) 78.9(65.0, 103.1) 0.218 eGFR (mL/min/1.73 m2) a 102.3(74.8, 122.4) 107.0(88.9, 124.5) 109.5(66.5, 127.7) 93.0(62.5, 120.3) 0.156 GFR>60, n (%) 71(87.7) 34(94.4) 10(83.3) 27(81.8) 0.249 GFR 40–60, n (%) 10(12.3) 2(5.6) 2(16.7) 6(18.2) Cholesterol (mmol/L) 7.1(5.6, 9.2) 8.2(5.9, 9.7) 6.8(5.0, 9.3) 6.8(5.6, 8.8) 0.374 Anti-PLA2R antibody positivity, n (%) b 59(72.8) 26(72.2) 7(58.3) 26(78.8) 0.392 Anti-PLA2R antibodies( >150U/mL) 20(24.7) 9(25.0) 2(16.7) 9(27.3) 0.765 Anti-PLA2R antibodies (U/mL) 61.0(17.5, 147.8) 79.3(18.1, 149.3) 29.2(8.6, 82.9) 70.8(24.0, 182.5) 0.451 Absolute values of CD19(/ml) 215.7(120.2, 334.2) 320.5(198.0, 587.4) 211.7(143.2, 379.5) 156.0(103.8, 279.2) 0.022 low-risk, n (%) 3(3.7) 3(3.7) 0 0 intermediate-risk, n (%) 50(61.7) 21(25.9) 9(11.1) 20(24.7) 0.580 high-risk, n (%) 28(34.6) 12(14.8) 3(3.7) 13(16.0) 0.654 The hemoglobin level in the ineffective group was lower than that in the initial treatment group (P<0 001) and the relapse group (P=0 016) Patients in the initial treatment group had higher 3.5 Adverse events (39.5, 331.4) vs. 47.5 (11.5, 104.7) U/mL, P=0.010], total protein [42.9 (38.4, 51.2) vs. 48.6 (43.0, 54.2) g/L, P=0.019], globulin [21.2 ± 2.8 vs. 23.7 ± 3.5 g/L, P=0.012], and C3 [1.1 (1.0, 1.1) vs. 1.2 (1.1, 1.3) g/L, P=0.035] (Table 4). Univariate logistic regression analysis showed that anti-PLA2R antibody titer (OR=0.994, P=0.005), cholesterol (OR=0.807, P=0.040), and blood creatinine (OR=0.979, P=0.033) were risk factors for nonremission, whereas total protein (OR=1.104, P=0.026) and globulin (OR= 1.256, P=0.017) were protective factors, and a high anti-PLA2R antibody titer (OR=0.993, P=0.032) was an independent risk factor for nonremission (Table 5). Adverse events occurred in 18 (22.2%) participants during the study, and serious adverse events occurred in 5 (6.2%) patients (2 patients hospitalized for herpes zoster and 3 patients hospitalized for pulmonary infection). The most common adverse reactions were infusion reactions, including rash, erythema, pruritus, runny nose, and irritability, all of which resolved spontaneously after the infusion was completed. Only patients with severe herpes zoster and pulmonary infections received systemic therapy, and all patients made a full recovery (Table 6). Frontiers in Immunology frontiersin.org 05 Zhang et al. 10.3389/ﬁmmu.2023.1156470 FIGURE 1 Flow chart of selection for patients with PMN receiving rituximab therapy. A total of 81 PMN patients were enrolled, with 36 patients receiving rituximab as the initial therapy, 12 receiving RTX as a secondary therapy after relapse, and 33 receiving RTX as an alternative therapy after failure of other immunosuppressant therapy. During follow-up, 2 patients had worsening renal function, and 1 patient relapsed, both from the ineffective group. No patients entered ESKD. Flow chart of selection for patients with PMN receiving rituximab therapy. A total of 81 PMN patients were enrolled, with 36 patients receiving rituximab as the initial therapy, 12 receiving RTX as a secondary therapy after relapse, and 33 receiving RTX as an alternative therapy after failure of other immunosuppressant therapy. During follow-up, 2 patients had worsening renal function, and 1 patient relapsed, both from the ineffective group. No patients entered ESKD. The primary outcome was complete remission at 12 months. The overall remission rate was 80.2% (65/81). The initial therapy group had a higher remission rate than the ineffective group at 9 months (72.2%vs.48.5%, P=0.044) and at 12 months (88.9% vs. 66.7%, P=0.025). The relapse group had a higher remission rate than the ineffective group at 12 months (91.7%vs.66.7%, P=0.036). Values in bold represent P<0.05. 4 Discussion decreased signiﬁcantly or turned negative, renal function remained relatively stable, and no patients progressed to ESKD. The results of this study emphasize the necessity of antibody clearance to achieve and improve clinical remission. The 2021 KDIGO guidelines (5) began recommending RTX for management of PMN, and the efﬁcacy and safety have been conﬁrmed for clinical application. However, its therapeutic effects have been less researched in Asian populations, especially in the Chinese population. This study further conﬁrmed the therapeutic efﬁcacy and safety of RTX in the treatment of Chinese PMN patients in a retrospective analysis, which showed that most patients achieved clinical remission, anti-PLA2R antibody levels The results showed an overall remission rate of 80.2% at month 12, slightly higher than that in previous studies on the efﬁcacy of RTX (21–26). When rituximab was administered as an initial therapy, the clinical remission rate was 88.9%, which was better than the previous remission rate of 69.1% reported by Ruggenenti et al. (21), the 60% in the MENTOR study (22), the 62% in the RI- TABLE 2 Complete remission or composite (complete or partial remission) at 3–12 months by intention-to-treat analysis. Study Time Points No. of Patients with Remission/Total No. (%) P Total Initial therapy Relapse Ineffective Complete remission 3 months 1/81(1.2) 1/36(2.7) 0(0) 0(0) 0.531 6 months 5/81(6.2) 3/36(8.3) 1/12(8.3) 1/33(3.0) 0.622 9 months 12/81(14.8) 7/36(19.4) 3/12(25.0) 2/33(6.1) 0.165 12 months 21/81(25.9) 12/36(33.3) 3/12(25.0) 6/33(18.2) 0.365 Complete or partial remission 3 months 31/81(32.3) 17/36(47.2) 6/12(50.0) 8/33(24.4) 0.097 6 months 39/81(48.1) 21/36(58.3) 7/12(58.3) 11/33(33.3) 0.086 9 months 52/81(64.2) 26/36(72.2) 10/12(83.3) 16/33(48.5) 0.039 12 months 65/81(80.2) 32/36(88.9) 11/12(91.7) 22/33(66.7) 0.038 The primary outcome was complete remission at 12 months. The overall remission rate was 80.2% (65/81). The initial therapy group had a higher remission rate than the ineffective group at 9 months (72 2% s 48 5% P 0 044) and at 12 months (88 9% s 66 7% P 0 025) The relapse group had a higher remission rate than the ineffecti e group at 12 months (91 7% s 66 7% P 0 036) 06 Frontiers in Immunology frontiersin.org Zhang et al. 10.3389/ﬁmmu.2023.1156470 FIGURE 2 Kaplan–Meier estimates of the composite remission (complete or partial) in the initial therapy,relapse and ineffective group. The initial therapy group and the ineffective group have a P-value of 0.013. FIGURE 2 Kaplan–Meier estimates of the composite remission (complete or partial) in the initial therapy,relapse and ineffective group. 4 Discussion The initial therapy group and the ineffective group have a P-value of 0.013. CYCLO study (23), and the 64.9% at month 12 in the GEMRITUX study (24). Compared to the initial therapy group, the ineffective group had a lower clinical remission rate, with 66.7% of patients achieving remission, which was better than the remission rate of 50.0% reported by Ruggenenti et al. (25) and 41.7% reported in the Peking University First Hospital study (26). The remarkably greater remission rates in the initial therapy group of the current study compared with those of previous trials may be explained by several reasons. First, the inclusion of patients differed, with the current study enrolling patients with a relatively higher eGFR, relatively more low- and intermediate-risk patients, and relatively fewer high- risk patients. Second, Chinese patients have a smaller body surface area than Western patients, and for patients receiving a 1 g x 2 dose regimen, the same dose of drug is relatively more available in patients with a smaller body surface area, i.e., a longer half-life and longer duration of action. Additionally, the patients in this study all received a standard regimen of full-dose RTX (375 mg/m2 × 4 doses or 1 g × 2 doses) at the time of the ﬁrst course of administration; additionally, most (70.4%) of the patients were given a second course of RTX infusion after 6 months, and full-dose treatment may facilitate a better treatment response (27). In fact, the issue of the optimal RTX dose in PMN therapy remains somewhat controversial, with RTX doses varying widely from a single dose of 375 mg/m2 to four doses of 375 mg/m2 in different studies, but B C D E A FIGURE 3 Serial levels of albumin (A), proteinuria (B), anti-PLA2R antibody (C), the absolute values of CD19 (D) ana serum creatinine (A) after rituximab treatment in patients who had been followed up for 12 months. Data are presented as the medians (interquartile range) over time (B, D) or mean ± SD (A, C, E). B A B A C E D D C E E FIGURE 3 Serial levels of albumin (A), proteinuria (B), anti-PLA2R antibody (C), the absolute values of CD19 (D) ana serum creatinine (A) after rituximab treatment in patients who had been followed up for 12 months. Data are presented as the medians (interquartile range) over time (B, D) or mean ± SD (A, C, E). TABLE 4 Composite comparisons of clinical features of patients with PMN between responders and nonresponders. n=81. 4 Discussion 07 07 Frontiers in Immunology frontiersin.org Zhang et al. 10.3389/ﬁmmu.2023.1156470 TABLE 3 Clinical characteristics of PMN patients after rituximab treatment for 12 months. TABLE 3 Clinical characteristics of PMN patients after rituximab treatment for 12 months. TABLE 3 Clinical characteristics of PMN patients after rituximab treatment for 12 months. Total Initial therapy Relapse Ineffective P Time to reach remission (months) 6.6 ± 3.7 6.0 ± 3.7 6.0 ± 3.5 7.7 ± 3.8 0.127 Proteinuria (g/24 h) 1.2(0.4, 2.5) 0.8(0.2, 2.3) 0.6(0.2, 0.9) 1.9(0.5, 3.7) 0.224 Albumin (g/L) 38.0(33.0, 40.7) 38.8(33.7, 40.9) 41.0(39.0, 43.0) 36.6(29.4, 38.4) 0.033 Anti-PLA2R antibody positivity, n(%) 4(4.9) 1(2.8) 0(0) 3(9.1) 0.334 Anti-PLA2R antibodies (U/mL) 2.0(2.0, 2.9) 2.0(2.0, 2.0) 2.0(2.0, 3.6) 2.0(2.0, 5.7) 0.143 Serum creatinine (mmol/L) 78.4(62.1, 90.3) 72.3(53.0, 81.4) 71.5(60.0, 91.1) 85.9(74.9, 101.0) 0.072 Absolute values of CD19 (/ml) 41.7(1.8, 166.5) 41.7(2.8, 85.9) 4.7(0.6, 146.4) 54.5(2.2, 255.6) 0.230 The serum albumin level in the relapse group was higher than that in the initial treatment group (P=0.047) and the ineffective group (P=0.009). PLA2R, phospholipase A2 receptor. Values in bold represent P<0.05. patients in the ineffective group. The above could explain the low clinical remission rate in the ineffective group. The absolute CD19 values were signiﬁcantly lower in the ineffective and relapse groups than in the initial groups, and we considered that previous immunosuppression had an immunosuppressive effect on the patients, reducing their immunity, although remission of proteinuria was not achieved. the mainstream consensus suggests that adequate doses of RTX are more efﬁcacious. There is a lack of randomized controlled cohort studies of low-dose versus full-dose RTX infusion in China and abroad, and further exploration of this issue is warranted (28). Patients in the ineffective group had a lower overall level of hemoglobin, and a higher proportion of the patients had an eGFR <60 mL/min/1.73 m2 than those in the initial therapy and relapse groups, indicating a more severe degree of renal injury. Previous studies have shown that patients with tubulointerstitial lesions and renal impairment respond worse than patients with normal renal function. Additionally, the signiﬁcantly lower globulin and absolute CD19 values indicated the relatively low overall immunity of the Anti-PLA2R antibodies are key markers of PMN (13, 24, 29) that can be effectively cleared by RTX. 4 Discussion The outcomes of this study showed that antibody levels decreased signiﬁcantly in antibody- positive patients at 12 months after RTX treatment; additionally, patients achieving clinical remission had slightly lower antibody TABLE 4 Composite comparisons of clinical features of patients with PMN between responders and nonresponders. n=8 Responders a n=65 Nonresponders n=16 Pd Male sex, n (%) 46 (70.8) 14 (87.5) 0.171 Age (years) 53.0 (36.5, 60.0) 51(33.3, 57.3) 0.643 Urine RBC/HPF 19.4(8.6, 34.1) 34.8(9.9, 70.0) 0.129 Proteinuria (g/24 h) 6.2(4.4, 10.4) 7.7(4.6, 12.9) 0.243 Hemoglobin (g/L) 134(113.5, 145.0) 127.0(110.0, 141.0) 0.285 Total protein (g/L) 48.6(43.0, 54.2) 42.9(38.4, 51.2) 0.019 Albumin (g/L) 25.0 ± 5.2 22.6 ± 6.2 0.129 Globulin (g/L) 23.7 ± 3.5 21.2 ± 2.8 0.012 Serum creatinine (mmol/L) 72.7(59.7, 85.0) 79.0(63.6, 123.5) 0.133 eGFR (mL/min/1.73 m2) b 100.2 ± 25.6 93.8 ± 39.2 0.542 Cholesterol (mmol/L) 6.9(5.5, 8.8) 8.4(6.6, 11.2) 0.059 Anti-PLA2R antibody positivity, n (%) c 45(69.2) 14(87.5) 0.141 Anti-PLA2R antibodies (U/mL) 47.5(11.5, 104.7) 204.5(39.5, 331.4) 0.010 C3 (g/L) 1.2(1.1, 1.3) 1.1(1.0, 1.1) 0.035 Absolute values of CD19 (/ml) 217.5(119.2, 390.3) 210.5(123.5, 326.0) 0.535 eGFR, estimated glomerular ﬁltration rate; PLA2R, phospholipase A2 receptor; C3, complement. a. Responders are deﬁned as patients who achieve clinical remission. b. eGFR was calculated according to the Chronic Kidney Disease Epidemiology Collaboration equation. c. Anti-PLA2R positivity deﬁned by a value >20 RU/ml. d. Values in bold represent P<0.05. Pd 08 Frontiers in Immunology frontiersin.org Zhang et al. 10.3389/ﬁmmu.2023.1156470 TABLE 5 Risk factors for nonremission in patients with PMN receiving rituximab therapy (logistic regression). g y g g Univariate analysis Multivariate analysis OR (95% CI) P OR (95% CI) P Male sex 2.891(0.598, 13.968) 0.186 Age (years) 1.004(0.965, 1.043) 0.854 Proteinuria (g/24 h) 0.962(0.871, 1.062) 0.446 ALT 1.086(0.990, 1.192) 0.080 AST 1.013(0.969, 1.059) 0.563 Total protein (g/L) 1.104(1.012, 1.204) 0.026 0.051(0.834, 1.324) 0.676 Albumin (g/L) 1.088(0.975, 1.215) 0.132 Globulin (g/L) 1.256(1.041, 1.514) 0.017 1.268(0.834, 1.926) 0.266 Cholesterol (mmol/L) 0.807(0.658, 0.990) 0.040 0.853(0.603, 1.208) 0.371 Serum creatinine (mmol/L) 0.979(0.959, 0.998) 0.033 0.984(0.960, 1.009) 0.203 eGFR (mL/min/1.73 m2) a 1.008(0.989, 1.027) 0.425 Anti-PLA2R antibody positivity, n (%) b 3.111(0.646, 14.991) 0.157 Anti-PLA2R antibodies (U/mL) 0.994(0.990, 0.998) 0.005 0.993(0.986, 0.999) 0.032 Absolute values of CD19 (/ml) 1.002(0.999, 1.005) 0.308 ALT, glutamic pyruvic transaminase; AST, glutamic oxaloacetic transaminase; eGFR, estimated glomerular ﬁltration rate; PLA2R, phospholipase A2 receptor. a. eGFR was calculated according to the Chronic Kidney Disease Epidemiology Collaboration equation. b. 4 Discussion Anti-PLA2R positivity was deﬁned by a value >20 RU/ml. c. Values in bold represent P<0.05. elimination in obtaining clinical remission. Remuzzi et al. (4) monitored 132 PMN patients undergoing RTX treatment in terms of their anti-PLA2R antibodies and found that patients with high antibody titers had a lower probability of achieving CR or PR than patients with low titers. Ruggenenti et al. (21) found a correlation between anti-PLA2R antibody levels at baseline and the levels than nonresponders prior to RTX treatment, and the clinical remission rates were distinctly different between antibody-depleted and nondepleted patients. Clinical remission rates were signiﬁcantly lower in patients with high antibody titers (>150 U/mL) than in patients with low titers. These data suggest the effectiveness of RTX in clearing anti-PLA2R antibodies and the importance of antibody TABLE 6 Adverse events in all patients with PMN receiving rituximab. TABLE 6 Adverse events in all patients with PMN receiving rituximab. Events Patients (n) No. of events (n) Any adverse event 18 25 Serious adverse events 5 6 Fatal 0 0 Nonfatal 5 6 Fever, pulmonary infection 2 2 Herpes zoster 3 4 Nonserious adverse events 13 19 Infusion reactions* 6 9 Swelling of the limb on the infusion side with joint pain 2 3 Diarrhoea 1 2 Ileus 1 1 Nausea and dizziness with transient tinnitus 1 1 Flustered 1 2 Weight loss 1 1 Infusion reactions include rash, erythema, pruritus, runny nose, and irritability. p g Events Patients (n) No. of events (n) Any adverse event 18 25 Serious adverse events 5 6 Fatal 0 0 Nonfatal 5 6 Fever, pulmonary infection 2 2 Herpes zoster 3 4 Nonserious adverse events 13 19 Infusion reactions 6 9 Swelling of the limb on the infusion side with joint pain 2 3 Diarrhoea 1 2 Ileus 1 1 Nausea and dizziness with transient tinnitus 1 1 Flustered 1 2 Weight loss 1 1 *Infusion reactions include rash, erythema, pruritus, runny nose, and irritability. 09 frontiersin.org Zhang et al. 10.3389/ﬁmmu.2023.1156470 10.3389/ﬁmmu.2023.1156470 10.3389/ﬁmmu.2023.1156470 clinical course of patients on RTX therapy, with antibody titers rising during clinical activity and falling before clinical remission. Therefore, anti-PLA2R antibodies are a valid identiﬁer for monitoring the effect of RTX treatment (22). In conclusion, treatment with RTX alone for PMN results in a high clinical remission rate and has a relatively low impact on patients’ renal function. 4 Discussion RTX has a high safety proﬁle and is less prone to adverse events. RTX is recommended as the preferred treatment option, and demonstrated efﬁcacy in patients with PMN who have relapsed and are not effectively relieved when treated with conventional immunotherapy. Reasonable and standardized treatment application and regular monitoring of anti-PLA2R antibody levels during therapy can help to improve the effectiveness of treatment and reduce the incidence of adverse events. Logistic analysis of the elements that may inﬂuence the treatment outcome found that anti-PLA2R antibody titer (OR=0.994, P=0.005), cholesterol (OR=0.807, P=0.040) and creatinine (OR=0.979, P=0.033) were risk factors for nonremission, total protein (OR=1.104, P=0.026) and globulin (OR=1.256, P=0.017) were protective factors, and a high anti-PLA2R antibody titer (OR=0.993, P=0.032) was an independent risk factor for nonremission. Brand J et al. (7) similarly proposed that serum total protein levels could be a protective factor in the treatment of PMN. Nonresponders had lower total protein, globulin and albumin levels and higher anti-PLA2R antibody titers, total cholesterol and creatinine levels than patients who achieved clinical remission. These outcomes also indirectly reafﬁrm that patients with less proteinuria, higher albumin and globulin, and lower anti-PLA2R antibodies, total serum cholesterol, and creatinine (i.e., patients at low to intermediate risk) are more likely to experience remission, suggesting that appropriate threshold migration treatment may beneﬁt more patients and lead to earlier clinical remission. Funding This study was supported by the grants from the Primary Research & Development Plan of Shandong Province (2018GSF118227) and Shandong Natural Science Foundation (General Program) (ZR2022MH322). the Science and Technology Plan (673 and 741) of Shizhong District of Jinan City, Clinical Medical Science and Technology Development Plan of Jinan City, Shandong Province (202019186), and Horizontal issues of Shandong University (6020121011). The funders had no role in the design, data collection, analysis, interpretation, writing, or the decision of submission. Cravedi et al. (34) evaluated the costs required for RTX versus glucocorticoids combined with cyclophosphamide treatment for 6 months, and while RTX incurred higher expenses, the latter was associated with more adverse events. By taking the costs of treating adverse events into account, the total costs for glucocorticoids combined with cyclophosphamide may exceed those for RTX. Treatment with cyclosporine A may lead to a total cost of treatment that is greater than the cost of treatment with RTX. Hamilton et al. (35) showed that RTX treatment regimens were relatively inexpensive after 5 years of PMN management; furthermore, RTX treatment expenditures were relatively lower the longer the duration of treatment. Despite the relatively higher cost of single-dose RTX, it remains a cost-effective regimen in the medium- to long-term management of PMN. Thus, rituximab should be recommended as a ﬁrst-line treatment for patients with PMN, rather than as remedial therapy, from both an efﬁcacy and an economic perspective. Author contributions SZ collected and analyzed the data, authored and reviewed drafts of the paper. JH, JD and MS collected the data, prepared ﬁgures and tables. ZL and YS analyzed the data and reviewed drafts of the paper. BC conceived and designed the study, analyzed the data, authored and reviewed drafts of the paper. All authors contributed to the article and approved the submitted version. Data availability statement The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Ethics statement The studies involving human participants were reviewed and approved by the Ethics Review Committee of Shandong Provincial Hospital in China (JNKJ: NO. 2020-3028). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Regarding adverse events, RTX was well tolerated by most of the patients, with 22.2% of patients experiencing adverse events and only 6.2% experiencing serious adverse events. The rate of adverse events in our study was lower than that of previous relevant studies, which reported values of 50-80% (19), and other studies reported serious adverse event rates of 0-17% (30, 31). This discrepancy may be due to the retrospective nature of this study, which may have resulted in the omission of some minor adverse events. In addition, the most common type of adverse reaction in previous studies was infusion reactions (7), which was relatively infrequent in the present study; this may have been due to the use of anti-allergy medication prior to infusion and the limited rate of infusion to avoid or ameliorate infusion-related events to some extent (32, 33). For serious adverse events, relevant data were carefully veriﬁed in this study, and none of them were malignant or fatal. This study provided additional evidence suggesting that RTX is safe for treating PMN. Conﬂict of interest The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Frontiers in Immunology frontiersin.org 10 Zhang et al. 10.3389/ﬁmmu.2023.1156470 10.3389/ﬁmmu.2023.1156470 Publisher’s note organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated References Kidney Disease: Improving Global Outcomes (Kdigo) Glomerulonephritis Work Group. Kdigo clinical practice guideline for glomerulonephritis. Kidney Int Suppl (2012) 2(2):139–274. 7. van den Brand J, Ruggenenti P, Chianca A, Hofstra JM, Perna A, Ruggiero B, et al. Safety of rituximab compared with steroids and cyclophosphamide for idiopathic membranous nephropathy. J Am Soc Nephrol (2017) 28(9):2729–37. doi: 10.1681/ASN.2016091022 25. Ruggenenti P, Chiurchiu C, Abbate M, Perna A, Cravedi P, Bontempelli M, et al. Rituximab for idiopathic membranous nephropathy: who can beneﬁt? Clin J Am Soc Nephrol (2006) 1(4):738–48. doi: 10.2215/CJN.01080905 8. Wu L, Lai J, Ling Y, Weng Y, Zhou S, Wu S, et al. A review of the current practice of diagnosis and treatment of idiopathic membranous nephropathy in China. Med Sci Monit (2021) 27:e930097. doi: 10.12659/MSM.930097 26. Wang X, Cui Z, Zhang YM, Qu Z, Wang F, Meng LQ, et al. Rituximab for non- responsive idiopathic membranous nephropathy in a Chinese cohort. Nephrol Dial Transplant (2018) 33(9):1558–63. doi: 10.1093/ndt/gfx295 27. Seitz-Polski B, Dahan K, Debiec H, Rousseau A, Andreani M, Zaghrini C, et al. High-dose rituximab and early remission in Pla2r1-related membranous nephropathy. Clin J Am Soc Nephrol (2019) 14(8):1173–82. doi: 10.2215/CJN.11791018 9. Alfaadhel T, Cattran D. Management of membranous nephropathy in Western countries. Kidney Dis (Basel) (2015) 1(2):126–37. doi: 10.1159/000437287 10. Tomas NM, Beck LHJr., Meyer-Schwesinger C, Seitz-Polski B, Ma H, Zahner G, et al. Thrombospondin type-1 domain-containing 7a in idiopathic membranous nephropathy. N Engl J Med (2014) 371(24):2277–87. doi: 10.1056/NEJMoa1409354 28. Moroni G, Depetri F, Del Vecchio L, Gallelli B, Rafﬁotta F, Giglio E, et al. Low- dose rituximab is poorly effective in patients with primary membranous nephropathy. Nephrol Dial Transplant (2017) 32(10):1691–6. doi: 10.1093/ndt/gfw251 11. Beck LHJr., Bonegio RG, Lambeau G, Beck DM, Powell DW, Cummins TD, et al. M-type phospholipase A2 receptor as target antigen in idiopathic membranous nephropathy. N Engl J Med (2009) 361(1):11–21. doi: 10.1056/NEJMoa0810457 29. Hoxha E, Thiele I, Zahner G, Panzer U, Harendza S, Stahl RA. Phospholipase A2 receptor autoantibodies and clinical outcome in patients with primary membranous nephropathy. J Am Soc Nephrol (2014) 25(6):1357–66. doi: 10.1681/ASN.2013040430 12. Allinovi M, Lugli G, Rossi F, Palterer B, Almerigogna F, Caroti L, et al. Accuracy of serum Pla2r antibody detected by indirect immunoﬂuorescence in diagnosing biopsy- proven primary membranous nephropathy: a single-center experience and a systematic review of the literature. J Nephrol (2023) 36(2):281–3. doi: 10.1007/s40620-022-01528-1 30. References 19. Fervenza FC, Cosio FG, Erickson SB, Specks U, Herzenberg AM, Dillon JJ, et al. Rituximab treatment of idiopathic membranous nephropathy. Kidney Int (2008) 73 (1):117–25. doi: 10.1038/sj.ki.5002628 1. Schieppati A, Mosconi L, Perna A, Mecca G, Bertani T, Garattini S, et al. Prognosis of untreated patients with idiopathic membranous nephropathy. New Engl J Med (1993) 329(2):85–9. doi: 10.1056/NEJM199307083290203 20. Ruggenenti P, Chiurchiu C, Brusegan V, Abbate M, Perna A, Filippi C, et al. Rituximab in idiopathic membranous nephropathy: a one-year prospective study. J Am Soc Nephrol (2003) 14(7):1851–7. doi: 10.1097/01.asn.0000071511.35221.b3 2. Glassock R. Diagnosis and natural course of membranous nephropathy. Semin Nephrol (2003) 23(4):324–32. doi: 10.1016/s0270-9295(03)00049-4 3. Polanco N, Gutierrez E, Covarsi A, Ariza F, Carreno A, Vigil A, et al. Spontaneous remission of nephrotic syndrome in idiopathic membranous nephropathy. J Am Soc Nephrol (2010) 21(4):697–704. doi: 10.1681/ASN.2009080861 3. Polanco N, Gutierrez E, Covarsi A, Ariza F, Carreno A, Vigil A, et al. Spontaneous remission of nephrotic syndrome in idiopathic membranous nephropathy. J Am Soc Nephrol (2010) 21(4):697–704. doi: 10.1681/ASN.2009080861 21. Ruggenenti P, Cravedi P, Chianca A, Perna A, Ruggiero B, Gaspari F, et al. Rituximab in idiopathic membranous nephropathy. J Am Soc Nephrol (2012) 23 (8):1416–25. doi: 10.1681/ASN.2012020181 4. Ruggenenti P, Debiec H, Ruggiero B, Chianca A, Pelle T, Gaspari F, et al. Anti- phospholipase A2 receptor antibody titer predicts post-rituximab outcome of membranous nephropathy. J Am Soc Nephrol (2015) 26(10):2545–58. doi: 10.1681/ ASN.2014070640 22. Fervenza FC, Appel GB, Barbour SJ, Rovin BH, Lafayette RA, Aslam N, et al. Rituximab or cyclosporine in the treatment of membranous nephropathy. N Engl J Med (2019) 381(1):36–46. doi: 10.1056/NEJMoa1814427 5. Kidney Disease: Improving Global Outcomes Glomerular Diseases Work G. Kdigo 2021 clinical practice guideline for the management of glomerular diseases. Kidney Int (2021) 100(4S):S1–S276. doi: 10.1016/j.kint.2021.05.021 23. Scolari F, Delbarba E, Santoro D, Gesualdo L, Pani A, Dallera N, et al. Rituximab or cyclophosphamide in the treatment of membranous nephropathy: the ri-cyclo randomized trial. J Am Soc Nephrol (2021). doi: 10.1681/ASN.2020071091 5. Kidney Disease: Improving Global Outcomes Glomerular Diseases Work G. Kdigo 2021 clinical practice guideline for the management of glomerular diseases. Kidney Int (2021) 100(4S):S1–S276. doi: 10.1016/j.kint.2021.05.021 24. Dahan K, Debiec H, Plaisier E, Cachanado M, Rousseau A, Wakselman L, et al. Rituximab for severe membranous nephropathy: a 6-month trial with extended follow- up. J Am Soc Nephrol (2017) 28(1):348–58. doi: 10.1681/ASN.2016040449 6. References Fervenza FC, Abraham RS, Erickson SB, Irazabal MV, Eirin A, Specks U, et al. Rituximab therapy in idiopathic membranous nephropathy: a 2-year study. Clin J Am Soc Nephrol (2010) 5(12):2188–98. doi: 10.2215/CJN.05080610 31. Cravedi P, Ruggenenti P, Sghirlanzoni MC, Remuzzi G. Titrating rituximab to circulating b cells to optimize lymphocytolytic therapy in idiopathic membranous nephropathy. Clin J Am Soc Nephrol (2007) 2(5):932–7. doi: 10.2215/CJN.01180307 13. Ruggenenti P, Fervenza FC, Remuzzi G. Treatment of membranous nephropathy: time for a paradigm shift. Nat Rev Nephrol (2017) 13(9):563–79. doi: 10.1038/nrneph.2017.92 14. Reff KC ME, Chambers KS, Chinn PC, Leonard JE, Raab R, Newman RA, et al. Depletion of b cells in vivo by a chimeric mouse human monoclonal antibody to Cd20. Blood (1994) 83(2):435–45. doi: 10.1182/blood.V83.2.435.435 32. Courville J, Nastoupil L, Kaila N, Kelton J, Zhang J, Alcasid A, et al. Factors inﬂuencing infusion-related reactions following dosing of reference rituximab and pf- 05280586, a rituximab biosimilar. BioDrugs (2021) 35(4):459–68. doi: 10.1007/s40259- 021-00487-6 15. Kattah AG FF. Rituximab: emerging treatment strategies of immune mediated glomerular disease. Expert Rev Clin Immunol (2012) 8(5):413–21. doi: 10.1586/eci.12.26 33. Fouda GE, Bavbek S. Rituximab hypersensitivity: from clinical presentation to management. Front Pharmacol (2020) 11:572863. doi: 10.3389/fphar.2020.572863 16. Ponticelli C, Moroni G. Rituximab or cyclosporine for membranous nephropathy. New Engl J Med (2019) 381(17):1688–9. doi: 10.1056/NEJMc1910393 34. Hamilton P, Kanigicherla D, Venning M, Brenchley P, Meads D. Rituximab versus the modiﬁed ponticelli regimen in the treatment of primary membranous nephropathy: a health economic model. Nephrol Dial Transplant (2018) 33(12):2145– 55. doi: 10.1093/ndt/gfy049 17. Shi Y, Zhang Q, Han X, Qin Y, Ke X, Su H, et al. Phase 1 studies comparing safety, tolerability, pharmacokinetics and pharmacodynamics of Hlx01 (a rituximab biosimilar) to reference rituximab in Chinese patients with Cd20-positive b-cell lymphoma. Chin J Cancer Res (2021) 33(3):405–16. doi: 10.21147/j.issn.1000-9604.2021.03.11 18. Shi Y, Song Y, Qin Y, Zhang Q, Han X, Hong X, et al. A phase 3 study of rituximab biosimilar Hlx01 in patients with diffuse Large b-cell lymphoma. J Hematol Oncol (2020) 13(1):38. doi: 10.1186/s13045-020-00871-9 35. Cravedi P, Remuzzi G, Ruggenenti P. Rituximab in primary membranous nephropathy: ﬁrst-line therapy, why not? Nephron Clin Pract (2014) 128(3-4):261–9. doi: 10.1159/000368589 11 Frontiers in Immunology frontiersin.org
https://openalex.org/W2519750373	https://www.nature.com/articles/srep33086.pdf	English	null	The dynamic nature of crystal growth in pores	Scientific reports	2,016	cc-by	7,629	The dynamic nature of crystal growth in pores N Jose R. A. Godinho1,2,3, Kirill M. Gerke4,5,6, Andrew G. Stack3 & Peter D. Lee1,2 The kinetics of crystal growth in porous media controls a variety of natural processes such as ore genesis and crystallization induced fracturing that can trigger earthquakes and weathering, as well as, sequestration of CO2 and toxic metals into geological formations. Progress on understanding those processes has been limited by experimental difficulties of dynamically studying the reactive surface area and permeability during pore occlusion. Here, we show that these variables cause a time-dependency of barite growth rates in microporous silica. The rate is approximately constant and similar to that observed on free surfaces if fast flow velocities predominate and if the time-dependent reactive surface area is accounted for. As the narrower flow paths clog, local flow velocities decrease, which causes the progressive slowing of growth rates. We conclude that mineral growth in a microporous media can be estimated based on free surface studies when a) the growth rate is normalized to the time-dependent surface area of the growing crystals, and b) the local flow velocities are above the limit at which growth is transport-limited. Accounting for the dynamic relation between microstructure, flow velocity and growth rate is shown to be crucial towards understanding and predicting precipitation in porous rocks. received: 18 April 2016 accepted: 16 August 2016 Published: 12 September 2016 Understanding mineral reactions in porous geological formations is currently a subject of intense study in Earth sciences as it underpins some of the most challenging environmental problems of modern society. Striking exam- ples include the possibility of permanently sequestering CO2 as carbonate minerals in deep sandstone formations and in ultramafic rocks1–3, and mineral sequestration of toxic metals into the bedrock, i.e., contaminants resulting from the unconventional extraction of natural gas and oil by hydraulic fracturing4,5. Additionally, studying the effect of microstructure and the spatial distribution of precipitates within a porous rock is essential to understand the formation of ore deposits, earthquakes induced by crystallization, as well as the degradation of stone building heritage and metasomatism6–9. g Mineral reactions depend, among other factors, on the reactive surface area and ion transport at the mineral-fluid interface, which in porous media depends largely on the permeability of the porous network10. For example, it has been demonstrated that the dissolution rates of magnesite within a porous column depend on the local microstructure, permeability and reactive surface area11,12. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports 1School of Materials, The University of Manchester, M13 9PL, Manchester, UK. 2Research Complex at Harwell, Rutherford Appleton Laboratory, OX11 0FA, Harwell, UK. 3Chemical Sciences Division, Oak Ridge National Laboratory, PO Box 2008, MS-6110, Oak Ridge, TN 37831 USA. 4The University of Melbourne, Department of Infrastructure Engineering, Parkville, VIC, 3010, Australia. 5CSIRO Land and Water, Glen Osmond, PB2, SA 5064, Australia. 6Institute of Physics of the Earth of Russian Academy of Sciences, Bolshaya Gruzinskaya 10, Moscow, 107031, Russia. Correspondence and requests for materials should be addressed to J.R.A.G. (email: jose.godinho@ manchester.ac.uk) received: 18 April 2016 accepted: 16 August 2016 Published: 12 September 2016 Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 The dynamic nature of crystal growth in pores N Dashed line (X), Continuous line and Dotted line (o) correspond to rates calculated using equation (2), measured on free surfaces for the same solution composition33 or calculated using the middle term of equation (1) (using Apore and (ΔV/Δt) measured here), respectively. See videos 1–2 for a full 3D reconstruction of the column. extraction efficiency and increased production costs4,20. Radioactivity from radium and strontium co-precipitated with barite, is a common hazard to oil platform workers and to communities living close to hydraulic fracturing sites21,22. The ability of barite to sequester hazardous metals can reduce radium mobility in the subsurface23–26, for example, to create a reactive barrier as part of a spent nuclear fuel disposal strategy, or for removing strontium and radium from fluids resulting from hydraulic fracturing5. Therefore, it is imperative that the kinetics of barite growth in porous media is better studied, hence enabling an enhancement in the recovery of hydrocarbons, whilst at the same time reducing its environmental impact. X-ray computed micro-tomography (XCT) has emerged in recent years as a technique capable of quantifying mineral reactions in porous media non-destructively27–32. The high flux and high signal to noise ratio of synchrotron X-ray based imaging allow fast scans of a sample, enabling new possibilities of quantifying the changing mineral phases of a porous structure with submicron resolution as a function of time. Using the powerful combination of 4D (3D + time) synchrotron based XCT, also named sCT, and numerical simulation of flow velocities, we study barite precipitation in a micro porous silica structure under continuous flow. The evolution of the 3D microstructure as a function of time allowed us to 1) test a dynamic model for surface area normalization of growth rates in porous media using input from sCT and an empirical equation for surface area variation as a function of time obtained for barite growth on a free surface; 2) study the growth rate variation as a function of time based on our dynamic model, which is compared with an existing static model and with growth on free surfaces; 3) relate the time-dependency of the growth rate with the changes of flow velocities, which are caused by the occlusion of pores. The dynamic nature of crystal growth in pores N These are dynamic properties that vary as the reaction modifies the pore microstructure, thus causing the reaction rates to be time-dependent13. For example, during mineral precipitation in a porous structure permeability decreases, which affects transport throughout the structure13. This time-dependency has been proposed to contribute to a key fundamental problem of modern geochemistry: the discrepancy of several orders of magnitude between reaction rates observed in the subsurface and those measured in laboratory experiments14.hi y p The growth rate of a mineral phase is easier to study on free surfaces (defined here as growth without spatial restrictions) than in porous media where the range of experimental techniques that can be used in situ is limited. Due to the difficulties of experimentally studying the evolution of porous microstructures during growth, reactive transport modeling has been the only tool capable of providing information about the dynamics of how crystals fill the pores15–18. Still, available modelling strategies require validation against experiments that track the reaction both temporally and spatially.t p y p y Barite often precipitates during oil and gas extraction due to the mixing of sulfate rich surface water with barium rich ground water4,19. The precipitates can reduce the permeability of the reservoir, causing a decrease in Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 1 www.nature.com/scientificreports/ Figure 1. Evolution of crystals and growth rates. (a) 3D subvolume at the beginning and at the end of the experiment; (b) 2D cross section (not segmented) at the end of the experiment, where barite is the brightest fraction; (c) variation of the growth rate as a function of time. Dashed line (X), Continuous line and Dotted line (o) correspond to rates calculated using equation (2), measured on free surfaces for the same solution composition33 or calculated using the middle term of equation (1) (using Apore and (ΔV/Δt) measured here), respectively. See videos 1–2 for a full 3D reconstruction of the column. Figure 1. Evolution of crystals and growth rates. (a) 3D subvolume at the beginning and at the end of the experiment; (b) 2D cross section (not segmented) at the end of the experiment, where barite is the brightest fraction; (c) variation of the growth rate as a function of time. The dynamic nature of crystal growth in pores N In summary, we study the link between the dynamic properties of a pore microstructure, such as surface area, flow velocity and permeability, and how their variation can cause important discrepancies between the growth rates measured in a porous structure and on a free surface. Results and Discussion Surface area normalization. × − A t t t ( ) 1 11 10 1 06 10 5 2 2 (4) Equation (3) is based on the hypothesis that a) Rfree = Rcrystal, thus the growth rate is not affected by the properties of the pore structure, b) on glass beads and for the solution composition and temperature of this experiment the nucleation density is constant (N = 0.29 crystals/μm2)29; c) the fluid composition is approximately homogeneous throughout the length of the column, thus the growth rate is the same for every crystal. The value of Rfree = 0.78 mmol/m2.h was obtained experimentally during growth of barite crystals on free surfaces in a solution with the same composition used here and was normalized to the surface area of the growing crystals29. Growth rates as a function of time. Barite crystals nucleate on the surface of SiO2 particles growing throughout the length of the experiment (Video 1), which causes a continuous alteration of the pore microstruc- ture (Supplementary Fig. S3). After 13.5 hours, 18% of the initial pore space is filled with barite. At the end of the experiment barite crystals can be found throughout the entire pore space (Fig. 1, Video 2) without significant var- iation between the center and the outer edges of the column, and between the inlet and the outlet (Supplementary Fig. S4). This suggests that at the millimeter scale, growth rates are similar throughout the length of the column. Therefore, we conclude that the ions consumed by crystal growth within the fluid residence time inside the pore network is insufficient to affect the growth rates, which is necessary to validate equation (3). Homogeneous nucle- ation in solution was not observed as expected29. p Next, we establish the link between the growth rate previously measured on free surfaces29 with the growth rates derived from this experiment (graphic in Fig. 1), either normalized to the initial pore surface area, Rpore, traditional static model using the middle term in equation (1); or normalized to the surface area of the crystals quantified statistically, Acrystal(t), dynamic method using equation (2). Four stages are identified. Stage 1 (initial 4 hours), barite crystals are too small to be accurately distinguished from noise. This is expected for a 1.24 μm voxel size and the growth rate of barite on free surfaces29. Results and Discussion Surface area normalization. Stage 2 (4–8.5 hours), Rcrystal (dashed line, X) and Rfree (continuous line) are similar and approximately constant. In contrast, Rpore (dotted line, o) oscillates over 100% of its initial value and is significantly higher than Rcrystal. Stage 3 (8.5–11 hours), Rcrystal progressively decreases down to five times lower than Rfree. Stage 4 (11–13.5 hours), Rcrystal remains approximately constant, whereby Rcrystal = 0.14 mmol/m2.h. crystal Results from Fig. 1 suggest that the method proposed to calculate the time-dependent reactive surface area is valid until the end of Stage 2, thus validating the hypothesis that the growth kinetics on free surfaces can be applied to predict the initial stages of growth in a permeable porous structure. In contrast, using the traditional method of normalizing the growth rates to the static value of surface area of the pore structure causes a significant discrepancy between Rpore and Rfree (see in Supplementary Figure S5 how this can affect the predictive calculation of crystal volumes precipitated). Similar to what has been shown for dissolution12, we suggest that this discrep- ancy can contribute to the wider geochemical problem that mineral reaction rates observed in nature do not agree with those measured in laboratory by several orders of magnitude14. We also suggest that more accurate methods of calculating Atotal(t) could be developed in future experiments to further refine the proposed dynamic method. For example, higher resolution sCT could be used to measure real values of Atotal(t). Nevertheless, using this dynamic method of surface area calculation, we demonstrate that Rfree = Rcrystal until effects inherent to growth within a pore structure become relevant to the growth kinetics, thus invalidating equation (3) (Stages 3–4). These effects are a consequence of microstructural changes in the pore network during precipitation (Supplementary Fig. S3 and Supplementary Table S1). Flow velocities. To explain the decrease of the growth rate during Stage 3, i.e. the divergent behavior between Rfree and Rcrystal, the flow velocities, permeability and Peclet number (Pe, the ratio between diffusive and advective time scales, see Supplementary Information) were calculated at five times between the beginning and the end of the experiment (Table 1 and see Supplementary Information for details on the flow simulations). The decrease in permeability, during pore occlusion with barite, is more significant during Stage 3. Results and Discussion Surface area normalization. The ability to analyze the transient stages of growth and to compare Rcrystal and Rfree is an essential progress towards the understanding of the fundamental mechanisms affecting crystal growth in porous media, which can enable the development of dynamic kinetic models that cannot be derived from traditional experimental approaches. = ∆ ∆ . R V t A t 1 ( ) (2) crystal crystal (2) Our method of estimating Acrystal (t) (units of L2) consists in multiplying the time-dependent surface area of single crystals (A(t), units of nm2/crystal) by the estimated total number of crystals in the entire pore structure, equation (3). A(t) is determined empirically, equation (4), t in hours. The estimated total number of crystals corresponds to the nucleation density, N (crystals/μm2), multiplied by the total pore surface area, Apore (units of L2). To account for the specific situations where a crystal completely fills a pore space or laterally overlaps with an adjacent crystal, and thus cannot grow more in a specific direction; Apore(t) is calculated for every scan using input from sCT by subtracting the surface area of pore-throats that are filled with crystals from Apore (Supplementary Fig. S1). Therefore, the area calculated using equation (3) only accounts for the surface area of crystals along the directions not restricted in space that are inherent to growth in porous media. = × × A t A t N A t ( ) ( ) ( ) (3) crystal pore = . × − . × − A t t t ( ) 1 11 10 1 06 10 (4) 5 2 2 = × × A t A t N A t ( ) ( ) ( ) (3) crystal pore = × × A t A t N A t ( ) ( ) ( ) (3) crystal pore = × × A t A t N A t ( ) ( ) ( ) crystal pore (3) = . × − . × − A t t t ( ) 1 11 10 1 06 10 (4) 5 2 2 = . × − . Results and Discussion Surface area normalization. Surface area normalization. To facilitate the understanding of the growth kinetics in porous media it is imperative to establish a method capable of linking it to the kinetics of growth on free surfaces. A major imped- iment to compare effectively the growth rates of a mineral phase inside a pore structure and on free surfaces is the necessity to normalize the rate to an equivalent reactive surface area. Growth rates measured on free surfaces are traditionally normalized using the surface area of the growing crystals, which can be measured experimen- tally33,34. However, available experimental methods do not allow the direct measurement of the surface area of crystals in pores as a function of time. Consequently, growth rates in a porous rock (Rpore) are traditionally nor- malized to the total surface area of the rock, which is either measured at the beginning of the experiment29 or calculated geometrically35. However, the surface area of crystals (Acrystal) vary with time and it’s reactivity is greater than the substrate surface (Apore). Therefore, for the same volume precipitated (ΔV) during a time interval (Δt), the estimate of growth rate in porous media (Rpore) and on free surfaces (Rfree) differ, depending on which surface area is used in the normalization, equation (1). ∆ ∆ = . = . V t R A R A (1) pore pore free crystal (1) here, we use sCT to quantify in situ barite precipitation (ΔV/Δt) in a microporous silica structure under continu- ous flow during 13.5 hours (Fig. 1). Even using sCT, the small crystal size restricts the accuracy at which Acrystal and its time-dependency can be measured directly36,37. Therefore, we developed a novel dynamic method of estimating the time-dependent surface area (Acrystal(t)) of barite crystals using input from sCT and based on empirical equa- tions used to calculate the surface area of barite crystals grown on free surfaces33. The result is a time-dependent growth rate in porous media (Rcrystal as defined in equation (2)) that can be directly compared to Rfree. Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 2 www.nature.com/scientificreports/ The ability to analyze the transient stages of growth and to compare Rcrystal and Rfree is an essential progress towards the understanding of the fundamental mechanisms affecting crystal growth in porous media, which can enable the development of dynamic kinetic models that cannot be derived from traditional experimental approaches. Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 Results and Discussion Surface area normalization. These regions contrast with areas where flow is slow throughout the experiment (dot box) and fewer crystals are observed. This could be because in areas with slow flow the nucleation density is lower than expected or the growth rate is too slow to form crystals large enough to be detected within the resolution of this experiment. These regions represent <2% of the overall pore surface area. Consequently, since our dynamic surface area model assumes homogeneous nucleation density and growth rate, the estimated surface area is slightly overestimated, which ultimately results in a small overestimation of the growth rate. This error can contribute to the oscillation of calculated growth rates during Stage 2, however, within the experimental variability the error is not sufficiently large to invalidate equation 3.l transport-controlled growth kinetics. Nevertheless, the final Pe is 6000, three orders of magnitude higher than the value at which growth rates should be considered limited by transport (Pe ~ 1)38. Therefore, ion transport in the largest pores, which remain open and interconnected throughout the experiment (Videos 2 and 3), is expected to be sufficient to maintain the overall solution composition constant over the length of the column (Pe > 100)38.hlli The flow simulations revealed that at the pore scale flow velocities can diverge significantly, which is linked to the local and overall growth rates (Fig. 2). A larger density of crystals is found in pore-throats where the flow velocities remain fast throughout the duration of the experiment (continuous box). Lower crystal densities are found in regions where the flow velocities are initially fast but decrease with time (dashed box). These regions contrast with areas where flow is slow throughout the experiment (dot box) and fewer crystals are observed. This could be because in areas with slow flow the nucleation density is lower than expected or the growth rate is too slow to form crystals large enough to be detected within the resolution of this experiment. These regions represent <2% of the overall pore surface area. Consequently, since our dynamic surface area model assumes homogeneous nucleation density and growth rate, the estimated surface area is slightly overestimated, which ultimately results in a small overestimation of the growth rate. Results and Discussion Surface area normalization. Distribution of flow velocities in a vertical cross section at the center of the column before h d ft 13 h SiO i l d i d b i i hi C l l f Figure 2. Distribution of flow velocities in a vertical cross section at the center of the column before growth and after 13.5 hours. SiO2 particles are represented in grey and barite in white. Color scale refers to flow velocities. Lines enclose areas where flow: remains fast during the entire experiment, whereby crystal density is high (continuous line); is fast at the beginning and slow after 13.5 hours, in contrast crystal density is low (dashed line); is slow at the beginning and after 13.5 hours, few crystals are observed (dotted line). The graphic shows the 3D statistical spatial relation between the flow velocities and barite precipitation. Figure 2. Distribution of flow velocities in a vertical cross section at the center of the column before growth and after 13.5 hours. SiO2 particles are represented in grey and barite in white. Color scale refers to flow velocities. Lines enclose areas where flow: remains fast during the entire experiment, whereby crystal density is high (continuous line); is fast at the beginning and slow after 13.5 hours, in contrast crystal density is low (dashed line); is slow at the beginning and after 13.5 hours, few crystals are observed (dotted line). The graphic shows the 3D statistical spatial relation between the flow velocities and barite precipitation. transport-controlled growth kinetics. Nevertheless, the final Pe is 6000, three orders of magnitude higher than the value at which growth rates should be considered limited by transport (Pe ~ 1)38. Therefore, ion transport in the largest pores, which remain open and interconnected throughout the experiment (Videos 2 and 3), is expected to be sufficient to maintain the overall solution composition constant over the length of the column (Pe > 100)38. The flow simulations revealed that at the pore scale flow velocities can diverge significantly, which is linked to the local and overall growth rates (Fig. 2). A larger density of crystals is found in pore-throats where the flow velocities remain fast throughout the duration of the experiment (continuous box). Lower crystal densities are found in regions where the flow velocities are initially fast but decrease with time (dashed box). Results and Discussion Surface area normalization. This is unexpected since dur- ing Stage 2 crystals grow faster than during Stage 3, which suggests that during Stage 3 a more significant occlu- sion of flow paths occurs, and consequently, the growth rates decrease. The decrease of Pe indicates a shift towards Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 3 www.nature.com/scientificreports/ Time (hours) 0 5.5 8.5 11 13.5 Peclet number (×103) 22.8 18.4 15.6 12.9 6.0 Permeability (×10−4) 5.49 4.50 3.88 2.33 1.47 Maximum velocity (μm/s) 712.2 700.9 695.6 653.0 611.1 Table 1. Properties of the pore structure after 0, 5.5, 8.5, 11 and 13.5 hours. Data corresponds to the 900 pixel side cube where flow velocities were simulated. Other statistics of the pore structure can be found in Supplementary Table S1. Time (hours) 0 5.5 8.5 11 13.5 Peclet number (×103) 22.8 18.4 15.6 12.9 6.0 Permeability (×10−4) 5.49 4.50 3.88 2.33 1.47 Maximum velocity (μm/s) 712.2 700.9 695.6 653.0 611.1 Table 1. Properties of the pore structure after 0, 5.5, 8.5, 11 and 13.5 hours. Data corresponds to the 900 pixel side cube where flow velocities were simulated. Other statistics of the pore structure can be found in Supplementary Table S1. Table 1. Properties of the pore structure after 0, 5.5, 8.5, 11 and 13.5 hours. Data corresponds to the 900 pixel side cube where flow velocities were simulated. Other statistics of the pore structure can be found in Supplementary Table S1. Table 1. Properties of the pore structure after 0, 5.5, 8.5, 11 and 13.5 hours. Data corresponds to the 900 pixel side cube where flow velocities were simulated. Other statistics of the pore structure can be found in Supplementary Table S1. Figure 2. Distribution of flow velocities in a vertical cross section at the center of the column before growth and after 13.5 hours. SiO2 particles are represented in grey and barite in white. Color scale refers to flow velocities. Lines enclose areas where flow: remains fast during the entire experiment, whereby crystal density is high (continuous line); is fast at the beginning and slow after 13.5 hours, in contrast crystal density is low (dashed line); is slow at the beginning and after 13.5 hours, few crystals are observed (dotted line). The graphic shows the 3D statistical spatial relation between the flow velocities and barite precipitation. re 2. Results and Discussion Surface area normalization. This error can contribute to the oscillation of calculated growth rates during Stage 2, however, within the experimental variability the error is not sufficiently large to invalidate equation 3.l q A 3D statistical spatial correlation between the distribution of crystals and the distribution of flow velocities (Fig. 2) shows that at the end of the experiment more crystals are present in volumes where the flow was ini- tially faster (see segmentation details in Supplementary Information). The increase of the percentage of pixels filled with barite as a function of the initial fluid velocity in those pixels is linear up to approximately 170 μm/s. Faster flow velocities do not seem to cause a significant change in the growth rate. Note that above 300 μm/s the points are scattered because there is less statistical data due to fewer pixels in each interval of flow velocities (Supplementary Fig. S2). pp y g Our results are concordant with previous literature showing that mineral reaction rates in porous media are strongly dependent on permeability and local flow velocities11,18. A linear relation between dissolution rates and flow velocities, when the kinetics is transport-limited, has also been proposed based on reactive transport mod- els39. Similarly, here we hypothesize that in subvolumes where the flow velocities are slower than approximately 170 μm/s growth is limited by transport, and above that velocity growth is limited by the reaction kinetics. Slower velocities may be insufficient to replenish the mineral-fluid interface with the same concentration of Ba2+/SO4 2− ions as the overall solution. Since during the experiment the flow velocity in some pore-throats decreases below 170 μm/s (Videos 2 and 3) we conclude that the overall decrease of growth rates during Stage 3 is caused by transport limitations in subvolumes of the column. This result is highly relevant as it is in contradiction with a growth rate limited by the reaction over the entire column, based on a high Pe during the experiment. Therefore, we conclude that parameters that represent the transport conditions averaged throughout a pore structure may Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 4 www.nature.com/scientificreports/ not be sufficient to interpret mineral reaction rates in heterogeneous porous media that is highly affected by local properties. Methods E i Experimental setup. The experiment consisted of flowing a solution supersaturated relative to barite through a microporous quartz column. The column with inner and outer diameters of 1.6 mm and 3.0 mm, respec- tively, was prepared by filling 10 mm of the column length with a mixture of 95 weight % borosilicate glass beads (3–150 μm) and 5 weight % natural quartz particles (180–300 μm). Column and filling were sintered at 645 °C for 5 minutes, forming a solid porous structure. Using a tomographic scan of the dry column, it was determined that the different sizes of beads and quartz particles formed an irregular pore structure with pore diameters up to 122 μm (Supplementary Fig. S6) and 39% fully interconnected porosity. μ ( pp y g ) y p y Synchrotron X-ray radiation at beam line 13-ID-B of the Advanced Photon Source was used to generate one tomogram of the column every 24 minutes. A total of 3200 projections were acquired per scan over 360 degrees with a beam intensity of 22 keV. The volume imaged was 2.38 × 2.38 × 1.49 mm (8.4 mm3) with a voxel size 1.24 μm, of which 3 mm3 corresponds to the inside of the column. This volume contains 71.3 mm2 of SiO2 surface area exposed to the fluid. Two solutions, a 0.23 mM Na2SO4 and a 0.23 mM BaCl2, where continuously flowed into a 5 mL chamber with magnetic stirring, forming a solution with a saturation index (SI) of 2.1, as defined in equation (S2), where aBa and aSO4 are the activity coefficients of barium and sulfate, respectively, and KSP is the solubility product of barite defined in PHREEQC-2. The experiment was performed at 21 ± 1 °C. The resulting solution was continuously fed to the column over the course of 13.5 hours with an initial flow rate of 60 mL/hr and a continuous pressure throughout the experiment applied by a Masterflex peristaltic pump from Cole-Parmer. Data analysis. Data was reconstructed using IDL software available at the beam line that generates stacks of 16-bit images, which were segmented using Avizo 8 and Fiji (see segmentation detail in Supplementary Information). Results and Discussion Surface area normalization. To generalize, mineral growth rates in a permeable heterogeneous microporous structure is similar to that expected on free surfaces if a) growth is not limited by transport, not only throughout the structure but also in the subvolumes with reduced permeability; and b) the rate is normalized to the evolving surface area of the growing crystals, e.g. using the dynamic method presented here, equations (1–3). The two points are closely related since the method used to calculate Acrystal(t) is only valid if the solution is homogeneous over the entire porous struc- ture, which is not the case if growth is transport limited in subvolumes of the structure. The presented dynamic surface area model is sufficient to establish the links between the evolving microporous network, flow properties and different kinetic stages. Now that we stablished the importance of accounting for the dynamic reactive surface area of crystals, it is clear that future research should focus on better ways to measure or calculate the reactive surface area as a function of time. Furthermore, we suggest the development of directional correlation functions to analyze the link between flow velocities and mineral growth rates40,41. Such combination of numerical and experimental methods is crucial to validate predictive reactive transport models and calculate the timeframe of key geological processes. Methods E i Reconstructed 3D stacks were segmented into three phases, each with a specific density that can be independently studied by thresholding their specific intensity: 1) fluid, fulfills the pore space; 2) SiO2, composes the solid matrix including glass beads, quartz particles and walls of the column; and 3) barite crystals, which nucleate on the SiO2 and grow into the fluid with time. The strong attenuation contrast between barite and SiO2 allows the quantification of the volume of barite precipitated at each time, which is used to calculate the average growth rate between scans. Simulation of flow velocities. The flow velocity field was simulated directly on voxelized images of the pore space resulting from segmentation of sCT scans42,43. In total we numerically analyzed five subvolumes obtained at different times: initial SiO2 porous structure at t = 0 hours and four subsequent stages with precipi- tated barite at t = 5.5, 8.5, 11 and 13.5 hours. Registered cubic subvolumes (edge length of 1.12 mm, corresponding to 9003 voxels) taken from the center of the column were used for all simulations. Within the initial subvolume the porosity was 35%, the maximum pore diameter was 89 μm and the average connectivity between pores was 7.3 pore-throats (see full statistical data for all subvolumes in Supplementary Information). Assuming creeping flow conditions (i.e., low Reynolds number (Re)), we numerically solve the so-called Stokes problem: η    ∆−∇ = = . p v v 0 div 0 (5) (5) where v is the velocity vector, η is the viscosity of the fluid, and p is the pressure field. No-slip boundary conditions are applied to all fluid-pore wall interfaces. Prescribed pressure boundary conditions assuming a linear pressure drop of 1 Pa per voxel and a viscosity of 10−3 Pa/s were used to solve equation (5). The flow rate measured at the beginning of the experiment was 60 ml/h. It was assumed that the constant pulses generated by the peristaltic pump, i.e. constant rotation speed, kept the pressure at the inlet approximately constant during the length of the experiment as a progressive decrease of the flow rate was observed. The four faces of the cube parallel to the flow direction were treated as solid walls (no-flow boundaries). This approach causes a distortion of the flow velocities in the voxels immediately adjacent to the lateral surfaces of the cube relative to the experimental flow conditions. Methods E i This effect is short-ranged and since we use a cubic domain’s size of 729 million voxels, which represents 62% of the total cross-sectional flow area, it is assumed to not affect the representativity of the overall experimental volume. Consequently, an adequate balance between representation of the experimental conditions and compu- tational resources needed to solve the Stokes problem is achieved (see details in Supplementary Information). where v is the velocity vector, η is the viscosity of the fluid, and p is the pressure field. No-slip boundary conditions are applied to all fluid-pore wall interfaces. Prescribed pressure boundary conditions assuming a linear pressure drop of 1 Pa per voxel and a viscosity of 10−3 Pa/s were used to solve equation (5). The flow rate measured at the beginning of the experiment was 60 ml/h. It was assumed that the constant pulses generated by the peristaltic pump, i.e. constant rotation speed, kept the pressure at the inlet approximately constant during the length of the experiment as a progressive decrease of the flow rate was observed. The four faces of the cube parallel to the flow direction were treated as solid walls (no-flow boundaries). This approach causes a distortion of the flow velocities in the voxels immediately adjacent to the lateral surfaces of the cube relative to the experimental flow conditions. This effect is short-ranged and since we use a cubic domain’s size of 729 million voxels, which represents 62% of the total cross-sectional flow area, it is assumed to not affect the representativity of the overall experimental volume. Consequently, an adequate balance between representation of the experimental conditions and compu- tational resources needed to solve the Stokes problem is achieved (see details in Supplementary Information). Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 5 www.nature.com/scientificreports/ References 1. DePaolo, D. J. & Cole, R. D. Geochemistry of Geologic Carbon Sequestration: An overview. Rev. Miner. Geochem. 77, 1–14, doi: 10.2138/rmg.2013.77.1 (2013). 1. DePaolo, D. J. & Cole, R. D. Geochemistry of Geologic Carbon Sequestration: An overview. Rev. Miner. Geochem. 77, 1–14, doi: 10.2138/rmg.2013.77.1 (2013). g ( ) 2. Boot-Handford, M. E. et al. Carbon capture and storage update. Energy Environ. Sci. 7, 130–189, doi: 10.1039/c3ee42350f (2014). C k h l Th l b f CO f l S T h l g 2. Boot-Handford, M. E. et al. Carbon capture and storage update g gy 3. Bourg, I. C., Beckingham, L. E. & Depaolo, D. J. The nanoscale basis of CO2 trapping for geologic storage. Environ. Sci. Technol. 49, 10265–10284, doi: 10.1021/acs.est.5b03003 (2015). 3. Bourg, I. C., Beckingham, L. E. & Depaolo, D. J. The nanoscale basis of CO2 trapping for geologic storage. Environ. Sci. Technol. 49 10265–10284, doi: 10.1021/acs.est.5b03003 (2015).i 4. Frenier, W. W. & Ziauddin, M. Formation, Removal, and Inhibition of Inorganic Scale in the Oilfield Environment. Society fo Petroleum Engineers: Richardson, USA (2008). g 5. Zhang, T., Gregory, K., Hammack, R. W. & Vidic, R. D. Co-precipitation of radium with barium and strontium sulfate and its impac on the fate of radium during treatment of produced water from unconventional gas extraction. Environ. Sci. Technol. 48, 4596–4603 doi: 10.1021/es405168b (2014). ( ) 6. Weatherley, D. K. & Henley, R. W. Flash vaporization during earthquakes evidenced by gold deposits. Nat. Geosci. 6, 294–298, doi: 10.1038/ngeo1759 (2013).h ( ) 6. Weatherley, D. K. & Henley, R. W. Flash vaporization during earthquakes evidenced by gold deposits. Nat. Geosci. 6, 294–298, doi 10.1038/ngeo1759 (2013).h / g ( ) 7. Gratier, J., Dysthe, D. & Renard, F. The role of pressure solution creep in the ductility of the earth’s upper crust. Adv. Geophys. 54, 47–179, doi: 10.1016/b978-0-12-380940-7.00002-0 (2013).l g 7. Gratier, J., Dysthe, D. & Renard, F. The role of pressure solution creep in the ductility of the earth’s upper crust. Adv. Geophys. 54 47–179, doi: 10.1016/b978-0-12-380940-7.00002-0 (2013).l 8. Steiger, D. M. Crystal growth in porous materials—II: Influence of crystal size on the crystallization pressure. J. Cryst. Growth 282 470–481, doi: 10.1016/j.jcrysgro.2005.05.008 (2005). j j y g 9. Putnis, A. & Austrheim, H. Fluid-induced processes: metasomatism and metamorphism. Geofluids 10, 254–269, doi 10.1111/j.1468-8123.2010.00285.x (2010). j 0. Emmanuel, S. & Berkowitz, B. Mixing-induced precipitation and porosity evolution in porous media. Adv. References Low sulfate seawater mitigates barite scale. Oil Gas J. 94 y g 21. Gray, P. R. NORM contamination in the petroleum industry. J. Petrol. Technol. 45, 12–16, doi: 10.2118/22880-PA (1993). 22. Olmstead, S. M., Muehlenbachs, L. A., Shih, J.-S., Chu, Z. & Krupnick, A. J. Shale Gas Development Impacts on Surface Water Quality in Pennsylvania. P. Natl. Acad. Sci. USA 110(13), 4962–4967, doi: 10.1073/pnas.1213871110 (2013). 3. Prieto, M., Astilleros, J. M. & Fernandez-Dias, L. Environmental remediation by crystallization of solid solutions. Elem. 9, 195–201 doi: 10.2113/gselements.9.3.195 (2013).h g 4. Martin, A. J., Crusius, J., McNee, J. J. & Yanful, E. K. The mobility of radium-226 and trace metals in pre-oxidized subaqueous uranium mill tailings. Appl. Geochem. 18, 1095–1110, doi: 10.1016/S0883-2927(02)00243-3 (2003). 25. Brant, F., Curti, E., Klinkenberg, M., Rozov, K. & Bosbach, D. Replacement of barite by a (Ba,Ra)SO4 solid solution at close-to- equilibrium conditions: A combined experimental and theoretical study. Geochim. Cosmochim. Acta 155, 1–15, doi: 10.1016/j. gca.2015.01.016 (2015). g 26. Klinkenberg, M., Brandt, F., Breuer, U. & Bosbach, D. Uptake of Ra during the Recrystallization of Barite: A Microscopic and Time of Flight-Secondary Ion Mass Spectrometry Study. Environ. Sci. Technol. 48, 6620–6627, doi: 10.1021/es405502e (2014). 7. Derluyn, H. et al. Recent micro-CT scanner developments at UGCT. Nucl. Instrum. Meth. B 324, 102–112, doi: 10.1016/j nimb.2013.10.051 (2014). 28. Noiriel, C., Steefel, C. I., Yang, L. & Ajo-Franklin, J. Upscaling calcite precipitation rates from pore to continuum scale. Chem. Geol. 318–319, 60–74, doi: 10.1016/j.chemgeo.2012.05.014 (2012). j g 9. Noiriel, C., Renard, F. & Doan, M. Gratier, Intense fracturing and fracture sealing induced by mineral growth in porous rocks J. Chem. Geol. 269, 197–209, doi: 10.1016/j.chemgeo.2009.09.018 (2010). j g 0. Noiriel, C. Resolving time-dependent Evolution of pore-scale structure, permeability and reactivity using X-ray microtomography Rev. Mineral. Geochem. 80, 247–286, doi: 10.2138/rmg.2015.80.08 (2015). Rev. Mineral. Geochem. 80, 247–286, doi: 10.2138/rmg.2015.80.08 ( g 31. Menke, H. P., Bijeljic, B. R., Andrew, M. G. & Blunt, M. J. Dynamic three-dimensional pore-scale imaging of reaction in a carbonate at reservoir conditions. Environ. Sci. Technol. 49, 4407–4414, doi: 10.1021/es505789f (2015). 2. Ellis, B. R. & Peters, C. A. 3D mapping of calcite and a demonstration of its relevance to permeability evolution in reactive fractures Adv. Water Resour 000, 1–8, doi: 10.1016/j.advwatres.2015.07.023 (2015).i 32. Ellis, B. R. & Peters, C. A. 3D mapping of calcite and a demonstration of its re Adv. References Water Resour. 28 337–344, doi: 10.1016/j.advwatres.2004.11.010 (2005).h j ( ) 1. Salehikhoo, F. & Li, L. The role of magnesite spatial distribution patterns in determining dissolution rates: When do they matter? G hi C hi A t 155 107 121 d i 10 1016/j 2015 01 035 (2015) j 1. Salehikhoo, F. & Li, L. The role of magnesite spatial distribution patterns in determining dissolution rates: When do they matter Geochim. Cosmochim. Acta 155, 107–121, doi: 10.1016/j.gca.2015.01.035 (2015). j 11. Salehikhoo, F. & Li, L. The role of magnesite spatial distribution patterns in determining diss Geochim. Cosmochim. Acta 155, 107–121, doi: 10.1016/j.gca.2015.01.035 (2015). 12. Li, L., Salehikhoo, F., Brantley, S. L. & Heidari, P. Spatial zonation limits magnesite dissolution in porous media. Geochim. Cosmochim. Acta 126, 555–573, doi: 10.1016/j.gca.2013.10.051 (2014). j g ( ) 13. Stack, A. G. Precipitation in Pores: A Geochemical Frontier. Rev. Miner. Geochem. 80, 165–190, doi: 10.2138/rmg.2015.80.05 (2 13. Stack, A. G. Precipitation in Pores: A Geochemical Frontier. Rev. Miner. Geochem. 80, 165–190, doi: 14. White, A. F. & Brantley, S. L. The effect of time on the weathering of silicate minerals: why do weathering rates differ in the laboratory and field? Chem. Geol. 202, 479–506, doi: 10.1016/j.chemgeo.2003.03.001 (2003). i 15. Steefel, C. I., DePaolo, D. J. & Lichtner, P. C. Reactive transport modeling: An essential tool and a new research approach for the Earth sciences. Earth Plan. Sci. Lett. 240, 539–558, doi: 10.1016/j.epsl.2005.09.017 (2005). j p 6. Molins, S. Reactive interfaces in direct numerical simulation of pore-scale processes. Rev. Min. Geochim. 80, 461–481, doi: 10.2138 rmg.2015.80.14 (2015). 7. Emmanuel, S. & Ague, J. J. Modeling the impact of nano-pores on mineralization in sedimentary rocks. Water Resour. Res. 45, 4, doi 10.1029/2008WR007170 (2009). 18. Li, L., Steefel, C. I. & Yang, L. Scale dependence of mineral dissolution rates within single pores and fractures. Geochim. Cosmochim. Acta 72, 360–377, doi: 10.1016/j.gca.2007.10.027 (2008).fl j g 9. Ferrar, K. J. et al. Assessment of effluent contaminants from three facilities discharging Marcellus Shale wastewater to surface water in Pennsylvania. Environ. Sci. Technol. 47, 3472–3481, doi: 10.1021/es301411q (2013). 19. Ferrar, K. J. et al. Assessment of effluent contaminants from three facilities discharging Marcellus Shale wast in Pennsylvania. Environ. Sci. Technol. 47, 3472–3481, doi: 10.1021/es301411q (2013). y q 20. Hardy, J. A. & Simm, I. Low sulfate seawater mitigates barite scale. Oil Gas J. 94, 64–67 (1996). & Simm, I. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 42. Vasilyev, R. V., Gerke, K. M., Karsanina, M. V. & Korost, D. V. Solution of the Stokes equation in three-dimensional geometry by the finite-difference method. Math. Mod. and Comp. Simul. 8, 63–72 (2016). 43 D H & Bl M J P k i f i i d h i Ph R E 80 036307 d i if 3. Dong, H. & Blunt, M. J. Pore-network extraction from micro-computerized-tomography images. Phys. Rev. E 80, 036307, doi 10.1103/PhysRevE.80.036307 (2009). Acknowledgementsh g The manuscript is based on experimental work supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division. Experimental measurements were performed at GeoSoilEnviroCARS (Sector 13), Advanced Photon Source (APS), Argonne National Laboratory. GeoSoilEnviroCARS is supported by the National Science Foundation - Earth Sciences (EAR- 1128799) and Department of Energy- GeoSciences (DE-FG02-94ER14466). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. The analysis was performed at the Diamond-Manchester Collaboration in the Research Complex at Harwell, funded in part by the EPSRC (EP/I02249X/1). K.M.G. acknowledges the support from RSF grant 14-17-00658 (pore-scale modelling) and RFBR 15-34-20989 мол_а_вед (porous media characterization). We thank Timofey Sizonenko for programming support. The underlying raw data is not shared online due to its size (>100Gb per scan). Author Contributions J.R.A.G. and A.G.S. planned and executed the experimental work; J.R.A.G. analyzed the experimental data; K.M.G. performed the flow simulations; All authors contributed to interpreting the results and to the writing of the paper. References Water Resour 000, 1–8, doi: 10.1016/j.advwatres.2015.07.023 (2015). is, B. R. & Peters, C. A. 3D mapping of calcite and a demonstration j 33. Godinho, J. R. A. & Stack, A. G. Growth Kinetics and Morphology of Barite Crystals Derived from Face-Specific Growth Rates. Cryst. Growth Des. 15, 2064–2071, doi: 10.1021/cg501507p (2015). Cryst. Growth Des. 15, 2064–2071, doi: 10.1021/cg501507p (2015 y g p 4. De Yoreo, J. J. & Vekilov, P. G. Principles of crystal nucleation and growth. Rev. Mineral. Geochem. 54, 57–93, doi: 10.2113/0540057 (2003). 5. Stack, A. G. et al. Pore-size-dependent calcium carbonate precipitation controlled by surface chemistry. Environ. Sci. Technol. 48 6177–6183, doi: 10.1021/es405574a (2014). 36. Lai, P., Moulton, K. & Krevor, S. Pore-scale heterogeneity in the mineral distribution and reactive surface area of porous rocks. Chem. Geol. 411, 260–273, doi: 10.1016/j.chemgeo.2015.07.010 (2015).fi 37. Seth, S. & Morrow, N. R. Efficiency of the Conversion of Work of Drainage to Surface Energy for Sandstone and Carbon Reserv. Eval. Eng. 10, 338–347, doi: SPE-102490-PA (2007).h g , , ( ) 38. Sheng, J. Modern Chemical Enhanced Oil Recovery: Theory and Practice, Gulf Professional Publishing. USA, 2010).fl g 38. Sheng, J. Modern Chemical Enhanced Oil Recovery: Theory and Practice, Gulf Professional P 38. Sheng, J. Modern Chemical Enhanced Oil Recovery: Theory 39. Molins, S., Trebotich, D., Steefel, C. I. & Shen, C. An investigation of the effect of pore scale flow on average geochemical reaction rates using direct numerical simulation. Water Resour. Res. 48, W03527, doi: 10.1029/2011WR011404 (2012). g 40. Gerke, K. M., Karsanina, M. V., Vasilyev, R. V. & Mallants, D. Improving pattern reconstruction using correlation func computed in directions. Europhys. Lett. 106, 66002, doi: 10.1209/0295-5075/106/66002 (2014). p p y 41. Gerke, K. M., Karsanina, M. V. & Mallants, D. Universal Stochastic Multiscale Image Fusion: An Example Application for Shale Rock. Sci. Rep. 5, 15880, doi: 10.1038/srep15880 (2015). Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 6 Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 Additional Information upplementary information accompanies this paper at http://www.nature.com/srepi Supplementary information accompanies this paper at http://www.nature.com/srepi Competing financial interests: The authors declare no competing financial interests.h Competing financial interests: The authors declare no competing financial interests. How to cite this article: Godinho, J. R. A. et al. The dynamic nature of crystal growth in pores. Sci. Rep. 6, 3086 d i 10 1038/ 33086 (2016) How to cite this article: Godinho, J. R. A. et al. The dynamic nature of crystal growth in pores. Sci. Rep. 6, 3086; doi: 10.1038/srep33086 (2016). How to cite this article: Godinho, J. R. A. et al. The dynamic nature of crystal growth in pores. Sci. Rep. 6, 33086; doi: 10.1038/srep33086 (2016). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ © The Author(s) 2016 Scientific Reports \| 6:33086 \| DOI: 10.1038/srep33086 7
https://openalex.org/W3105515791	https://diposit.ub.edu/dspace/bitstream/2445/172568/1/704973.pdf	English	null	Assessing Effectiveness of Colonic and Gynecological Risk Reducing Surgery in Lynch Syndrome Individuals	Cancers	2,020	cc-by	12,280	cancers cancers cancers Article Assessing Eﬀectiveness of Colonic and Gynecological Risk Reducing Surgery in Lynch Syndrome Individuals Nuria Dueñas 1,2,† , Matilde Navarro 1,2,3,†, Àlex Teulé 1, Ares Solanes 3, Mònica Salinas 1,2, Sílvia Iglesias 1,2, Elisabet Munté 1, Jordi Ponce 4 , Jordi Guardiola 5, Esther Kreisler 6, Elvira Carballas 7, Marta Cuadrado 8, Xavier Matias-Guiu 9, Napoleón de la Ossa 10,11 , Joan Lop 12 , Conxi Lázaro 1,2, Gabriel Capellá 1,2, Marta Pineda 1,2,‡ and Joan Brunet 1,2,13,,‡ 1 Hereditary Cancer Program, Catalan Institute of Oncology-IDIBELL, ONCOBELL, Hospitalet de Llobregat, 08908 Barcelona, Spain; nduenas@iconcologia.net (N.D.); mnavarrogarcia@iconcologia.net (M.N.); ateule@iconcologia.net (À.T.); msalinas@iconcologia.net (M.S.); siglesias@iconcologia.net (S.I.); emunter@iconcologia.net (E.M.); clazaro@iconcologia.net (C.L.); gcapella@iconcologia.net (G.C.); mpineda@iconcologia.net (M.P.) p g 2 Centro de Investigación Biomédica en Red de Cáncer (CIBERONC), Instituto Salud Carlos III, 28029 Madrid, Spain 3 Hereditary Cancer Program, Catalan Institute of Oncology, Badalona, 089016 Barcelona, Spain; asolanes@iconcologia.net 3 Hereditary Cancer Program, Catalan Institute of Oncology, Badalona, 089016 Barcelona, Spain; asolanes@iconcologia.net 4 Department of Gynecology, Bellvitge University Hospital, 08908 Hospitalet de Llobregat, 089016 Barcelona, Spain; jponce@bellvitgehospital.cat 5 4 Department of Gynecology, Bellvitge University Hospital, 08908 Hospitalet de Llobregat, 089016 Barcelona, Spain; jponce@bellvitgehospital.cat 5 Department of Gastroenterology, Bellvitge University Hospital, Hospitalet de Llobregat, 08908 Barcelo Spain; jguardiola@bellvitgehospital.cat 6 Department of General Surgery, Bellvitge University Hospital, Hospitalet de Llobregat, 08908 Barcelona, Spain; ekreisler@bellvitgehospital.cat p g p 7 Department of Gynecology, Trias i Pujol University Hospital, Badalona, 089016 Barcelona, Spain; ecarballas.germanstrias@gencat.cat 8 Department of General Surgery, Trias i Pujol University Hospital, Badalona, 089016 Barcelona, Spain; cuadradin@gmail.com 9 Department of Pathology, Bellvitge University Hospital, Hospitalet de Llobregat, 08908 Barcelona, Spain; fjmatiasguiu.lleida.ics@gencat.cat 10 Department of Pathology, Trias i Pujol University Hospital, Badalona, 089016 Barcelona, Spain; napoleondelaossa@gmail.com 11 Department of Pathology, Hospital General de Catalunya—Grupo Quironsalud, 08203 Barcelona, Spain 12 Department of Pathology, Hospital del Mar Institute for Medical Research, 08003 Barcelona, Spain; lopgros@gmail.com 13 Hereditary Cancer Program, Catalan Institute of Oncology-IDBIGI, 17007 Girona, Spain Correspondence: jbrunet@iconcologia.net; Tel.: +34-93-260-7959 † These authors contributed equally to this work. † These authors contributed equally to this work. ‡ These authors contributed equally to this work. ‡ These authors contributed equally to this work. Received: 14 October 2020; Accepted: 16 November 2020; Published: 18 November 2020 Received: 14 October 2020; Accepted: 16 November 2020; Published: 18 November 2020 Simple Summary: Colorectal and endometrial cancers are the most important life-threating risk in Lynch syndrome subjects, with incidences at 75 years as high as 40–60%. However, surveillance has shown to be ineﬀective. Risk reducing surgeries are an option in Lynch Syndrome (LS) individuals to decrease incidence of this type of cancers. In this manuscript, we have analyzed the rates of colorectal and gynecological cancer in 976 LS individuals after a mean follow-up of 10.2 years (patients under regular surveillance or after a risk reducing surgery). We can conﬁrm in the largest study published up to the present in a single-institution that risk reducing surgeries are eﬀective in decreasing incidence of colorectal and gynecological cancer in all LS carriers. Moreover, is the ﬁrst report showing a Cancers 2020, 12, 3419; doi:10.3390/cancers12113419 www.mdpi.com/journal/cancers www.mdpi.com/journal/cancers 2 of 17 Cancers 2020, 12, 3419 decrease in all-cause mortality cumulative incidence in females with Lynch syndrome that undergo gynecological risk reducing surgery. Abstract: Background: Colorectal (CRC) and endometrial cancer (EC) are the most common types of cancer in Lynch syndrome (LS). Risk reducing surgeries (RRS) might impact cancer incidence and mortality. Our objectives were to evaluate cumulative incidences of CRC, gynecological cancer and all-cause mortality after RRS in LS individuals. Methods: Retrospective analysis of 976 LS carriers from a single-institution registry. Primary endpoints were cumulative incidence at 75 years of cancer (metachronous CRC in 425 individuals; EC and ovarian cancer (OC) in 531 individuals) and all-cause mortality cumulative incidence, comparing extended (ES) vs. segmental surgery (SS) in the CRC cohort and risk reducing gynecological surgery (RRGS) vs. surveillance in the gynecological cohort. Results: Cumulative incidence at 75 years of metachronous CRC was 12.5% vs. 44.7% (p = 0.04) and all-cause mortality cumulative incidence was 38.6% vs. 55.3% (p = 0.31), for ES and SS, respectively. Cumulative, incidence at 75 years was 11.2% vs. 46.3% for EC (p = 0.001) and 0% vs. 12.7% for OC (p N/A) and all-cause mortality cumulative incidence was 0% vs. 52.7% (p N/A), for RRGS vs. surveillance, respectively. Conclusions: RRS in LS reduces the incidence of metachronous CRC and gynecological neoplasms, also indicating a reduction in all-cause mortality cumulative incidence in females undergoing RRGS. Keywords: Lynch syndrome; endometrial neoplasms; colorectal neoplasms; ovarian neoplasms; prophylactic surgical procedures; risk reduction; gynecological neoplasms; risk reducing surgery 2. Results Of all 976 carriers, 678 were diagnosed with at least one malignancy: in 410 (42.0%) the ﬁrst neoplasms were CRC (384 colonic (39.3%) and 26 rectal (2.7%)), and 125 (12.8%) were gynecological tumors (97 endometrial (9.9%), 28 ovarian (2.9%)) (Table 1). Mean age at diagnosis of ﬁrst cancer was 47.6 years (range 18–86). Table 1. Patient characteristics. 1. Introduction Lynch Syndrome (LS) is characterized by an inherited defect in the mismatch repair (MMR) genes ((MLH1, MSH2, MSH6, PMS2) or EPCAM gene deletions, resulting in silencing MSH2 gene in epithelial tissues)). It is the ﬁrst cause of inherited colorectal cancer (CRC) and endometrial cancer (EC) [1]. CRC cumulative incidences at 75 years by genes are 48.3–57.1%, 46.6–51.4%, 18.2–20.3% and 10.4% for MLH1, MSH2, MSH6 and PMS2 mutation carriers, respectively. Moreover, LS carriers have a high risk of developing multiple CRC and at a younger age [2,3]. Gynecological cancer risk is also increased. Endometrial cancer (EC) incidence is comparable to CRC, being the cumulative incidences at 75 years by genes 37, 48.9, 41.1 and 12.8% for MLH1, MSH2, MSH6 and PMS2 mutation carriers, respectively. Ovarian cancer (OC) risk is also signiﬁcantly increased with cumulative incidences at 75 years of 11, 17.4, 10.8 and 3% for MLH1, MSH2, MSH6 and PMS2 mutation carriers, respectively. The modestly increased risk of CRC and gynecological cancer of PMS2 mutation carriers is not evident before 50 years of age [2,3]. It should be noted, however, that the prognosis of Lynch associated tumors is generally good [4]. Follow-up with regular colonoscopies every 1–3 years demonstrated a signiﬁcant reduction in CRC-related mortality in the LS population [5,6]. Of note, interval cancers (tumors diagnosed between scheduled colonoscopies) continue to be diagnosed, and shorter colonoscopy intervals have not resulted in a reduction in incidence, a lowering of stage or an improvement in 10 year survival [7–10]. To reduce the incidence of metachronous CRC, extended colectomy is considered an option as it is a safe procedure without a signiﬁcant impact on quality of life despite altering bowel function [11,12]. Retrospective series and two meta-analyses reported a four-fold reduction in the incidence of metachronous CRC with no advantage in overall survival [13–25]. Most international guidelines recommend discussing the pros and cons of extended surgery with CRC aﬀected LS individuals before performing colorectal surgery for a ﬁrst neoplasm [26–30]. Screening for EC and OC is not well established. Patients are usually diagnosed due to their symptoms, even if they had a prior normal screening with pelvic ultrasound [31,32]. Even though 3 of 17 Cancers 2020, 12, 3419 endometrial biopsy is a better diagnostic tool, it is a badly tolerated procedure in women under surveillance for a long period of time [33]. 2. Results All Individuals Colorectal Cancer Cohort Gynecological Cancer Cohort Extended Surgery Segmental Surgery RRGS 1 Non-RRGS 2 TOTAL n = 976 n = 29 n = 261 n = 66 n = 465 Sex Male 445 (45.6%) 18 (62.1%) 146 (55.9%) Female 531 (54.4%) 11 (37.9%) 115 (44.1%) 66 (100%) 465 (100%) Mean age 54.1 y 7 (18–95) 56.4 y 7 (32–79) 59.9 y 7 (29–96) 57.3 y 7 (40–85) 54.8 y 7 (18–95) MMR gene MLH1 480 (49.2%) 19 (65.5%) 137 (52.5%) 33 (50.0%) 226 (48.6%) MSH2 262 (26.8%) 5 (17.2%) 77 (29.5%) 19 (28.8%) 127 (27.3%) MSH6 165 (16.9%) 1 (3.4%) 30 (11.5%) 10 (15.2%) 75 (16.1%) PMS2 48 (4.9%) 1 (3.4%) 14 (5.4%) 3 (4.5%) 22 (4.7%) EPCAM 21 (2.2%) 3 (10.3%) 3 (1.1%) 1 (1.5%) 15 (3.2%) Death 221 (22.6%) 4 (13.8%) 41 (15.7%) 0 (0%) 98 (21.1%) Mean age at death (range) 58.4 y 7 (25–89) 55.0 y 7 (44–71) 50.8 y 7 (29–84) 60.5 y 7 (25–89) First cancer diagnosis 3 678 (69.5%) 29 (100%) 261 (100%) 33 (50.0%) 277 (59.6%) Colon 384 (39.3%) 26 (89.7%) 214 (82.0%) 22 (33.3%) 119 (25.6%) Endometrial 97 (9.9%) 1 (3.4%) 17 (6.5%) 4 (6.1%) 86 (18.5%) Ovarian 28 (2.9%) 1 (3.4%) 3 (1.1%) 0 (0%) 25 (5.4%) Rectum 26 (2.7%) 0 (0%) 19 (7.3%) 3 (4.5%) 9 (1.9%) Other GI 4 26 (2.7%) 0 (0%) 2 (0.8%) 0 (0%) 10 (2.2%) Urologic 5 14 (1.4%) 0 (0%) 0 (0%) 2 (3.0%) 5 (1.1%) Other non-LS 103 (10.6%) 1 (3.4%) 6 (2.3%)) 2 (3.0%) 23 (4.9%) Mean age at ﬁrst cancer diagnosis (range) 47.6 y 7 (18–86) 46.3 y 7 (25–79) 45.9 y 7 (18–83) 46.2 y 7 (28–66) 49.0 y 7 (18–86) Mean age at surgery of study (range) 6 46.0 y 7 (25–79) 46.9 y 7 (18–83) 49.1 y 7 (36–72) 50.1 y 7 (28–80) 1 RRGS Risk reducing gynecological surgery, 2 non-RRGS Non-risk reducing gynecological surgery, 3 First cancer diagnosis First neoplasm developed by the subjects of our series, sorted from more to less frequent, 4 Other GI tumors stomach, small bowel and bile ducts, 5 Urologic tumors bladder and urinary tract, y years, 6 Mean age at surgery of study Mean age at ﬁrst colorectal cancer surgery (Extended surgery or segmental surgery) for the CRC cohort and mean age at gynecological surgery (risk reducing gynecological surgery or surgery for gynecological cancer) for the gynecological cancer cohort, 7 y years. 1. Introduction Schmeler et al., showed that risk reducing hysterectomy and oophorectomy in females with LS reduces the incidence of gynecological cancer by 100%, without major surgical complications [34]. In this context, international guidelines recommend discussing risk reducing gynecological surgery (RRGS) at age 35–40 after completion of childbearing in MLH1, MSH2 and MSH6 carriers [26–28,32]. Doubts exist regarding PMS2 carriers because of their lower cancer risk [3]. To our knowledge, no study has evaluated the impact of this procedure on mortality in LS subjects. Our objective was to evaluate the outcomes of risk reducing surgery in a single-institution LS cohort. We have evaluated cumulative incidences of metachronous CRC and all-cause mortality comparing segmental surgery (SS) vs. extended surgery (ES), as well as cumulative incidences of EC, OC and all-cause, EC-speciﬁc and OC-speciﬁc mortality comparing RRGS with standard gynecological follow-up. 2. Results Table 1. Patient characteristics. 1 RRGS Risk reducing gynecological surgery, 2 non-RRGS Non-risk reducing gynecological surgery, 3 First cancer diagnosis First neoplasm developed by the subjects of our series, sorted from more to less frequent, 4 Other GI tumors stomach, small bowel and bile ducts, 5 Urologic tumors bladder and urinary tract, y years, 6 Mean age at surgery of study Mean age at ﬁrst colorectal cancer surgery (Extended surgery or segmental surgery) for the CRC cohort and mean age at gynecological surgery (risk reducing gynecological surgery or surgery for gynecological cancer) for the gynecological cancer cohort, 7 y years. 2.1.1. Incidence of Colorectal Cancer Of all 976 carriers, the CRC cohort included 425 individuals with at least one previous CRC (55.8% men, 44.2% women); all of which were adenocarcinomas. Mean age at ﬁrst CRC diagnosis 4 of 17 Cancers 2020, 12, 3419 was 47.6 years (range 18–86). While 312 of the 425 carriers (73.4%) had a single cancer, 113 (26.6%) had more than one CRC (42 (9.9%) synchronous CRC and 71 (16.7%) metachronous CRC). The mean lapse between the ﬁrst and the second CRC was 10.5 years (range 5–18). No diﬀerences regarding metachronous and synchronous CRC were evident between MLH1 and MSH2 carriers (p > 0.5). No conclusions could be drawn for MSH6, PMS2 or EPCAM mutation carriers because of the small number of patients included (Table 2). Table 2. Number of colorectal cancer and type of second colorectal cancer according to mutated genes. Table 2. Number of colorectal cancer and type of second colorectal cancer according to mutated genes. Characteristics TOTAL n = 425 (100%) MLH1 n = 239 (56.2%) MSH2 n = 112 (26.4%) MSH6 n = 50 (11.8%) PMS2 n = 17 (4%) EPCAM n = 7 (1.6%) Number of CRC One 312 (73.4%) 181 (75.7%) 77 (68.8%) 38 (76.0%) 14 (82.4%) 2 (28.6%) 2 or more 113 (26.6%) 58 (24.3%) 35 (31.2%) 12 (24.0%) 3 (17.6%) 5 (71.4%) Mean age at ﬁrst CRC 2 diagnosis (range) 47.6 y 1 (18–86) 45.2 y 1 (18–86) 46.7 y 1 (21–83) 56.3 y 1 (33–78) 58.8 y 1 (38–72) 44.8 y 1 (33–61) Type of second CRC 2 Synchronous 42 (9.9%) 21 (8.8%) 9 (8.0%) 7 (14%) 2 (11.8%) 3 (42.9%) Metachronous 71 (16.7%) 37 (15.5%) 26 (23.2%) 5 (10%) 1 (5.9%) 2 (28.6%) 632 colorectal cancers were diagnosed in 425 subjects. 1 y years, 2 CRC colorectal cancer. Table 2. Number of colorectal cancer and type of second colorectal cancer according to m Data relating to ﬁrst surgery were available in 290 of the 425 cases: 29 subjects (10.0%) underwent ES (79.3% subtotal colectomies) at a mean age of 46.0 years (range 25–79). and 261 (90.0%) underwent SS (64.0% right hemicolectomy and 11.9% left hemicolectomy) at a mean age of 46.9 years (range 18–83). 2.1.1. Incidence of Colorectal Cancer No diﬀerences were observed in the proportion of carriers of MLH1, MSH2 and MSH6 gene mutations regarding gender, age at diagnosis, mean age at surgery and stage of ﬁrst CRC between surgery groups (p > 0.5) (Tables 1 and 3, Table S1). Cumulative incidence at 75 years of metachronous CRC was 12.5% in the ES group vs. 44.7% in the SS group (p = 0.04), signifying an 84% reduction in the risk of developing CRC when ES were performed (Figure 1A). Signiﬁcant diﬀerences in the rate of metachronous CRC were observed: 1 out of 29 (3.4%) in the ES group vs. 62 out of 261 (23.8%) in the SS group (p < 0.0001). (Table 3). Metachronous CRC in the SS group were mostly stage I (30.6%) and II (32.3%). One patient in the ES group developed a metachronous stage I rectal cancer (Table S1). Table 3. Number of second colorectal cancer according to mutated gene and type of surgery performed. Characteristics TOTAL n = 290 (29 1/261 2) MLH1 n = 156 (19 1/137 2) MSH2 n = 83 (5 1/78 2) MSH6 n = 31 (1 1/30 2) PMS2 n = 14 (1 1/13 2) EPCAM n = 6 (3 1/3 2) ONE CRC n = 192 Extended surgery 16/29 (55.2%) 13/19 (68.4%) 3/5 (60%) 0/1 (0%) 0/1 (0%) 0/3 (0%) Segmental surgery 176/261 (67.4%) 92/137 (67.2%) 50/78 (64.1%) 21/30 (70%) 12/13 (92.4%) 1/3 (33.3%) SYNCHRONOUS CANCER n = 35 Extended surgery 12/29 (41.4%) 6/19 (31.6%) 2/5 (40.0%) 1/1 (100%) 1/1 (100%) 2/3 (66.7%) Segmental surgery 23/261 (8.8%) 11/137 (8.0%) 7/78 (9.0%) 4/30 (13.3%) 0/13 (0%) 1/3 (33.3%) METACHRONOUS CANCER n = 63 Extended surgery 1/29 (3.4%) 0/19 (0%) 0/5 (0%) 0/1 (0%) 0/1 (0%) 1/3 (33.3%) Segmental surgery 62/261 (23.8%) 34/138 (24.8%) 21/78 (26.9%) 5/30 (16.7%) 1/13 (7.7%) 1/3 (33.3%) Information about surgery was available in 290 subjects: 192 with one CRC, 35 with synchronous CRC and 63 individuals with metachronous CRC. Percentages of type of second CRC are calculated according to the total number of individuals in each surgery group. 1 Extended surgery, 2 Segmental surgery. Table 3. Number of second colorectal cancer according to mutated gene and type of surgery performed. Cancers 2020, 12, 3419 surgery. 5 of 17 5 of 17 Figure 1. 2.1.2. Colorectal Cancer Mortality Follow-up for patients was 10.9 years (range 0–28) in the ES group vs. 14.7 years (range 0–47) in the SS group. Death occurred in 3 of 29 (10.3%) patients treated with ES vs. 73 of 261 (27.9%) treated with SS. All-cause mortality cumulative incidence was 38.6% in the ES group vs. 55.3% in the SS group, (p = 0.31) (Figure 1B). Follow-up for patients was 10.9 years (range 0–28) in the ES group vs. 14.7 years (range 0–47) in the SS group. Death occurred in 3 of 29 (10.3%) patients treated with ES vs. 73 of 261 (27.9%) treated with SS. All-cause mortality cumulative incidence was 38.6% in the ES group vs. 55.3% in the SS group, (p = 0.31) (Figure 1B). 2.2. Gynecological Cancer Cohort 2.1.1. Incidence of Colorectal Cancer Colorectal cancer cohort: (A) Cumulative incidence at 75 years of metachronous colorectal cancer in Lynch syndrome subjects comparing extended surgery and segmental surgery: Cumulative incidence at 75 years of metachronous colorectal cancer was 12.5% for extended surgery vs. 37.3% for Figure 1. Colorectal cancer cohort: (A) Cumulative incidence at 75 years of metachronous colorectal cancer in Lynch syndrome subjects comparing extended surgery and segmental surgery: Cumulative incidence at 75 years of metachronous colorectal cancer was 12.5% for extended surgery vs. 37.3% for segmental surgery (p = 0.004); (B) All-cause mortality cumulative incidence in Lynch syndrome subjects comparing extended surgery and segmental surgery: All-cause mortality cumulative incidence was 38.6% for extended surgery vs. 55.3% for segmental surgery (p = 0.31). Figure 1. Colorectal cancer cohort: (A) Cumulative incidence at 75 years of metachronous colorectal cancer in Lynch syndrome subjects comparing extended surgery and segmental surgery: Cumulative incidence at 75 years of metachronous colorectal cancer was 12.5% for extended surgery vs. 37.3% for Figure 1. Colorectal cancer cohort: (A) Cumulative incidence at 75 years of metachronous colorectal cancer in Lynch syndrome subjects comparing extended surgery and segmental surgery: Cumulative incidence at 75 years of metachronous colorectal cancer was 12.5% for extended surgery vs. 37.3% for segmental surgery (p = 0.004); (B) All-cause mortality cumulative incidence in Lynch syndrome subjects comparing extended surgery and segmental surgery: All-cause mortality cumulative incidence was 38.6% for extended surgery vs. 55.3% for segmental surgery (p = 0.31). 2.2.1. Incidence of Gynecological Cancer Characteristics TOTAL 150/531 (28.2%) MLH1 52/259 (20.1%) MSH2 58/146 (39.7%) MSH6 32/85 (37.6%) PMS2 6/25 (24.0%) EPCAM 2/16 (1.3%) Localization of gynecological cancer Endometrial 114 (76.0%) 40 (76.9%) 45 (77.6%) 24 (75.0%) 4 (66.7%) 1 (50.0%) Ovarian 27 (18.0%) 9 (17.3%) 10 (17.2%) 5 (15.6%) 2 (33.3%) 1 (50.0%) Endometrial + ovarian 9 (6.0%) 3 (5.8%) 3 (5.2%) 3 (9.4%) 0 (0.0%) 0 (0.0%) Mean age at gynecological cancer diagnosis (range) 49.9 y 1 (28–80) 47.6 y 1 (31–78) 46.4 y 1 (28–80) 51.0 y 1 (38–79) 53.5 y 1 (42–66) 38.0 y 1 (38–38) 150 females with LS developed a gynecological cancer. Percentages of gynecological cancers in each mutated gene group are calculated according to the total number of females in the cohort carrier of each mutated gene. 1 y years. Cancers 2020, 12, x FOR PEER REVIEW 7 of 2 non-RRGS groups respectively. This signifies a 74 and 100% risk reduction of developing EC and OC when RRGS were performed (Figure 2A,B). Table 5. Number of gynecological cancer according to mutated genes and type of surgery performed i ll LS Table 4. Number and location of gynecological cancer according to mutated genes. 150 females with LS developed a gynecological cancer. Percentages of gynecological cancers in each mutated gene group are calculated according to the total number of females in the cohort carrier of each mutated gene. 1 y years. Table 5. Number of gynecological cancer according to mutated genes and type of surgery performed Table 5. Number of gynecological cancer according to mutated genes and type of surgery performed in all LS women. Characteristics TOTAL n = 531 (66 1/465 2) MLH1 n = 259 (33 1/226 2) MSH2 n = 146 (19 1/127 2) MSH6 n = 85 (10 1/75 2) PMS2 n = 25 (3 1/22 2) EPCAM n = 16 (1 1/15 2) ENDOMETRIAL CANCER n = 123 RRGS 1 6 (9.1%) 4 (12.1%) 2 (10.5%) 0 (0%) 0 (0%) 0 (0%) non-RRGS 2 117 (25.2%) 39 (17.3%) 46 (36.2%) 27 (36.0%) 4 (18.2%) 1 (6.7%) OVARIAN CANCER n = 36 RRGS 1 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) non-RRGS 2 36 (7.7%) 12 (5.3%) 13 (10.2%) 8 (10.7%) 2 (9.1%) 1 (6.7%) 1 RRGS risk reducing gynecological surgery, 2 non-RRGS: non-risk reducing gynecological surgery. 2.2.1. Incidence of Gynecological Cancer A total of 159 gynecological cancers were diagnosed in 150 of the 531 females with LS (28.2%): 114 EC (76.0%), 27 OC (18.0%) and 9 synchronous EC and OC (6.0%). Mean age at diagnosis was 49.9 years (range 28–80). Histology was predominantly endometrioid (Table S3). A higher percentage of gynecological cancer was identiﬁed in MSH2 (39.7%) and MSH6 (37.6%) carriers compared with MLH1 (20.1%) (p < 0.05) (Table 4). Sixty-six women (12.4%) underwent RRGS at a mean age of 49.1 years (range 36–72) (57 HBSO, 1 hysterectomy plus salpingectomy and 8 hysterectomies). In the RRGS group, 6 women (9.1%) were diagnosed with an incidental EC identiﬁed in the pathological analysis of the surgical specimen. All 6 individuals were asymptomatic and had normal screening prior to surgery. All cancers were non-metastatic and three of them were pTis. However, in two cases extended surgical procedure were required after the diagnosis of EC. On the other hand, 117 EC were diagnosed in 465 (25.2%) women in the non-RRGS group at a mean age of 50.1 years (range 28–80), mostly stage I (n = 57, 48.7%). However, 4 women had stage IV EC (3.4%). No OC were diagnosed in the RRGS group (0%) vs. 36 in the non-RRGS group (7.7%): 16 stage I, 2 stage II and 7 stage III. No stage IV OC were diagnosed (Table 5 and Table S2). Cumulative incidence at 75 years of gynecological cancer was 11.2% in the RRGS group vs. 53.5% in the non-RRGS group (p < 0.001). Cumulative incidences at 75 years of each cancer were: 11.2% vs. 46.3% of EC (p = 0.001) and 0% vs. 12.7% (p N/A) of OC, in the RRGS and the non-RRGS groups respectively. This signiﬁes a 74 and 100% risk reduction of developing EC and OC when RRGS were performed (Figure 2A,B). 6 of 17 Cancers 2020, 12, 3419 Table 4. Number and location of gynecological cancer according to mutated genes. 2.2.1. Incidence of Gynecological Cancer Characteristics TOTAL n = 531 (66 1/465 2) MLH1 n = 259 (33 1/226 2) MSH2 n = 146 (19 1/127 2) MSH6 n = 85 (10 1/75 2) PMS2 n = 25 (3 1/22 2) EPCAM n = 16 (1 1/15 2) ENDOMETRIAL CANCER n = 123 RRGS 1 6 (9.1%) 4 (12.1%) 2 (10.5%) 0 (0%) 0 (0%) 0 (0%) non-RRGS 2 117 (25.2%) 39 (17.3%) 46 (36.2%) 27 (36.0%) 4 (18.2%) 1 (6.7%) OVARIAN CANCER n = 36 RRGS 1 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) non-RRGS 2 36 (7.7%) 12 (5.3%) 13 (10.2%) 8 (10.7%) 2 (9.1%) 1 (6.7%) 1 RRGS risk reducing gynecological surgery, 2 non-RRGS: non-risk reducing gynecological surgery. Table 5. Number of gynecological cancer according to mutated genes and type of surgery performed in all LS women. Characteristics TOTAL n = 531 (66 1/465 2) MLH1 n = 259 (33 1/226 2) MSH2 n = 146 (19 1/127 2) MSH6 n = 85 (10 1/75 2) PMS2 n = 25 (3 1/22 2) EPCAM n = 16 (1 1/15 2) Figure 2. Cumulative incidence at 75 years of gynecological cancer in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: (A) Cumulative incidence at 75 years of endometrial cancer. Cumulative incidence at 75 years of endometrial cancer was 11.2% for risk reducing gynecological surgery vs. 46.3% for non-risk reducing gynecological surgery (p = 0.001); (B) Cumulative incidence at 75 years of ovarian cancer. Cumulative incidence at 75 years of ovarian cancer was 0.0% for risk reducing gynecological surgery vs. 12.7% for non-risk reducing gynecological surgery (p = not assessable). Figure 2. Cumulative incidence at 75 years of gynecological cancer in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: (A) Cumulative incidence at 75 years of endometrial cancer. Cumulative incidence at 75 years of endometrial cancer was 11.2% for risk reducing gynecological surgery vs. 46.3% for non-risk reducing gynecological surgery (p = 0.001); (B) Cumulative incidence at 75 years of ovarian cancer. Cumulative incidence at 75 years of ovarian cancer was 0.0% for risk reducing gynecological surgery vs. 12.7% for non-risk reducing gynecological surgery (p = not assessable). Figure 2. 2.2.1. Incidence of Gynecological Cancer Cumulative incidence at 75 years of gynecological cancer in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: (A) Cumulative incidence at 75 years of endometrial cancer. Cumulative incidence at 75 years of endometrial cancer was 11.2% for risk reducing gynecological surgery vs. 46.3% for non-risk reducing gynecological surgery (p = 0.001); (B) Cumulative incidence at 75 years of ovarian cancer. Cumulative incidence at 75 years of ovarian cancer was 0.0% for risk reducing gynecological surgery vs. 12.7% for non-risk reducing gynecological surgery (p = not assessable). Figure 2. Cumulative incidence at 75 years of gynecological cancer in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: (A) Cumulative incidence at 75 years of endometrial cancer. Cumulative incidence at 75 years of endometrial cancer was 11.2% for risk reducing gynecological surgery vs. 46.3% for non-risk reducing gynecological surgery (p = 0.001); (B) Cumulative incidence at 75 years of ovarian cancer. Cumulative incidence at 75 years of ovarian cancer was 0.0% for risk reducing gynecological surgery vs. 12.7% for non-risk reducing gynecological surgery (p = not assessable). 3. Discussion 3. Discussion Most international guidelines recommend discussing the pros and cons of risk reducing surgery in affected LS individuals with CRC, and also considering risk reducing gynecological surgery (RRGS) in asymptomatic females with LS at age 35–40 after completion of childbearing. However, the strength of the evidence supporting these recommendations is still weak due to the limited sample size. This work evaluating risk reducing surgery (extended surgery after a first CRC and risk reducing gynecological surgery) is the largest single-center study published up to the present and points that in LS population these strategies have an impact on decreasing the cumulative incidence of cancer at 75 years and the cumulative incidence of mortality Most international guidelines recommend discussing the pros and cons of risk reducing surgery in aﬀected LS individuals with CRC, and also considering risk reducing gynecological surgery (RRGS) in asymptomatic females with LS at age 35–40 after completion of childbearing. However, the strength of the evidence supporting these recommendations is still weak due to the limited sample size. This work evaluating risk reducing surgery (extended surgery after a ﬁrst CRC and risk reducing gynecological surgery) is the largest single-center study published up to the present and points that in LS population these strategies have an impact on decreasing the cumulative incidence of cancer at 75 years and the cumulative incidence of mortality. of cancer at 75 years and the cumulative incidence of mortality. As reported in the literature, LS subjects have a higher risk of developing one or more CRC with cumulative incidences by the age 75 years of 48.3–57.1% for MLH1, 46.6–51.4% for MSH2, 18.2–20.3% for MSH6 and 12.8% for PMS2 [2,3]. The reason for this increased incidence of CRC was initially explained by accelerated carcinogenesis [35], since performing colonoscopies every 1–3 years vs. no follow-up showed a reduction both in incidence and mortality in LS individuals [5,6]. However, recent studies are raising doubts regarding the secondary prevention of CRC with annual colonoscopies since no differences were found in the incidence of advanced CRC with respect to longer intervals of colonoscopies [7–10]. Moreover, some CRC have been proven to evolve from a non-polypous precursor lesion at MMR-deficient crypts via somatic mutations in the CTNNB1 gene [36,37]. A recent publication identified differences in carcinogenesis between MMR genes, showing different risks of advanced adenoma (7.7% in MLH1 vs. 3. Discussion 3. Discussion 17.8% in MSH2 carriers) but a similar proportion of cancer (11.3% in MLH1 vs. 11.4% in MSH2 carriers) and different patterns of CTNNB1 somatic mutations (50% in CRC from MLH1 carriers vs. 7% in MSH2 carriers), suggesting that CRC in MLH1 mutation carriers may evolve more frequently through non-polypous precursor lesions and, therefore, are not prevented by colonoscopies [38]. As reported in the literature, LS subjects have a higher risk of developing one or more CRC with cumulative incidences by the age 75 years of 48.3–57.1% for MLH1, 46.6–51.4% for MSH2, 18.2–20.3% for MSH6 and 12.8% for PMS2 [2,3]. The reason for this increased incidence of CRC was initially explained by accelerated carcinogenesis [35], since performing colonoscopies every 1–3 years vs. no follow-up showed a reduction both in incidence and mortality in LS individuals [5,6]. However, recent studies are raising doubts regarding the secondary prevention of CRC with annual colonoscopies since no diﬀerences were found in the incidence of advanced CRC with respect to longer intervals of colonoscopies [7–10]. Moreover, some CRC have been proven to evolve from a non-polypous precursor lesion at MMR-deﬁcient crypts via somatic mutations in the CTNNB1 gene [36,37]. A recent publication identiﬁed diﬀerences in carcinogenesis between MMR genes, showing diﬀerent risks of advanced adenoma (7.7% in MLH1 vs. 17.8% in MSH2 carriers) but a similar proportion of cancer (11.3% in MLH1 vs. 11.4% in MSH2 carriers) and diﬀerent patterns of CTNNB1 somatic mutations (50% in CRC from MLH1 carriers vs. 7% in MSH2 carriers), suggesting that CRC in MLH1 mutation carriers may evolve more frequently through non-polypous precursor lesions and, therefore, are not prevented by colonoscopies [38]. therefore, are not prevented by colonoscopies [38]. Based on this increased risk of presenting both a first CRC and a metachronous CRC [2–4], it is important to discuss the extent of surgical resection for the first CRC. Several studies and two meta- analyses have reviewed this issue (Table 6) [13–25]. In the meta-analyses by Heneghan [16] and Anele, [17] the incidence of metachronous CRC was significantly lower in the ES group (6.8% and 6%) than in the SS group (23.5 and 22.8%) but no differences in 10 year survival were found. Similar to the Based on this increased risk of presenting both a ﬁrst CRC and a metachronous CRC [2–4], it is important to discuss the extent of surgical resection for the ﬁrst CRC. 2.2.2. Gynecological Cancer Mortality 2.2.2. Gynecological Cancer Mortality Follow-up for patients with and without RRGS was 8.7 years (range 0–43) and 10.4 years (range 0–45), respectively. Death occurred in 0 out of 66 (0%) women in the RRGS group vs. 98 out of 465 (21.1%) in the non-RRGS group, 11 due to EC (2.4%) and 6 due to OC (1.3%). All-cause mortality cumulative incidence was 0% in the RRGS group vs. 52.7% in the non-RRGS group (p N/A) (Figure 3). EC-specific mortality cumulative incidence was 0% in the RRGS group vs. 5.9% in the non-RRGS group (p N/A) and OC-specific mortality cumulative incidence was 0% in the RRGS group vs. 2.6% Follow-up for patients with and without RRGS was 8.7 years (range 0–43) and 10.4 years (range 0–45), respectively. Death occurred in 0 out of 66 (0%) women in the RRGS group vs. 98 out of 465 (21.1%) in the non-RRGS group, 11 due to EC (2.4%) and 6 due to OC (1.3%). All-cause mortality cumulative incidence was 0% in the RRGS group vs. 52.7% in the non-RRGS group (p N/A) (Figure 3). EC-speciﬁc mortality cumulative incidence was 0% in the RRGS group vs. 5.9% in the non-RRGS 7 of 17 7 of 17 Cancers 2020, 12, 3419 group (p N/A) and OC-speciﬁc mortality cumulative incidence was 0% in the RRGS group vs. 2.6% in the non-RRGS group (p N/A). Figure 3. All-cause mortality cumulative incidence in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: All-cause mortality cumulative incidence was 0.0% for 52.7% for risk reducing gynecological surgery vs. 52.7% for non- risk reducing gynecological surgery (p = not assessable). Figure 3. All-cause mortality cumulative incidence in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: All-cause mortality cumulative incidence was 0.0% for 52.7% for risk reducing gynecological surgery vs. 52.7% for non-risk reducing gynecological surgery (p = not assessable). Figure 3. All-cause mortality cumulative incidence in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: All-cause mortality cumulative incidence was 0.0% for 52.7% for risk reducing gynecological surgery vs. 52.7% for non- risk reducing gynecological surgery (p = not assessable). Figure 3. All-cause mortality cumulative incidence in females with Lynch syndrome comparing risk reducing gynecological surgery and non-risk reducing gynecological surgery: All-cause mortality cumulative incidence was 0.0% for 52.7% for risk reducing gynecological surgery vs. 2.2.2. Gynecological Cancer Mortality 2.2.2. Gynecological Cancer Mortality 52.7% for non-risk reducing gynecological surgery (p = not assessable). 3. Discussion 3. Discussion Several studies and two meta-analyses have reviewed this issue (Table 6) [13–25]. In the meta-analyses by Heneghan [16] and Anele, [17] the incidence of metachronous CRC was signiﬁcantly lower in the ES group (6.8% and 6%) than in the SS group (23.5 and 22.8%) but no diﬀerences in 10 year survival were found. Similar to 8 of 17 Cancers 2020, 12, 3419 the published literature, cumulative incidence at 75 years of metachronous CRC in our series was lower in the ES than in the SS group (12.5 vs. 44.7%, p = 0.04). Moreover, all-cause mortality cumulative incidence was lower in the ES group (38.6%) than in the SS group (55.3%). However, this diﬀerence was not signiﬁcant (p = 0.31). In agreement with previous reports [2,3], CRC in our series were diagnosed at younger ages in individuals carrying mutations in highly penetrant MMR genes (mean age: 45.2 and 46.7 years for MLH1 and MSH2 vs. 56.3 and 58.8 years for MSH6 and PMS2, respectively). The rates of metachronous CRC were also higher in these individuals (15.5 and 23.2% in MLH1 and MSH2 vs. 10.0 and 5.9% in MSH6 and PMS2, respectively) (Table 2). The 10 year follow-up results of the double-blind randomized CAPP2 trial have been recently published and the study shows that treatment with 600 mg aspirin daily vs. placebo decreases the incidence of CRC. The rates of CRC were 9% in the aspirin group vs. 13% in the placebo group (HR = 0.65; 95% CI 0.43–0.97; p = 0.035) with no signiﬁcant diﬀerences in adverse events or compliance between intervention groups [39]. To date, no study has compared extended surgery vs. segmental surgery plus chemoprevention with aspirin, so currently we cannot conclude that one approach is better than the other. Based on the latest evidence, extensive colonic surgery could be strongly recommended to MLH1 and MSH2 mutation carriers, but there are doubts regarding extending this recommendation to MSH6 mutation carriers [30]. In our series, 30.3% of MSH6 mutation carriers developed at least one CRC, and metachronous cancers were diagnosed in 0% of the individuals in the ES group vs. 16.7% in the SS group. For this reason, we recommend caution in these individuals, especially in those that develop CRC at younger ages. As for elderly subjects (>75–80) and rectal cancer, considering associated morbimortality, segmental surgery plus endoscopic surveillance seems the best option. 3. Discussion 3. Discussion Chemoprevention with aspirin has to be individualized in this situation, taking into account interactions with other comorbidities and treatments. In light of the literature published so far, this decision must be considered by a multidisciplinary team, discussing the pros and cons of both types of surgeries in every LS individual with a recent diagnosis of CRC. It is also important to highlight that ES does not completely prevent the risk of metachronous CRC, since rectal or sigmoid tissue is maintained to preserve functional outcome (as seen in one patient in our series). The risk of metachronous rectal cancer in these subjects has been reported as 3–12% at 10–12 years [4,14,18,40]. For this reason, regular endoscopic surveillance of the remaining colon or rectum should be maintained to reduce the incidence of CRC and its related mortality. In females with LS, gynecological cancer risk has been proven to equal or exceed the risk of CRC and it can be the sentinel malignancy in up to 35–40% of cases, with cumulative incidences at 75 years of age of EC and OC of 37% and 11% for MLH1, 48.9 and 17.4% for MSH2, 41.1 and 10.8% for MSH6 and 12.8 and 3% for PMS2, respectively [2,3]. Regarding EC, the results from our series also match those previously reported: gynecological tumors were diagnosed in 28.2% of the women and were the ﬁrst neoplasms in 21.7% of women who developed cancer (Table 4 and Table S2). However, the number of gynecological cancers in females carrying an MSH6 mutation were lower than expected, probably because most of our patients were identiﬁed by clinical criteria (Amsterdam and Bethesda), which are focused mainly on CRC, which is less frequent than EC in MSH6 carriers [2]. 9 of 17 Cancers 2020, 12, 3419 Table 6. Studies comparing colorectal cancer incidence and survival between extended surgery and segmental surgery in Lynch syndrome population. Author Year Collected Data/Type of Study n (ES 1/SS 2) Population Follow-Up (Years) Rate of mCRC 3 (ES 1/SS 2) 10 Years Overall Survival (ES 1/SS 2) Vasen [13] 1993 Retrospective Multicentric International 54 (17 1/37 2) Ams 4 5.8 (1–10) 11.8% 1 vs. 21.6% 2 (p = 0.394) n.r. 6 De Vos tot Nederveen WH [14] 2002 Retrospective Multicentric National 97 (29 1/68 2) LS 5 (MLH1, MSH2, MSH6) ES 1: 5 (1–15) SS 2: 6.8 (0–15) 3.5 1 vs. 11.8% 2 (p > 0.05) n.r. 3. Discussion 3. Discussion 6 Kalady MF [18] 2010 Retrospective Single-institution 296 (43 1/253 2) Ams 4 (11 LS 5 conﬁrmed) 8,7 (n.r) 8.0 1 vs. 25.0% 2 (p = 0.016) n.r. 6 Natarajan N [19] 2010 Retrospective Single-institution 106 (37 1,7/69 2) LS 5 (MLH1, MSH2) 12 (5–20) 10.8 1 vs. 33.3%2 (p = 0.006) 86.5 1 vs. 76.8% 2 (p = 0.239) Parry S [20] 2011 Retrospective Multicentric International 382 (50 1/332 2) LS 5 (MLH1, MSH2, MSH6, PMS2) ES 1: 8 (1–30) SS 2: 9 (1–40) 0 vs. 22.3% 2 (p = 0.019) 98 vs. 97% (p = 0.692) Stupart DA [21] (Stupart et al., 2011) 2011 Retrospective Single-institution 60 (21 1/39 2) LS 5 (MLH1, MSH2) ES 1: 8 (0–34) SS 2: 6 (1–30) 9.5 1 vs. 20.5% 2 (p = 0.346) 76 1 vs. 62% 2 (p = 0.222) Aronson M [22] 2015 Retrospective Single-institution 105 (29 1/76 2) LS 5 (MLH1, MSH2, MSH6, PMS2) 6.2 (0–55) 10.3 1 vs 28.9% 2 (p = 0.071) n.r. 6 Kim TJ [23] 2017 Retrospective Single-institution 106 (30 1/76 2) LS 5 (MLH1, MSH2, MSH6, EPCAM) ES 1: 5.7 (1–13) SS 2: 6.4 (0–14) 0 1 vs. 17.1% 2 (p = 0.038) 82.9 1 vs. 83.3% 2 (p = 0.659) 9 Hiatt MJ [24] 2017 Retrospective Single-institution 64 8 (16 1/48 2) LS 5 (MLH1, MSH2, MSH6, EPCAM) n.r. 6 6.3 1 vs. 27.0% 2 (p n.r. 6) 81.0 1 vs. 82.8% 2 (p = 0.471) Renkonen- Sinisalo L [25] 2017 Retrospective Multicentric National 242 (98 1/144 2) LS 5 (MLH1, MSH2, MSH6) 15.0 (0–32) 5.1 1 vs. 25.0% 2 (p < 0.001) 47.2 1 vs. 41.1% 2 (p = 0.83) 10 Roh SJ [15] 2020 Retrospective Single-institution 87 (51 1/36 2) Ams 4 ES 1: 7.7 (n.r) SS 2: 6.6 (n.r) 5.9 1 vs. 2.8% 2 (p = 0.637) n.r. 6 Heneghan HM [16] 2015 Meta-analysis 948 (168 1/780 2) LS 5 + Ams 4 8.9 (5–12) 6.8 1 vs. 23.5% 2 (p < 0.005) 89.8 1 vs. 90.7% 2 (p = 0.085) Anele CC [17] 2017 Meta-analysis 871 (166 1/705 2) LS 5 7.6 (6–12) 6 1 vs. 22.8% 2 (p < 0.0001) n.r. 6 CURRENT REPORT 2020 Retrospective Single-institution 293 (29 1/264 2) LS 5 (MLH1, MSH2, MSH6, PMS2, EPCAM) ES 1: 10.9 (0–28) SS 2: 14.7 (0–47) 3.4 1 vs. 23.8% 2 (p < 0.0001) n.r. 1 ES extended surgery, 2 SS segmental surgery, 3 mCRC metachronous colorectal cancer, 4 Ams families which meet Amsterdam criteria, 5 LS Lynch syndrome patients, 6 n.r. not reported, 7 extended right hemicolectomy is included in extended surgery, 8 only considered when initial tumor is right-sided tumor, 9 15 years overall survival, 10 25 years overall survival. 3. Discussion 3. Discussion 6 1 ES extended surgery, 2 SS segmental surgery, 3 mCRC metachronous colorectal cancer, 4 Ams families which meet Amsterdam criteria, 5 LS Lynch syndrome patients, 6 n.r. not reported, 7 extended right hemicolectomy is included in extended surgery, 8 only considered when initial tumor is right-sided tumor, 9 15 years overall survival, 10 25 years overall survival. Table 6. Studies comparing colorectal cancer incidence and survival between extended surgery and segmental surgery in Lynch syndrome population. incidence and survival between extended surgery and segmental surgery in Lynch syndrome population. Cancers 2020, 12, 3419 10 of 17 Our study conﬁrms that in females with LS, RRGS is a beneﬁcial procedure for reducing the incidences of EC and OC compared with regular follow-up. In our analysis of 531 women with LS, EC were signiﬁcantly reduced when performing RRGS (25.2 vs. 9.1%) and no OC were diagnosed in the RRGS group (7.7 vs. 0%). It is worth mentioning that our series also includes 85 female carriers of a mutation in the MSH6 gene, 32 of whom (37.6%) developed gynecological cancer (24 EC, 5 OC and 3 synchronous EC and OC), all from the non-RRGS group (Tables 4 and 5). We replicated the observations of Schmeler et al. [34], that included 315 females with LS (138 MLH1, 175 MSH2 and 3 MSH6); 61 of whom underwent RRGS at a mean age of 41 years. Incidences of EC and OC were similar to our study (Table 7). Even though the incidence of EC is considered 0% in the study by Schmeler et al., three females were incidentally diagnosed of EC at the time of prophylactic surgery and were included in the non-RRGS group. In our series, all the six cancers diagnosed in the RRGS group were incidental ﬁndings during the pathological analysis of the surgical specimen and two of them required extended oncological surgery. We have maintained these six diagnoses in the RRGS group because we consider that it better reﬂects routine clinical practice and is a more accurate consideration regarding the risks and beneﬁts of RRGS. The mean age at RRGS in our series was 49.1 years, older than expected, since in our institution RRGS were clearly recommended after 2015 and was then oﬀered to women under surveillance. Older age at RRGS could explain the higher incidence of EC in females in the RRGS group. 3. Discussion 3. Discussion No primary peritoneal cancer was identiﬁed in our analysis. The magnitude of the risk of primary peritoneal cancer in females with LS is unknown, but it is probably low as only ﬁve cases have been described so far [41]. Table 7. Studies comparing gynecological cancer incidence and survival between risk reducing gynecological surgery and surveillance in Lynch syndrome population. Table 7. Studies comparing gynecological cancer incidence and survival between risk reducing gynecological surgery and surveillance in Lynch syndrome population. Table 7. Studies comparing gynecological cancer incidence and survival between risk reducing gynecological surgery and surveillance in Lynch syndrome population. Author Year Collected Data/Type of Study n (RRGS 1/Non-RRGS 2) Follow-Up (Years) (RRGS 1/Non-RRGS 2) Rate EC 3 (RRGS 1/Non-RRGS 2) Rate OC 4 (RRGS 1/Non-RRGS 2) 10 Years Overall Survival (RRGS 1/Non-RRGS 2) Schmeler KM [34] 2006 Retrospective Multicentric National (USA) 315 (61 1/254 2) 13.3 1 (0.5–38) 7.4 2 (0.1–35) 0 1 vs. 33.0% 2 (p < 0.001) 0 1 vs. 5.5% 2 (p = 0.09) n.r. 6 CURRENT REPORT 2020 Retrospective Single-institution 531 (66 1/465 2) 8.7 1 (0–43) 10.4 2 (0–45) 9.1 1 vs. 25.2% 2 (p < 0.001) 0 1 vs. 7.7% 2 (p N/A 5) n.r. 6 1 RRGS risk reducing gynecological surgery, 2 non-RRGS non- risk reducing gynecological surgery, 3 EC Endometrial cancer, 4 OC Ovarian cancer, 5 N/A not assessable, 6 n.r. not reported. Our present estimates, show a reduction in all-cause, EC-speciﬁc and OC-speciﬁc mortality cumulative incidence in the RRGS group (all-cause mortality of 0 vs. 52.7%, EC-speciﬁc mortality of 0 vs. 5.9% and OC-speciﬁc mortality of 0 vs. 2.6% for RRGS and non-RRGS group, respectively). It must be taken into account that oncologic treatments have evolved since the start of this study, so it is possible that survival in females with LS diagnosed with gynecological cancers will improve in the future. So far, screening options for EC include TVUS, endometrial sampling and Ca125, even though most women are diagnosed due to their clinical symptoms. Therefore, EC surveillance does not imply neither a reduction in stage at diagnosis nor a survival improvement. Moreover, surveillance for OC has proven to be unsuccessful [32,33]. In consequence, main international clinical guidelines recommend considering prophylactic hysterectomy with or without bilateral salpingoophorectomy around the age of 35–40 years and after fulﬁlling childbearing in MLH1 and MSH2 carriers, and later (at 40 years) for MSH6. 3. Discussion 3. Discussion We have calculated the cumulative incidence of mortality (all-cause, EC-speciﬁc and OC-speciﬁc) using chronological age and not time-on-study as the time scale, because we considered that the follow-up 5 or 10 years after a preventive surgery performed at a mean age of 45–50 years is not usually a long enough follow-up time and, furthermore, it does not take into account the age of the patient to calculate mortality [43,44]. Even more so, if we consider that, in general, the prognosis for these tumors is considered to be good [2,3]. However, this approach can be argued as unadjusted age scale can lead to biases. In the CRC cohort, information about cause of death was lacking or it was not reliable in a high number of patients and only all-cause mortality cumulative incidence could be analyzed. It would have been interesting to analyze the cumulative incidence of metachronous CRC-speciﬁc mortality. However, considering the study follow-up time and the improvements in OS and DFS of the new combinations of oncologic treatments over the last years, it is possible that if diﬀerences in mortality were found, they were due to new treatment of the metachronous CRC and not surgical treatment of the ﬁrst CRC. A possible limitation of our study is that patients were allocated by the surgeon to segmental or extended surgery groups based on clinical criteria (age at CRC diagnosis, localization of the tumor, existence of synchronous cancer, patient’s preference, etc.) without information regarding LS condition, MMR protein expression or BRAF mutation in tumor. We can state, however, that there were no statistically signiﬁcant diﬀerences between groups regarding gender, mutated gene, age at CRC diagnosis or CRC stage (Tables 1–3 and Table S1). A similar situation occurred in the gynecological cancer cohort, where individuals were included in each group based on patient’s preference, family history and existence of a benign pathology that required gynecological surgery and, after 2015 it was then oﬀered to all females with LS. We veriﬁed that no signiﬁcant diﬀerences existed between mean age at gynecological surgery or proportion of carriers for each gene (Tables 1, 4 and 5). It must be considered that biases can also exist regarding diﬀerences between time of follow-up and number of patients included in each surgery group, both in the colonic and gynecological cohort. 3. Discussion 3. Discussion To avoid early menopause complications, diﬀerent approaches can be considered: oestrogenic hormonal replacement therapy (HRT) until the natural age of menopause, or a two-step surgery with hysterectomy after childbearing is complete followed by salpingoophorectomy after menopause [32]. More data are required to extend this recommendation to PMS2 carriers because of their lower gynecological cancer risk, which is not increased before 50 years of age [2,3]. 11 of 17 Cancers 2020, 12, 3419 Nevertheless, it seems reasonable that if RRGS is considered, it can be oﬀered after the natural age of menopause [26–28,32]. In line with these results, a recently published survey conducted by the prospective Lynch Syndrome database (PLSD) show that most of the referral centers included worldwide (91–95%) oﬀer RRGS to carriers of pathogenic variants in MLH1, MSH2 and MSH6 but less (67%) to carriers of pathogenic variants in PMS2. Most of the centers (71%) also recommend oestrogen-only HRT between 35-55 years, approximately [42]. Nevertheless, it seems reasonable that if RRGS is considered, it can be oﬀered after the natural age of menopause [26–28,32]. In line with these results, a recently published survey conducted by the prospective Lynch Syndrome database (PLSD) show that most of the referral centers included worldwide (91–95%) oﬀer RRGS to carriers of pathogenic variants in MLH1, MSH2 and MSH6 but less (67%) to carriers of pathogenic variants in PMS2. Most of the centers (71%) also recommend oestrogen-only HRT between 35-55 years, approximately [42]. Important strengths of our analysis are the internal robustness due to the use of prospectively acquired data and the long-term follow-up and that all patients were treated in one single comprehensive cancer institution with the same protocol. Furthermore, this is the ﬁrst report analyzing the impact of both colonic and gynecological risk reducing surgeries in the same LS series. Here, we describe for the ﬁrst time a decrease in all-cause, EC-speciﬁc and OC-speciﬁc mortality cumulative incidence after RRGS. We were not able to calculate statistical signiﬁcance in the cumulative incidence of mortality in the gynecological cancer cohort, because no woman died in the RRGS group after 8.7 years of mean follow-up vs. 98 women in the non-RRGS group after 10.4 years of mean follow-up. 4.1. Study Sample This is a retrospective cohort study from the Catalan Institute of Oncology Hereditary Cancer Program that included 976 individuals (531 women, 445 men) belonging to 234 LS families: 826 proven carriers (84.6%) and 150 obligate carriers (15.4%) of a pathogenic mutation in one of the MMR genes (480 MLH1, 262 MSH2, 165 MSH6, 48 PMS2 and 21 EPCAM). All the families were visited at the Genetic Counselling Units of our center between 1999 and 2016. Mean follow-up was 10.2 years (range 0–47) (Table 1). All patients underwent appropriate genetic counselling prior to all genetic tests and gave informed consent for genetic analysis and internal Ethics Committee approved this study. This study was performed in accordance with the Declaration of Helsinki. 4.2. Data Collection Data supporting the results are stored in the clinical database of the Hereditary Cancer Program. If surgery was performed, information regarding indication of surgery (prophylactic vs. non-prophylactic), extension of surgery, date of surgery and pathological note was recorded. Information about surgery and pathological ﬁndings were veriﬁed by medical reports. Demographic, personal characteristics, genogram and personal history of cancer were collected. When missing data occurred, the subject was excluded from the analysis that involved the missing data point, but the subject was included in other analysis where complete data were available. Clinical information before LS diagnosis was reviewed retrospectively and included in the clinical database, while analyzed observations after the identiﬁcation of a MMR gene mutation were prospective. After LS diagnosis, colonic and gynecological follow-up or RRGS was indicated. Follow-up recommendation in our institution included a 2-year interval colonoscopy with or without chromoendoscopy, beginning at 20–25 years (MLH1 and MSH2 carriers) or at 25–30 years (MSH6 and PMS2 carriers) unless an earlier onset cancer existed in the family. After 40 years of age, colonoscopy was performed annually. In women, yearly gynecological examination with transvaginal ultrasound (TVUS) was recommended after 30 years of age. 3. Discussion 3. Discussion Another possible limitation is the existence of a cohort eﬀect, as patients were included in the study between 1969 and 2016 and during these years, life expectancy has increased and oncological treatments, surgical procedures and endoscopic techniques have improved, achieving improvements in diagnoses and increasing patients’ survival. Therefore, we must bear in mind that, due to the small number of patients and the nature of retrospective studies, certain selection biases and risk overestimations cannot be ruled out. For that reason, although risk reducing surgery has shown a tendency to reduce mortality in LS individuals, these results are not statistically signiﬁcant and must be conﬁrmed in prospective studies. Another limitation of our study is that neither quality of life nor morbidity was evaluated. Other similar publications did not ﬁnd important diﬀerences in generic quality of life in ES, unless total 12 of 17 Cancers 2020, 12, 3419 colectomy with ileorectal anastomosis was performed [11,12]. Two studies evaluating risk-reducing HBSO in females with LS concluded that surgery does not have a negative eﬀect on quality of life, especially in those using hormonal therapy [45,46]. Any signiﬁcant diﬀerences in surgical morbidity of CRC and gynecological surgery were not reported in several prospective studies [11,12,45,46]. It should be considered that if a 35–40-year female with LS who has fulﬁlled childbearing has to be intervened for a CRC, HBSO should be oﬀered in the same surgical act to avoid further complications [32]. 4.5. Deﬁnitions Segmental surgery for CRC was deﬁned as right or left hemicolectomy, sigmoid colectomy, transverse colectomy, abdominoperineal resection or low anterior resection. Extended surgery was deﬁned as total colectomy with ileorectal anastomosis or subtotal colectomy with ileosigmoid anastomosis. Synchronous CRC was deﬁned as a CRC identiﬁed at the same moment or during the 6 months following the diagnosis of a primary CRC. Metachronous CRC was deﬁned as a CRC diagnosed at least 6 months after ﬁrst CRC diagnosis. Females with LS were classiﬁed depending on the treatment received: RRGS vs. non-RRGS. RRGS was deﬁned as surgery in a women not previously diagnosed with gynecological cancer. Surgery was either hysterectomy, hysterectomy plus salpingectomy or hysterectomy plus bilateral salpingoophorectomy (HBSO), whichever was considered best in a gynecological assessment. All women had a presurgical evaluation consisting of a transvaginal ultrasound. In case of suspicion, endometrial biopsies and/or complementary imaging examinations were performed. All-cause mortality considered any cause of death. EC-speciﬁc mortality and OC-speciﬁc mortality calculations considered women with stage IV EC or OC and whose death were attributable to progression of these neoplasms. 4.4. Gynecologicall Cancer Cohort In the gynecological cancer cohort, all women diagnosed with LS underwent gynecological screening with a yearly gynecological consultation that included anamnesis, physical examination and transvaginal ultrasound, starting at 30 years of age. Since 2003 RRGS were recommended to females with LS when there existed a family history of gynecological cancer or they presented symptoms caused by benign lesions. However, this surgery has been actively recommended since 2015, following international guidelines [27,28,32]. Data about gynecological cancer, gynecological surgery and cause of death were available in all the individuals. Cumulative incidence at 75 years of gynecological cancer (endometrial or ovarian) and all-cause, EC-speciﬁc and OC-speciﬁc mortality cumulative incidences were compared among females who had undergone RRGS and those who had not. For cumulative incidences at 75 years of EC and OC endpoint, individuals were followed until diagnosis of EC, OC, RRGS or censoring at death or date of last assessment. For all-cause, EC-speciﬁc and OC-speciﬁc mortality cumulative incidence endpoint, individuals were followed until death or date of last assessment. 4.3. Colorectal Cancer Cohort In the CRC cohort, all patients had at least one previous CRC. Colorectal surgeries for ﬁrst cancers were decided by surgeons at diagnosis, considering clinical criteria (age of the subject, localization of the tumor and existence of a synchronous cancer). At that moment, most of the patients were not diagnosed with LS as LS screening in the colectomy specimen based on Jerusalem criteria was implemented in our center at the beginning of January 2016. Data about ﬁrst CRC surgery were available in 290 of the 425 patients. Only the surgical treatment for the ﬁrst CRC was considered. Information on the cause of death was missing in a high number of patients or it was unreliable, so this information was not considered. Each individual was counted only once, regardless of how many cancers he or she might have had. Cumulative incidence at 75 years of metachronous CRC and all-cause mortality cumulative incidence were compared between SS and ES. For cumulative incidence at 75 years of metachronous CRC endpoint, individuals were followed until diagnosis of metachronous CRC or censoring at death or date of last assessment. For all-cause mortality cumulative incidence endpoint, individuals were followed until death or date of last assessment. Cancers 2020, 12, 3419 13 of 17 13 of 17 4.6. Mutation Testing Patients referred for MMR mutation analysis were suspected of having LS because they fulﬁlled LS clinical criteria (Amsterdam or revised Bethesda guidelines) or, as of January 2016, Jerusalem criteria [47]. Point mutation analyses of MMR genes were performed by PCR ampliﬁcation of exonic regions and exon-intron boundaries followed by Sanger sequencing or by next generation sequencing. Genomic rearrangements in MMR genes were analyzed by multiplex ligation dependent probe ampliﬁcation kits (MRC-Holland). Variant classiﬁcation was completed according to Insight guidelines [48]. Laboratory reports identifying a pathogenic variant in any MMR gene were required for conﬁrmation. Untested individuals were considered obligate MMR mutation carriers whenever the evidence indicated that they were the only possible transmitters of the mutation to their oﬀspring. 5. Conclusions In conclusion, this study conﬁrms that colonic and gynecological risk reducing surgeries are eﬀective decreasing the incidence of metachronous CRC and gynecological cancer in LS patients. This beneﬁt was seen in all LS subjects; however, caution is still needed for MSH6 and PMS2 mutation carriers. This is the ﬁrst report pointing to a reduction in the all-cause mortality cumulative incidence in females with LS that undergo RRGS. Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6694/12/11/3419/s1, Table S1: Stage of ﬁrst and second metachronous colorectal cancer comparing segmental to extended surgery in LS individuals with colorectal cancer, Table S2: Stage of endometrial and ovarian cancer according to type of surgery in all LS women with gynecological cancer, Table S3: Histology of gynecological cancers diagnosed in the gynecological cohort. Author Contributions: Conceptualization: N.D., M.N., M.P., J.B.; provide cases and corresponding data: N.D., M.N., A.S., M.S., S.I., E.M., J.P., J.G., E.K., E.C., M.C., X.M.-G., N.d.l.O.; facilitate germline data: C.L., M.P.; evaluation and statistical analysis: N.D., M.N., E.M., J.L., G.C., M.P., J.B.; writing original draft: N.D., M.N.; writing review and editing: N.D., M.N., À.T., E.M., J.L., C.L., G.C., M.P., J.B. All authors contributed to manuscript editing. All authors have read and agreed to the published version of the manuscript. Funding: This research has been funded by the Instituto de Salud Carlos III and co-funded by European Social Fund—ESF investing in your future—(grant CM19/00099), the Catalan-Balearic Society of Oncology (2018 grant of the Catalan-Balearic Society of Oncology), the Spanish Ministry of Economy and Competitiveness and co-funded by FEDER funds—A way to build Europe—(grant SAF2015-68016-R), CIBERONC and the Government of Catalonia (grants 2017SGR1282 and PERIS SLT002/16/0037). We thank the CERCA Program/Generalitat de Catalunya for institutional support. Acknowledgments: First of all, we thank the patients who participated in this study and their families. We also thank all the members of the Units of Genetic Counselling and Genetic Diagnostics of the Hereditary Cancer Program of the Catalan Institute of Oncology (ICO-IDIBELL) for their contribution as well as the BufaLynch Association for their support and funding of ICO’s Lynch Syndrome Database. The authors would also like to acknowledge the Department of Medicine at the Universitat Autònoma de Barcelona. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. 4.7. Statistical Analysis The Kaplan–Meier method was used to estimate cumulative incidence at 75 years of cancer and all-cause mortality cumulative incidence, EC-speciﬁc mortality cumulative incidence and OC-speciﬁc mortality cumulative incidence. Age rather than time-on-study was used as the time scale, as recommended [43,44]. Comparisons of proportions between groups were assessed using the test for equality of proportions. Log-rank and Peto–Peto test compared survival curves between the diﬀerent 14 of 17 Cancers 2020, 12, 3419 surgical techniques. To control for confounding variables, adjustments for speciﬁc mutated MMR gene were performed. In all tests, the level of statistical signiﬁcance was set at p < 0.05. Statistical analyses were performed using R version 3.2.1 and 3.4.0. surgical techniques. To control for confounding variables, adjustments for speciﬁc mutated MMR gene were performed. In all tests, the level of statistical signiﬁcance was set at p < 0.05. Statistical analyses were performed using R version 3.2.1 and 3.4.0. 5. Conclusions Ethics Approval and Consent to Participate: All patients underwent appropriate genetic counselling prior to all genetic tests and gave informed consent for genetic analysis and internal Ethics Committee approved this study (code PR225/11). This study was performed in accordance with the Declaration of Helsinki. References 1. Moreira, L.; Balaguer, F.; Lindor, N.; De La Chapelle, A.; Hampel, H.; Aaltonen, L.A.; Hopper, J.L.; Le Marchand, L.; Gallinger, S.; Newcomb, P.A.; et al. Identiﬁcation of Lynch Syndrome Among Patients with Colorectal Cancer. JAMA 2012, 308, 1555–1565. [CrossRef] [PubMed] 1. Moreira, L.; Balaguer, F.; Lindor, N.; De La Chapelle, A.; Hampel, H.; Aaltonen, L.A.; Hopper, J.L.; Le Marchand, L.; Gallinger, S.; Newcomb, P.A.; et al. Identiﬁcation of Lynch Syndrome Among Patients with Colorectal Cancer. JAMA 2012, 308, 1555–1565. [CrossRef] [PubMed] 2. Dominguez-Valentin, M.; Sampson, J.R.; Seppälä, T.T.; Broeke, S.W.T.; Be, J.-P.P.; Nakken, S.; Engel, C.; Aretz, S.; Jenkins, M.A.; Sunde, L.; et al. Correction: Cancer risks by gene, age, and gender in 6350 carriers of pathogenic mismatch repair variants: Findings from the Prospective Lynch Syndrome Database. Genet. Med. 2020, 22, 15–25. [CrossRef] [PubMed] 3. ten Broeke, S.W.; Van Der Klift, H.M.; Tops, C.M.; Aretz, S.; Bernstein, I.; Buchanan, D.D.; De La Chapelle, A.; Capella, G.; Clendenning, M.; Engel, C.; et al. Cancer Risks for PMS2-Associated Lynch Syndrome. J. Clin. Oncol. 2018, 36, 2961–2968. [CrossRef] [PubMed] 3. ten Broeke, S.W.; Van Der Klift, H.M.; Tops, C.M.; Aretz, S.; Bernstein, I.; Buchanan, D.D.; De La Chapelle, A.; Capella, G.; Clendenning, M.; Engel, C.; et al. Cancer Risks for PMS2-Associated Lynch Syndrome. J. Clin. Oncol. 2018, 36, 2961–2968. [CrossRef] [PubMed] 4. Møller, P.; Seppälä, T.; Bernstein, I.; Holinski-Feder, E.; Sala, P.; Evans, D.G.; Lindblom, A.; Macrae, F.; Blanco, I.; Sijmons, R.; et al. Incidence of and survival after subsequent cancers in carriers of pathogenic MMR variants with previous cancer: A report from the prospective Lynch syndrome database. Gut 2017, 66, 1657–1664. [CrossRef] 15 of 17 Cancers 2020, 12, 3419 5. Vasen, H.; Abdirahman, M.; Brohet, R.; Langers, A.M.J.; Kleibeuker, J.H.; Van Kouwen, M.; Koornstra, J.J.; Boot, H.; Cats, A.; Dekker, E.; et al. One to 2-Year Surveillance Intervals Reduce Risk of Colorectal Cancer in Families with Lynch Syndrome. Gastroenterology 2010, 138, 2300–2306. [CrossRef] 6. Järvinen, H.J.; Aarnio, M.; Mustonen, H.; Aktan–Collan, K.; Aaltonen, L.A.; Peltomäki, P.; De La Chapelle, A.; Mecklin, J.P. Controlled 15-year trial on screening for colorectal cancer in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 2000, 118, 829–834. [CrossRef] 7. Seppälä, T.T.; Ahadova, A.; Dominguez-Valentin, M.; Macrae, F.; Evans, D.G.; Therkildsen, C.; Sampson, J.; Scott, R.; Burn, J.; Möslein, G.; et al. References Lack of association between screening interval and cancer stage in Lynch syndrome may be accounted for by over-diagnosis; a prospective Lynch syndrome database report. Hered. Cancer Clin. Pract. 2019, 17, 1–8. [CrossRef] [PubMed] 8. Seppälä, T.; in collaboration with The Mallorca Group; Pylvänäinen, K.; Evans, D.G.; Järvinen, H.; Renkonen-Sinisalo, L.; Bernstein, I.; Holinski-Feder, E.; Sala, P.; Lindblom, A.; et al. Colorectal cancer incidence in path_MLH1 carriers subjected to diﬀerent follow-up protocols: A Prospective Lynch Syndrome Database report. Hered. Cancer Clin. Pract. 2017, 15, 1–10. [CrossRef] [PubMed] 9. Engel, C.; Vasen, H.F.; Seppälä, T.; Aretz, S.; Bigirwamungu-Bargeman, M.; De Boer, S.Y.; Bucksch, K.; Büttner, R.; Holinski-Feder, E.; Holzapfel, S.; et al. No Diﬀerence in Colorectal Cancer Incidence or Stage at Detection by Colonoscopy among 3 Countries with Diﬀerent Lynch Syndrome Surveillance Policies. Gastroenterology 2018, 155, 1400–1409. [CrossRef] [PubMed] 10. Dominguez-Valentin, M.; Seppälä, T.T.; Sampson, J.R.; Macrae, F.; Winship, I.; Evans, D.G.; Scott, R.J.; Burn, J.; Möslein, G.; Bernstein, I.; et al. Survival by colon cancer stage and screening interval in Lynch syndrome: A prospective Lynch syndrome database report. Hered. Cancer Clin. Pract. 2019, 8, 4–9. [CrossRef] [PubMed] 11. Haanstra, J.F.; Cappel, W.H.D.V.T.N.; Gopie, J.P.; Vecht, J.; Vanhoutvin, S.A.L.W.; Cats, A.; Van Der Zaag-Loonen, H.J.; Langers, A.M.J.; Bergmann, J.H.W.; Van De Meeberg, P.C.; et al. Quality of Life After Surgery for Colon Cancer in Patients with Lynch Syndrome. Dis. Colon Rectum 2012, 55, 653–659. [CrossRef] 12. Urso, E.D.L.; Celotto, F.; Giandomenico, F.; Gavaruzzi, T.; Del Bianco, P.; Lotto, L.; Spolverato, G.; Pucciarelli, S.; Bao, Q.R.; Emanuele, D.U. Analysis of morbidity and mortality, quality of life and bowel function after total colectomy with ileorectal anastomosis versus right and left hemicolectomy: A study to optimise the treatment of lynch syndrome and attenuated polyposis coli. Eur. J. Surg. Oncol. 2020, 46, 1613–1619. [CrossRef] [PubMed] 13. Vasen, H.F.; Mecklin, J.-P.; Watson, P.; Utsunomiya, J.; Bertario, L.; Lynch, P.; Svendsen, L.B.; Cristofaro, G.; Müller, H.; Khan, M.P.; et al. Surveillance in hereditary nonpolyposis colorectal cancer. Dis. Colon Rectum 1993, 36, 425–429. [CrossRef] [PubMed] 14. De Vos tot Nederveen Cappel, W.H.; Nagengast, F.M.; Griﬃoen, G.; Menko, F.H.; Taal, B.G.; Kleibeuker, J.H.; Vasen, H.F. Surveillance for Hereditary Nonpolyposis Colorectal Cancer. Dis. Colon Rectum 2002, 45, 1588–1594. [CrossRef] [PubMed] 15. Roh, S.J.; Hong, Y.H.; Kim, B.C.; Chang, H.J.; Han, K.S.; Hong, C.W.; Sohn, D.K.; Park, S.C.; Lee, D.W.; Kim, B.; et al. References Analysis of metachronous colorectal neoplasms and survival following segmental or extended resection in patients with hereditary non-polyposis colorectal cancer. Int. J. Color. Dis. 2020, 35, 1273–1282. [CrossRef] [PubMed] 16. Heneghan, H.M.; Martin, S.T.; Winter, D.C. Segmental vs extended colectomy in the management of hereditary nonpolyposis colorectal cancer: A systematic review and meta-analysis. Color. Dis. 2015, 17, 382–389. [CrossRef] [PubMed] 17. Anele, C.; Adegbola, S.O.; Askari, A.; Rajendran, A.; Clark, S.K.; Latchford, A.; Faiz, O. Risk of metachronous colorectal cancer following colectomy in Lynch syndrome: A systematic review and meta-analysis. Color. Dis. 2017, 19, 528–536. [CrossRef] 18. Kalady, M.F.; McGannon, E.; Vogel, J.D.; Manilich, E.; Fazio, V.W.; Church, J.M. Risk of colorectal adenoma and carcinoma after colectomy for colorectal cancer in patients meeting Amsterdam criteria. Ann. Surg. 2010, 252, 507–511. [CrossRef] 19. Natarajan, N.; Watson, P.; Silva-Lopez, E.; Lynch, H.T. Comparison of Extended Colectomy and Limited Resection in Patients with Lynch Syndrome. Dis. Colon Rectum 2010, 53, 77–82. [CrossRef] 20. Parry, S.; Win, A.K.; Parry, B.; Macrae, F.A.; Gurrin, L.C.; Church, J.M.; Baron, J.A.; Giles, G.G.; Leggett, B.A.; Winship, I.; et al. Metachronous colorectal cancer risk for mismatch repair gene mutation carriers: The advantage of more extensive colon surgery. Gut 2011, 60, 950–957. [CrossRef] 16 of 17 16 of 17 Cancers 2020, 12, 3419 21. Stupart, D.; Goldberg, P.A.; Baigrie, R.J.; Algar, U.; Ramesar, R. Surgery for colonic cancer in HNPCC: Total vs segmental colectomy. Color. Dis. 2011, 13, 1395–1399. [CrossRef] [PubMed] 22. Aronson, M.; Holter, S.; Semotiuk, K.; Winter, L.; Pollett, A.; Gallinger, S.; Cohen, Z.; Gryfe, R. DNA Mismatch Repair Status Predicts Need for Future Colorectal Surgery for Metachronous Neoplasms in Young Individuals Undergoing Colorectal Cancer Resection. Dis. Colon Rectum 2015, 58, 645–652. [CrossRef] [PubMed] 23. Kim, T.J.; Kim, E.R.; Hong, S.N.; Kim, Y.-H.; Huh, J.W.; Park, Y.A.; Cho, Y.B.; Yun, S.H.; Kim, H.C.; Lee, W.Y.; et al. Survival Outcome and Risk of Metachronous Colorectal Cancer After Surgery in Lynch Syndrome. Ann. Surg. Oncol. 2016, 24, 1085–1092. [CrossRef] [PubMed] 24. Hiatt, M.J.; Casey, M.J.; Lynch, H.T.; Snyder, C.L.; Stacey, M.; Walters, R.W. Eﬃcacy of proximal colectomy for surgical management of right-sided ﬁrst colorectal cancer in Lynch Syndrome mutation carriers. Am. J. Surg. 2018, 216, 99–105. [CrossRef] [PubMed] 25. Renkonen-Sinisalo, L.; Seppälä, T.T.; Järvinen, H.J.; Mecklin, J.-P. Subtotal Colectomy for Colon Cancer Reduces the Need for Subsequent Surgery in Lynch Syndrome. Dis. Colon Rectum 2017, 60, 792–799. [CrossRef] [PubMed] 26. References Giardiello, F.M.; Allen, J.I.; Axilbund, J.E.; Boland, C.R.; Burke, C.A.; Burt, R.W.; Church, J.M.; Dominitz, J.A.; Johnson, D.A.; Kaltenbach, T.; et al. Guidelines on Genetic Evaluation and Management of Lynch Syndrome. Dis. Colon Rectum 2014, 109, 1159–1179. [CrossRef] 27. Syngal, S.; Brand, R.E.; Church, J.M.; Giardiello, F.M.; Hampel, H.L.; Burt, R.W. ACG Clinical Guideline: Genetic Testing and Management of Hereditary Gastrointestinal Cancer Syndromes. Am. J. Gastroenterol. 2015, 110, 223–262. [CrossRef] 28. Stjepanovic, N.; Moreira, L.; Carneiro, F.; Balaguer, F.; Cervantes, A.; Balmaña, J.; Martinelli, E. Hereditary gastrointestinal cancers: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann. Oncol. 2019, 30, 1558–1571. [CrossRef] 29. Monahan, K.J.; Bradshaw, N.; Dolwani, S.; DeSouza, B.; Dunlop, M.G.; East, J.E.; Ilyas, M.; Kaur, A.; Lalloo, F.; Latchford, A.; et al. Guidelines for the management of hereditary colorectal cancer from the British Society of Gastroenterology (BSG)/Association of Coloproctology of Great Britain and Ireland (ACPGBI)/United Kingdom Cancer Genetics Group (UKCGG). Gut 2019, 69, 411–444. [CrossRef] 30. Seppälä, T.T.; Latchford, A.; Negoi, I.; Soares, A.S.; Jimenez-Rodriguez, R.; Sánchez-Guillén, L.; Evans, D.G.; Ryan, N.; Crosbie, E.J.; Dominguez-Valentin, M.; et al. European guidelines from the EHTG and ESCP for Lynch syndrome: An updated third edition of the Mallorca guidelines based on gene and gender. BJS 2020. online ahead of print. [CrossRef] 31. Møller, P.; Seppälä, T.; Bernstein, I.; Holinski-Feder, E.; Sala, P.; Evans, D.G.; Lindblom, A.; Macrae, F.; Blanco, I.; Sijmons, R.; et al. Cancer incidence and survival in Lynch syndrome patients receiving colonoscopic and gynaecological surveillance: First report from the prospective Lynch syndrome database. Gut 2017, 66, 464–472. [CrossRef] [PubMed] 32. Crosbie, E.J.; Ryan, N.A.J.; Arends, M.J.; Bosse, T.; Burn, J.; Cornes, J.M.; Crawford, R.; Eccles, D.; Frayling, I.M.; Ghaem-Maghami, S.; et al. The Manchester International Consensus Group recommendations for the management of gynecological cancers in Lynch syndrome. Genet. Med. 2019, 21, 2390–2400. [CrossRef] [PubMed] 33. Helder-Woolderink, J.; De Bock, G.; Sijmons, R.; Hollema, H.; Mourits, M.; De Bock, G.H. The additional value of endometrial sampling in the early detection of endometrial cancer in women with Lynch syndrome. Gynecol. Oncol. 2013, 131, 304–308. [CrossRef] [PubMed] 34. Schmeler, K.M.; Lynch, H.T.; Chen, L.; Munsell, M.F.; Soliman, P.T.; Clark, M.B.; Daniels, M.S.; White, K.G.; Boyd-Rogers, S.G.; Conrad, P.G.; et al. Prophylactic Surgery to Reduce the Risk of Gynecologic Cancers in the Lynch Syndrome. N. Engl. J. Med. 2006, 354, 261–269. [CrossRef] 35. References Van Der Groep, P.; Van Der Wall, E.; Van Diest, P.J. Pathology of hereditary breast cancer. Cell. Oncol. 2011, 34, 71–88. [CrossRef] 36. Ahadova, A.; Gallon, R.; Gebert, J.; Ballhausen, A.; Endris, V.; Kirchner, M.; Stenzinger, A.; Burn, J.; Doeberitz, M.V.K.; Bläker, H.; et al. Three molecular pathways model colorectal carcinogenesis in Lynch syndrome. Int. J. Cancer 2018, 143, 139–150. [CrossRef] 37. Ahadova, A.; Doeberitz, M.V.K.; Bläker, H.; Kloor, M. CTNNB1-mutant colorectal carcinomas with immediate invasive growth: A model of interval cancers in Lynch syndrome. Fam. Cancer 2016, 15, 579–586. [CrossRef] 17 of 17 Cancers 2020, 12, 3419 17 of 17 38. Engel, C.; Ahadova, A.; Seppälä, T.T.; Aretz, S.; Bigirwamungu-Bargeman, M.; Bläker, H.; Bucksch, K.; Büttner, R.; Cappel, W.T.D.V.T.N.; Endris, V.; et al. Associations of Pathogenic Variants in MLH1, MSH2, and MSH6 With Risk of Colorectal Adenomas and Tumors and With Somatic Mutations in Patients with Lynch Syndrome. Gastroenterology 2020, 158, 1326–1333. [CrossRef] 39. Burn, J.; Sheth, H.; Elliott, F.; Reed, L.; Macrae, F.; Mecklin, J.-P.; Möslein, G.; McRonald, E.F.; Bertario, L.; Evans, D.G.; et al. Cancer prevention with aspirin in hereditary colorectal cancer (Lynch syndrome), 10-year follow-up and registry-based 20-year data in the CAPP2 study: A double-blind, randomised, placebo-controlled trial. Lancet 2020, 395, 1855–1863. [CrossRef] 40. Rodríguez-Bigas, M.A.; Vasen, H.F.; Pekka-Mecklin, J.; Myrhøj, T.; Rozen, P.; Bertario, L.; Järvinen, H.J.; Jass, J.R.; Kunitomo, K.; Nomizu, T.; et al. Rectal Cancer Risk in Hereditary Nonpolyposis Colorectal Cancer After Abdominal Colectomy. Ann. Surg. 1997, 225, 202–207. [CrossRef] 41. Ghezzi, F.; Uccella, S.; Cromi, A.; Bogani, G.; Donadello, N.; Riva, C. Primary Peritoneal Cancer in Lynch Syndrome. Int. J. Gynecol. Pathol. 2013, 32, 163–166. [CrossRef] [PubMed] 42. Dominguez-Valentin, M.; Seppälä, T.T.; Engel, C.; Aretz, S.; Macrae, F.; Winship, I.; Capella, G.; Thomas, H.; Hovig, E.; Nielsen, M.; et al. Risk-Reducing Gynecological Surgery in Lynch Syndrome: Results of an International Survey from the Prospective Lynch Syndrome Database. J. Clin. Med. 2020, 158, 1326–1333. 3. Kom, E.L.; Graubard, B.I.; Midthune, D. Time-to-Event Analysis of Longitudinal Follow-up of a Sur Choice of the time-scale. Am. J. Epidemiol. 1997, 145, 72–80. [CrossRef] [PubMed] 44. Hurley, M. A reference relative time-scale as an alternative to chronological age for cohorts with long follow-up. Emerg. Themes Epidemiol. 2015, 12, 1–12. [CrossRef] [PubMed] 45. Etchegary, H.; Dicks, E.; Watkins, K.; Alani, S.; Dawson, L. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Decisions about prophylactic gynecologic surgery: A qualitative study of the experience of female Lynch syndrome mutation carriers. Hered. Cancer Clin. Pract. 2015, 13, 1–8. [CrossRef] [PubMed] 46. Moldovan, R.; Keating, S.; Clancy, T. The impact of risk-reducing gynaecological surgery in premenopausal women at high risk of endometrial and ovarian cancer due to Lynch syndrome. Fam. Cancer 2014, 14, 51–60. [CrossRef] 47. Boland, C.R.; Shike, M. Report From the Jerusalem Workshop on Lynch Syndrome-Hereditary Nonpolyposis Colorectal Cancer. Gastroenterology 2010, 138, 2197.e1–2197.e7. [CrossRef] 48. Thompson, B.A.; Insight, O.B.O.; Spurdle, A.B.; Plazzer, J.-P.; Greenblatt, M.S.; Akagi, K.; Al-Mulla, F.; Bapat, B.; Bernstein, I.; Capellá, G.; et al. Application of a 5-tiered scheme for standardized classiﬁcation of 2,360 unique mismatch repair gene variants in the InSiGHT locus-speciﬁc database. Nat. Genet. 2014, 46, 107–115. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aﬃliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4225733803	https://ri.conicet.gov.ar/bitstream/11336/185058/2/CONICET_Digital_Nro.3d9591c4-4bf4-4572-815e-b92c65c80f4b_B.pdf	English	null	Effective description of bistability and irreversibility in apoptosis	Physical review. E	2,021	cc-by	11,972	Effective description of bistability and irreversibility in apoptosis Sol M. Fernández Arancibia,1 Hernán E. Grecco ,2,3,4 and Luis G. Morelli 1,2,4,* 1Instituto de Investigación en Biomedicina de Buenos Aires (IBioBA) – CONICET/Partner Institute of the Max Planck Society, Polo Cientíﬁco Tecnológico, Godoy Cruz 2390, Buenos Aires C1425FQD, Argentina 2Department of Physics, FCEN, University of Buenos Aires, Ciudad Universitaria, 1428 Buenos Aires, Argentina 3IFIBA, CONICET, 1428 Buenos Aires, Argentina 4Max Planck Institute for Molecular Physiology, Department of Systemic Cell Biology, Otto-Hahn-Strasse 11, Dortmund D-44227, Germany Apoptosis is a mechanism of programmed cell death in which cells engage in a controlled demolition and prepare to be digested without damaging their environment. In normal conditions, apoptosis is repressed until it is irreversibly induced by an appropriate signal. In adult organisms, apoptosis is a natural way to dispose of damaged cells and its disruption or excess is associated with cancer and autoimmune diseases. Apoptosis is reg- ulated by a complex signaling network controlled by caspases, specialized enzymes that digest essential cellular components and promote the degradation of genomic DNA. In this work, we propose an effective description of the signaling network focused on caspase-3 as a readout of cell fate. We integrate intermediate network interactions into a nonlinear feedback function acting on caspase-3 and introduce the effect of pro-apoptotic stimuli and regulatory elements as a saturating activation function. We show that activation dynamics in the theory is similar to previously reported experimental results. We compute bifurcation diagrams and obtain cell fate maps describing how stimulus intensity and feedback strength affect cell survival and death fates. These fates overlap within a bistable region that depends on total caspase concentration, regulatory elements, and feedback nonlinearity. We study a strongly nonlinear regime to obtain analytical expressions for bifurcation curves and fate map boundaries. For a broad range of parameters, strong stimuli can induce an irreversible switch to the death fate. We use the theory to explore dynamical stimulation conditions and determine how cell fate depends on stimulation temporal patterns. This analysis predicts a critical relation between transient stimuli intensity and duration to trigger irreversible apoptosis. We derive an analytical expression for this critical relation, valid for short stimuli. Our description provides distinct predictions and offers a framework to study how this signaling network processes different stimuli to make a cell fate decision. DOI: 10.1103/PhysRevE.104.064410 PHYSICAL REVIEW E 104, 064410 (2021) PHYSICAL REVIEW E 104, 064410 (2021) Corresponding author: lmorelli@ibioba-mpsp-conicet.gov.ar Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI. Open access publication funded by the Max Planck Society. Published by the American Physical Society Corresponding author: lmorelli@ibioba-mpsp-conicet.gov.ar Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI. Open access publication funded by the Max Planck Society. I. INTRODUCTION Symbols: Ci and C∗ i are inactive and active forms of caspase i; E and I (green) are extrinsic and intrinsic stimuli; Bar and XIAP (red) are inhibitors of apoptosis. (b) Reduced model centered on caspase-3. Other network interactions are repre- sented as an effective positive feedback, and S (green) stands for a generic stimulus. (a),(b) Arrows indicate direct (solid) and indirect (dashed) caspase activation. Blunt arrows (red) indicate inhibition. Hollow arrows indicate apoptotic outcome. (a) Apoptosis Bar XIAP MOMP E I a strong enough stimulus would trigger a switch to the death state. Additionally, to avoid genomic instability and tumor propagation, commitment to cell death should be irreversible once triggered. This complex signaling network has been modeled with different degrees of detail, from very detailed descriptions spanning the whole network [17–20] to focused descriptions of some of the key modules that compose the net- work [21–24]. However, the complexity of this network may obscure the key features that provide it with both tolerance to weak stimuli and commitment after apoptosis initiation, and how such features and stimulus intensity combine to determine a cell fate map. Additionally, it is unclear how such features regulate this fate decision in the presence of more complex, time-dependent stimuli. (b) (a) (b) Apoptosis S 1. (a) Schematic representation of the core signaling n FIG. 1. (a) Schematic representation of the core signaling net- work controlling apoptosis. Symbols: Ci and C∗ i are inactive and active forms of caspase i; E and I (green) are extrinsic and intrinsic stimuli; Bar and XIAP (red) are inhibitors of apoptosis. (b) Reduced model centered on caspase-3. Other network interactions are repre- sented as an effective positive feedback, and S (green) stands for a generic stimulus. (a),(b) Arrows indicate direct (solid) and indirect (dashed) caspase activation. Blunt arrows (red) indicate inhibition. Hollow arrows indicate apoptotic outcome. In this work, we propose an effective description of caspase dynamics that is endowed with bistability and irreversibility. We use this theory to construct fate maps that reveal how these properties are regulated, and to study the response to time-dependent stimulation protocols. We focus on caspase-3 as a readout of cell fate since its drastic increase signals the onset of apoptosis. We integrate intermediate steps of the net- work into a nonlinear feedback function centered on caspase-3 together with a direct stimulus producing caspase-3 activa- tion; see Fig. 1(b). I. INTRODUCTION Inactive caspase precursors, termed procaspases, lack the capacity to cleave their substrates. Procaspases are conformed by two subunits and become active caspases by cleavage of the intersubunit linker [10]. Caspases involved in apoptosis can be broadly classiﬁed into two groups: initiators caspase-8 and caspase-9, and effectors caspase-6, caspase-3, and caspase-7. Initiator caspases cleave effector procaspases, activating them [10]. Effector caspases are considered the executioners of apoptosis: they cleave structural proteins, leading to a con- trolled cell demolition. Cells have intrinsic control mechanisms that can trigger cell death programs, such as apoptosis [1]. During apoptosis, cells undergo an orderly demolition to avoid inﬂammation in the surrounding tissues [2]. Apoptosis plays key roles during embryonic development, shaping tissues [3,4], and in adult organisms, removing damaged cells [5] and controlling tissue homeostasis [1]. Impairment or malfunction of apoptosis is associated with a wide range of developmental abnormalities and diseases such as cancer and autoimmune disorders [6,7]. Both extrinsic and intrinsic stimuli can activate these en- zymes, eliciting the sequence of events that trigger apoptosis. An extrinsic stimulus is the binding of an extracellular ligand, generally from the tumor necrosis factor family, to speciﬁc transmembrane death receptors [11]. This promotes the as- sembly of a membrane bound signaling complex that cleaves and activates procaspase-8 [12]; see Fig. 1(a). Intrinsic stim- uli are of intracellular origin, such as cell damage following γ and UV radiation, oxidative stress, oncogene activation, and other events that compromise the cell integrity [11]. These intrinsic stimuli induce mitochondrial outer membrane permeabilization (MOMP), a process that results in the re- lease of pro-apoptotic proteins into the cytosol promoting Apoptosis is regulated by a signaling network composed of specialized enzymes termed caspases, that cleave essential components of the cell and promote DNA degradation [2,8,9]. *Corresponding author: lmorelli@ibioba-mpsp-conicet.gov.ar Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI. Open access publication funded by the Max Planck Society. 064410-1 Published by the American Physical Society 2470-0045/2021/104(6)/064410(10) PHYSICAL REVIEW E 104, 064410 (2021) ARANCIBIA, GRECCO, AND MORELLI (b) (a) Apoptosis S Apoptosis Bar XIAP MOMP E I FIG. 1. (a) Schematic representation of the core signaling net- work controlling apoptosis. II. THEORY Caspase-3, an effector caspase at the onset of apoptosis that digests cellular components, is a good indicator of cell state. Thus, we focus the theory on this particular enzyme and describe the network interaction as nonlinear feedback and activation functions. We interpret the steady state of active caspase-3 concentration as the readout of the signaling net- work: steady high concentrations represent cell death, while steady low concentrations represent cell survival. Importantly, besides cleaving essential cellular substrates, effector caspase-3 also activates an effector caspase-6, that in turn activates initiator caspase-8 [16]; see Fig. 1(a). This mechanism closes a self-amplifying positive feed- back loop within the enzymatic network: caspase-3 activates procaspase-6, caspase-6 activates procaspase-8, and caspase-8 activates procaspase-3. Interestingly, besides a direct activa- tion of procaspase-3, caspase-8 can also promote MOMP, subsequently activating procaspase-3 through a less direct path. This second path also closes into a positive feedback loop: caspase-3 activates procaspase-6, caspase-6 activates procaspase-8, caspase-8 induces MOMP and procaspase-9 activation, and caspase-9 activates procaspase-3. Thus, these two paths involve procaspase-3 into parallel positive feedback loops. We propose a two-variable description for the concen- trations of the inactive and active caspase-3, C3 and C∗ 3, respectively, ˙C3 = b −dC3 −ksS C3 C3 + K −ψ C∗ 3 h C∗ 30 h + C∗ 3 h C3 C3 + K , (1) ˙C∗ 3 = −dC∗ 3 + ksS C3 C3 + K + ψ C∗ 3 h C∗ 30 h + C∗ 3 h C3 C3 + K . (2) ˙C3 = b −dC3 −ksS C3 C3 + K −ψ C∗ 3 h C∗ 30 h + C∗ 3 h C3 C3 + K , (1) ˙C∗ 3 = −dC∗ 3 + ksS C3 C3 + K + ψ C∗ 3 h C∗ 30 h + C∗ 3 h C3 C3 + K . (2) (2) This complex signaling network controlling apoptosis inte- grates information from the environment and cell state to take a critical cell fate decision. In normal conditions, a certain tolerance to weak stimuli should be possible, to avoid initi- ating self-destruction due to random ﬂuctuations. Bistability would allow for this tolerance of the survival state since only Inactive caspase-3 is synthesized at a constant basal rate b and degraded at a rate d [Eq. (1)]. Here we assume that active caspase-3 is degraded at the same rate d for simplic- ity [Eq. (2)]. I. INTRODUCTION We describe regulatory interactions from Bar and XIAP implicitly through saturating kinetics. We ﬁrst explore the dynamics of caspase activation in the theory and show similarities with previous experiments. We then obtain a cell fate map indicating regions of cell survival, cell death, and the bistable region where these overlap. We determine how this fate map changes with parameters that have a biological interpretation. Furthermore, we show that above some critical feedback strength within the bistable region, the system can switch irreversibly from survival to death state. Finally, we use the theory to simulate dynamic stimulation conditions and predict cell fate outcomes in different scenarios. procaspase-9 activation. Both caspase-8 and caspase-9 cleave and activate effector procaspase-3 and procaspase-7, initiating apoptosis. Given the structural and functional similarity of caspase-3 and caspase-7, here we focus on caspase-3 alone for simplicity [13]. To prevent the cell from dying in normal conditions, in- hibitory molecules keep the onset of apoptosis in check. Proteins inhibiting apoptosis act at several points in the network, binding to active caspases and suppressing their function. The bifunctional apoptosis regulator (Bar) prevents caspase-8 from activating caspase-3, both by sequestering procaspase-8 and by inhibiting MOMP [14]; see Fig. 1(a). The X-chromosome-linked inhibitor of apoptosis protein (XIAP) binds to a caspase-3 active site, inhibiting its activity and promoting its degradation [15]. II. THEORY Both extrinsic and intrinsic stimuli promote the activation of different initiator caspases that subsequently 064410-2 PHYSICAL REVIEW E 104, 064410 (2021) EFFECTIVE DESCRIPTION OF BISTABILITY AND … leads to effector caspase activation, here represented by caspase-3. Thus, in the model, we consider a single activation mechanism, represented by a loss term in Eq. (1) and a cor- responding gain term in Eq. (2). Caspase-3 activation occurs at a rate ks modulated by stimulus intensity S. Since conver- sion of caspase-3 into its active form depends on substrate availability, we assume saturating kinetics in this term. The functional form of this term is linearly increasing for low cas- pase concentrations and saturating for large concentrations, C3 ≫K. This concentration scale K effectively represents inhibitory mechanisms mediated by proteins such as Bar; see Fig. 1(a). The last term corresponds to activation of caspase-3 due to the positive feedback from its active form mediated by caspase-6 and caspase-8. The feedback strength ψ could be linked to caspase-6 or caspase-8 concentrations, not explicitly included in this description. This positive feedback com- prises several steps and molecular reactions of the underlying network. These reactions involve the formation of macro- molecular complexes that may introduce cooperative effects and nonlinearity. Therefore, feedback is represented here by a sigmoidal Hill function, with Hill exponent h describing the underlying nonlinearities in the network. The feedback kicks in when active caspase concentration C∗ 3 exceeds the threshold C∗ 30. The concentration scale C∗ 30 accounts for regulatory pro- teins that prevent effector caspase activity, such as XIAP. A saturating function modulates the feedback term, accounting for substrate depletion as in the stimulus term. threshold. Renaming the variable t′ →t to alleviate notation, we arrive at the dimensionless equations ˙x = β −x −σ x x + κ −ϕ yh 1 + yh x x + κ , (8) ˙y = −y + σ x x + κ + ϕ yh 1 + yh x x + κ , (9) (8) (9) where dimensionless parameter β is the basal synthesis rate, σ the intensity of the stimulus, ϕ the feedback strength, and κ the relative concentration of inhibitory proteins Bar and XIAP; see Eq. (6). q Introducing the total caspase-3 concentration z = x + y, from Eqs. (8) and (9) we obtain ˙z = β −z, (10) (10) with solution z(t) = β −[β −z(0)]e−t. II. THEORY (4) (5) We deﬁne dimensionless groups II. THEORY (11) (11) Thus, the total caspase-3 concentration is decoupled from other variables and has a ﬁxed point z = β that is approached exponentially within one unit of dimensionless time. Here, we consider a regime in which the cell is in a stationary state of the network upon receiving different stimuli. Then, before the stimulus is applied, the total caspase concentration had enough time to settle in this steady state where z = β is con- stant. In this regime, we can eliminate the inactive caspase-3 concentration described by x in terms of this total caspase-3 concentration, x = β −y. Active caspase-3 dynamics is then described by the single equation Before considering the steady states of this signaling net- work, we render the system dimensionless to reduce the number of parameters. We introduce a concentration scale C0 and a timescale T , and deﬁne dimensionless variables x, y, and t′: ˙y = −y + σ β −y κ + β −y + ϕ yh 1 + yh β −y κ + β −y. (12) (12) C3 = C0x, C∗ 3 = C0y, and t = Tt′. (3) (3) This dimensionless model contains ﬁve parameters: the basal synthesis rate β setting the steady state total caspase concen- tration, the stimulus intensity σ, the feedback strength ϕ, the relative concentration of inhibitory proteins κ, and the expo- nent of the Hill function h controlling feedback steepness. Multiplying Eqs. (1) and (2) by T/C0 to render them dimen- sionless, we obtain, in terms of dimensionless variables, ∂x ∂t′ = bT C0 −dT x −ksST C0 x x + K/C0 −ψT C0 yh C∗ 30 h/C0h + yh x x + K/C0 , ∂y ∂t′ = −dTy + ksST C0 x x + K/C0 + ψT C0 yh C∗ 30 h/C0h + yh x x + K/C0 . In the following, we analyze whether this model is compat- ible with properties observed in the biological context such as bistability and irreversibility. We obtain the dynamics of y(t) by numerical integration of Eq. (12) with the ordinary dif- ferential equation (ODE) integrator from the SciPy integrate submodule. We construct numerical bifurcation diagrams in PYTHON. Stable ﬁxed points are obtained as steady states of the dynamics. Unstable ﬁxed points are obtained by ﬁnding the corresponding root of ˙y = 0, using the fsolve function from the SciPy optimize submodule [25]. III. DYNAMICS OF CASPASE ACTIVATION Phase portrait and bifurcation diagrams. (a)–(c) Phase portrait showing ˙y vs y (solid black line), for three different values of the total caspase concentration β. Stable and unstable ﬁxed points are indicated as full and open blue dots, respectively. (d)–(i) Single parameter bifurcation diagrams showing stable (solid blue lines) and unstable (dashed blue lines) ﬁxed points for (d) the total caspase concentration β, (e) the relative concentration of inhibitory proteins κ, (f) the feedback strength ϕ, and (g)–(i) the stimulus intensity 0 3 y -0.5 0 0.5 ̇y (a) 0 3 y -0.5 0 0.5(b) 0 3 y -0.5 0 0.5(c) 2 5 β 0 3 y (d) 3 11 κ 0 3 y (e) 2 8 φ 0 3 y (f) 0 3 σ 0 3 y (g) 0 3 σ 0 3 y (h) 0 3 σ 0 3 y (i) FIG. 3. Phase portrait and bifurcation diagrams. (a)–(c) Phase portrait showing ˙y vs y (solid black line), for three different values of the total caspase concentration β. Stable and unstable ﬁxed points are indicated as full and open blue dots, respectively. (d)–(i) Single parameter bifurcation diagrams showing stable (solid blue lines) and unstable (dashed blue lines) ﬁxed points for (d) the total caspase concentration β, (e) the relative concentration of inhibitory proteins κ, (f) the feedback strength ϕ, and (g)–(i) the stimulus intensity σ. Gray lines are the analytical solutions obtained in the strongly nonlinear regime; blue lines correspond to h = 5. (g)–(i) correspond to different values of the feedback strength ϕ. Parameters are as in Table I. −2 0 14 time 0 3 y (a) σ = 9 σ = 1.2 σ = 1 −2 0 14 time 0 3 y (b) φ = 9 φ = 6 φ = 1 −2 0 14 time −2 0 2 ̇y (c) feedback stimulus degradation 0 2 time 0 3 y Td (d) 101 102 103 σ 0 2 Td (e) FIG. 2. Dynamics of caspase activation. Temporal evolution of active caspase concentration y(t) for (a) different stimulus inten- sity and (b) feedback strength values, as indicated in plot labels. (c) Derivative ˙y(t) (thick teal line) and separate contributions of degradation, stimulus, and feedback terms, as indicated in the label. The derivative corresponds to the thick blue curve in (a) and (b). (d) Schematic deﬁnition of the delay time Td. III. DYNAMICS OF CASPASE ACTIVATION The red dot marks the point where y reaches half its maximum value. (e) Delay time Td for different stimulus intensity values. Parameters are β = 4, κ = 5, ϕ = 6, σ = 1.2, and h = 5, unless otherwise speciﬁed. −2 0 14 time 0 3 y (a) σ = 9 σ = 1.2 σ = 1 −2 0 14 time 0 3 y (b) φ = 9 φ = 6 φ = 1 −2 0 14 time −2 0 2 ̇y (c) feedback stimulus degradation 0 2 time 0 3 y Td (d) 101 102 103 σ 0 2 Td (e) 0 5(b) 3 (a) 0 5(c) 3(e) 0 3 σ 0 3 y (g) 0 3 σ 0 3 y (h) 3(i) 0 3 σ 0 3 y (i) FIG. 2. Dynamics of caspase activation. Temporal evolution of active caspase concentration y(t) for (a) different stimulus inten- sity and (b) feedback strength values, as indicated in plot labels. (c) Derivative ˙y(t) (thick teal line) and separate contributions of degradation, stimulus, and feedback terms, as indicated in the label. The derivative corresponds to the thick blue curve in (a) and (b). (d) Schematic deﬁnition of the delay time Td. The red dot marks the point where y reaches half its maximum value. (e) Delay time Td for different stimulus intensity values. Parameters are β = 4, κ = 5, ϕ = 6, σ = 1.2, and h = 5, unless otherwise speciﬁed. σ σ σ FIG. 3. Phase portrait and bifurcation diagrams. (a)–(c) Phase portrait showing ˙y vs y (solid black line), for three different values of the total caspase concentration β. Stable and unstable ﬁxed points are indicated as full and open blue dots, respectively. (d)–(i) Single parameter bifurcation diagrams showing stable (solid blue lines) and unstable (dashed blue lines) ﬁxed points for (d) the total caspase concentration β, (e) the relative concentration of inhibitory proteins κ, (f) the feedback strength ϕ, and (g)–(i) the stimulus intensity σ. Gray lines are the analytical solutions obtained in the strongly nonlinear regime; blue lines correspond to h = 5. (g)–(i) correspond to different values of the feedback strength ϕ. Parameters are as in Table I. activation dynamics displays an inﬂection point: after a slow concentration increase, a sharper increase takes place and active caspase-3 concentration attains large values. III. DYNAMICS OF CASPASE ACTIVATION The timing of this bump coincides with the inﬂection, iden- tifying the feedback term as the cause triggering the second, sharper increase in active caspase-3 concentration. We noticed that increasing the stimulus intensity causes the activation inﬂection to happen earlier; see Fig. 2(a). To quantify this effect, we introduce the delay time Td as the time elapsed from the stimulus onset to the point where y attains half its maximum value; see Fig. 2(d). This delay time decreases with increasing stimulus intensity, similar to experimental observations and comprehensive models of the caspase net- work [18]; see Fig. 2(e). The low and high active caspase-3 concentration regimes that we observe in the long term dynamics could be inter- preted as a cell survival and apoptotic fates, respectively. Below, we formalize this further, introducing the ﬁxed points of the dynamics and obtaining fate maps. III. DYNAMICS OF CASPASE ACTIVATION Further increasing stimulus intensity produces an earlier and sharper onset of activation. To determine the role of feedback in the dynamics of caspase-3 activation, we ﬁxed the stimulus inten- sity σ and changed the feedback strength ϕ; see Fig. 2(b). For low feedback strength, active caspase-3 concentration remains low. With larger feedback strength, a sharper increase takes place after the inﬂection point in activation dynamics. This inﬂection and sharper increase of active caspase-3 concen- tration occurs earlier as feedback strength grows further. To unravel the cause of this inﬂection in the dynamics, we plotted the derivative ˙y together with the separate contributions of the three terms in Eq. (12): degradation, stimulus, and feedback; see Fig. 2(c). This analysis shows that after applying the stim- ulus, the derivative gradually decreases and reaches a plateau, that is followed with a bump when the feedback term turns on. The timing of this bump coincides with the inﬂection, iden- tifying the feedback term as the cause triggering the second, sharper increase in active caspase-3 concentration. We noticed that increasing the stimulus intensity causes the activation inﬂection to happen earlier; see Fig. 2(a). To quantify this effect, we introduce the delay time Td as the time elapsed from the stimulus onset to the point where y attains half its maximum value; see Fig. 2(d). This delay time decreases with increasing stimulus intensity, similar to experimental activation dynamics displays an inﬂection point: after a slow concentration increase, a sharper increase takes place and active caspase-3 concentration attains large values. Further increasing stimulus intensity produces an earlier and sharper onset of activation. To determine the role of feedback in the dynamics of caspase-3 activation, we ﬁxed the stimulus inten- sity σ and changed the feedback strength ϕ; see Fig. 2(b). For low feedback strength, active caspase-3 concentration remains low. With larger feedback strength, a sharper increase takes place after the inﬂection point in activation dynamics. This inﬂection and sharper increase of active caspase-3 concen- tration occurs earlier as feedback strength grows further. To unravel the cause of this inﬂection in the dynamics, we plotted the derivative ˙y together with the separate contributions of the three terms in Eq. (12): degradation, stimulus, and feedback; see Fig. 2(c). This analysis shows that after applying the stim- ulus, the derivative gradually decreases and reaches a plateau, that is followed with a bump when the feedback term turns on. III. DYNAMICS OF CASPASE ACTIVATION β ≡bT C0 , σ ≡ksST C0 , ϕ ≡ψT C0 , κ ≡K C0 , (6) We study the dynamics of caspase activation by numerical integration of Eq. (12). This equation describes the response to stimuli of active caspase concentration, a quantity that can be accessed experimentally in living single cells by means of reporters [26] and sensors [27]. We apply a constant stim- ulus at t = 0 and observe the temporal evolution of active caspase-3 concentration y(t). For low stimulus intensity, this concentration increases only slightly up to a steady state where ˙y = 0; see Fig. 2(a). With increasing stimulus intensity, (6) and set the time and concentration scales and set the time and concentration scales T d ≡1 and C∗ 30 C0 ≡1. (7) (7) Thus, time is measured in terms of the inverse degradation rate, and concentrations in terms of the feedback activation 064410-3 ARANCIBIA, GRECCO, AND MORELLI PHYSICAL REVIEW E 104, 064410 (2021) −2 0 14 time 0 3 y (a) σ = 9 σ = 1.2 σ = 1 −2 0 14 time 0 3 y (b) φ = 9 φ = 6 φ = 1 −2 0 14 time −2 0 2 ̇y (c) feedback stimulus degradation 0 2 time 0 3 y Td (d) 101 102 103 σ 0 2 Td (e) FIG. 2. Dynamics of caspase activation. Temporal evolution of active caspase concentration y(t) for (a) different stimulus inten- sity and (b) feedback strength values, as indicated in plot labels. (c) Derivative ˙y(t) (thick teal line) and separate contributions of degradation, stimulus, and feedback terms, as indicated in the label. The derivative corresponds to the thick blue curve in (a) and (b). (d) Schematic deﬁnition of the delay time Td. The red dot marks the point where y reaches half its maximum value. (e) Delay time Td for different stimulus intensity values. Parameters are β = 4, κ = 5, ϕ = 6, σ = 1.2, and h = 5, unless otherwise speciﬁed. 0 3 y -0.5 0 0.5 ̇y (a) 0 3 y -0.5 0 0.5(b) 0 3 y -0.5 0 0.5(c) 2 5 β 0 3 y (d) 3 11 κ 0 3 y (e) 2 8 φ 0 3 y (f) 0 3 σ 0 3 y (g) 0 3 σ 0 3 y (h) 0 3 σ 0 3 y (i) FIG. 3. IV. PHASE PORTRAIT AND BIFURCATION DIAGRAM In the one-dimensional representation given by Eq. (12), the ﬁxed points of the dynamics are the solutions to ˙y = 0; see Figs. 3(a)–3(c). Increasing the parameter β we ﬁnd (i) a unique stable ﬁxed point with low active caspase-3 concen- tration [Fig. 3(a)], (ii) two stable ﬁxed points separated by an unstable ﬁxed point [Fig. 3(b)], and (iii) one stable ﬁxed point with high active caspase-3 concentration [Fig. 3(c)]. Thus, the model has a built-in bistable regime, that is, there are parameter values that allow for two possible steady states; see Fig. 3(b). We interpret a steady state with low active caspase-3 concentration as a cell survival outcome, and a state with high active caspase-3 concentration as an apoptotic fate. 064410-4 PHYSICAL REVIEW E 104, 064410 (2021) EFFECTIVE DESCRIPTION OF BISTABILITY AND … TABLE I. Fig. 3 parameters. In all cases, h = 5. Fig. 3 parameters Panel β κ ϕ σ (a) 2 5 6 1 (b) 3 5 6 1 (c) 5 5 6 1 (d) [2,5] 5 6 1 (e) 4 [3,11] 6 1 (f) 4 5 [2,8] 1 (g) 4 5 2 [0,3] (h) 4 5 4 [0,3] (i) 4 5 6 [0,3] (f) 0 7 0 3 0 7 0 3 = 3 = 4 = 9 0 7 0 3 = 1 = 5 = 9 h 2 3 5 10 50 1 7 0 3(a) 1 7 0 3(b) 1 7 0 3 death bistability survival (c) 0 2 y (d) (e) FIG. 4. Bistability and irreversibility in the two-parameter bifur- cation diagram for the stimulus intensity σ and feedback strength ϕ. (a), (b) Stationary value of active caspase concentration y as indicated in the color scale bar. White lines delimit the bistable region: solid white lines mark the bifurcations and dashed white lines are shown for reference. (a) and (b) highlight the two possible values of the steady states in the bistable region located in the bottom right corner. Black triangles mark the feedback strength values from Figs. 3(g)– 3(i). (c) Cell fate map with bistable region shaded in gray. Vertical red line indicates the irreversibility threshold for feedback strength ϕi. (d)–(f) Bistable region for different values of (d) the exponent of the feedback Hill function h, (e) the total caspase concentration β, and (f) the relative concentration of inhibitory proteins κ. Thick gray lines in (d) correspond to Eq. IV. PHASE PORTRAIT AND BIFURCATION DIAGRAM (15) obtained in the strongly nonlinear regime h ≫1, here labeled as ∞. (a)–(f) Unless otherwise speciﬁed, parameters are β = 4, κ = 5, h = 5. TABLE I. Fig. 3 parameters. In all cases, h = 5. (f) 0 7 0 3 0 7 0 3 = 3 = 4 = 9 0 7 0 3 = 1 = 5 = 9 h 2 3 5 10 50 1 7 0 3(a) 1 7 0 3(b) 1 7 0 3 death bistability survival (c) 0 2 y (d) (e) 1 7 0 3(a) 1 7 0 3(b) 1 7 0 3 death bistability survival (c) 0 y 0 7 0 3 = 3 = 4 = 9 2 y (e) 0 7 0 3 h 2 3 5 10 50 (d) (f) 0 7 0 3 = 1 = 5 = 9 (f) 3 In the strongly nonlinear regime h ≫1, it is possible to obtain analytical expressions for ﬁxed points. In this situation, the feedback term yh/(1 + yh) approximates a Heaviside step function H(y −1) at y = 1. Although this introduces a dis- continuity in Eq. (12), it also simpliﬁes the solutions to ˙y = 0; see Appendix A. For y > 1, the cell death solution is FIG. 4. Bistability and irreversibility in the two-parameter bifur- FIG. 4. Bistability and irreversibility in the two-parameter bifur- cation diagram for the stimulus intensity σ and feedback strength ϕ. (a), (b) Stationary value of active caspase concentration y as indicated in the color scale bar. White lines delimit the bistable region: solid white lines mark the bifurcations and dashed white lines are shown for reference. (a) and (b) highlight the two possible values of the steady states in the bistable region located in the bottom right corner. Black triangles mark the feedback strength values from Figs. 3(g)– 3(i). (c) Cell fate map with bistable region shaded in gray. Vertical red line indicates the irreversibility threshold for feedback strength ϕi. (d)–(f) Bistable region for different values of (d) the exponent of the feedback Hill function h, (e) the total caspase concentration β, and (f) the relative concentration of inhibitory proteins κ. Thick gray lines in (d) correspond to Eq. (15) obtained in the strongly nonlinear regime h ≫1, here labeled as ∞. (a)–(f) Unless otherwise speciﬁed, parameters are β = 4, κ = 5, h = 5. IV. PHASE PORTRAIT AND BIFURCATION DIAGRAM yD = 1 2 β + κ + σ + ϕ − (β + κ + σ + ϕ)2 −4β(σ + ϕ) , (13) (13) while for y < 1, the feedback term vanishes and the cell sur- vival solution is yS = 1 2 β + κ + σ − (β + κ + σ )2 −4βσ . (14) (14) This analysis shows that in the strongly nonlinear regime, the system is bistable over wide ranges of parameter values; see gray lines in Figs. 3(d)–3(i). While this strongly nonlinear regime provides valuable insight, it is important to consider lower values of h for their biological relevance. Away from bifurcations, the steady states for ﬁnite h coincide with the an- alytical solutions obtained for the strongly nonlinear regime; see blue lines in Figs. 3(d)–3(i). Near bifurcations, the dif- ference with the strongly nonlinear case becomes prominent. However, the existence of a bistable regime is preserved with moderate nonlinearity. This analysis shows that in the strongly nonlinear regime, the system is bistable over wide ranges of parameter values; see gray lines in Figs. 3(d)–3(i). While this strongly nonlinear regime provides valuable insight, it is important to consider lower values of h for their biological relevance. Away from bifurcations, the steady states for ﬁnite h coincide with the an- alytical solutions obtained for the strongly nonlinear regime; see blue lines in Figs. 3(d)–3(i). Near bifurcations, the dif- ference with the strongly nonlinear case becomes prominent. However, the existence of a bistable regime is preserved with moderate nonlinearity. above a critical stimulus intensity, the cell goes into the death state without the possibility to return [28]. V. FATE MAPS We systematically explored the parameters that allow for this kind of behavior constructing two-parameter bifurcation diagrams; see Fig. 4. The space deﬁned by the feedback strength and stimulus intensity is divided into three regions: a region where active caspase concentration is low, a region where it is high, and a region where both low and high steady state values coexist; see Figs. 4(a) and 4(b). We can interpret this as a cell fate map with a cell survival region, a cell death region, and a region of coexistence where both fates are possible; see Fig. 4(c). This bistable region is delimited by two bifurcation curves that converge into a single point, forming a cusp. Beyond a critical feedback strength ϕi, the switch between survival and death becomes irreversible: after visiting a high caspase concentration state, reducing σ from large values cannot cause a transition back from high to low active caspase concentrations. We are interested in the response of the system to stimuli, here parametrized by the stimulus intensity σ. For low feed- back strength ϕ, we observe that an increasing stimulus causes a smooth crossover from low to high active caspase concentra- tions; see Fig. 3(g). A larger feedback strength causes a region of bistability in terms of the stimulus intensity σ; see Fig. 3(h). This means that low stimuli will not affect the output of the system signiﬁcantly, yet increasing the stimulus over some critical threshold can cause an abrupt change of the active caspase concentration and trigger cell death. Once in this large active caspase concentration state, a moderate reduction in the stimulus does not bring the cell back to the survival state. However, a strong reduction in the stimulus could still bring this state back for this feedback strength. For even larger val- ues of the feedback strength, the switch becomes irreversible since even reducing the stimulus back to zero cannot bring back the cell from the death state; see Fig. 3(i). Thus, a large feedback strength allows for irreversible apoptotic switch: The bistable region exists even for low nonlinearity and its size increases with nonlinearity h; see Fig. 4(d). V. FATE MAPS In the strongly nonlinear regime h ≫1, we can obtain analytical expressions for the bifurcation curves delimiting the bistable 064410-5 PHYSICAL REVIEW E 104, 064410 (2021) ARANCIBIA, GRECCO, AND MORELLI 0 time 0 0 (t) (a) 1 7 0 3 0 (b) = 1 1 7 (c) = 3 1 7 (d) = 20 0 5 0 4 0 (g) 0 2 0 3 y (e) 0 2 (f) survival death 10 1 100 101 103 0 (h) 1 1 2 2 3 3 4 4 5 FIG. 5. Transient stimulation. (a) Time-dependent stimulus of intensity σ0 and duration τ. (b)–(d) Cell fate maps for the stimulus in- tensity σ0 and feedback strength ϕ, with increasing stimulus duration τ as indicated. The solid white line marks the critical lower boundary of the cell death region, σc(ϕ). The dashed white line indicates the bistable region for reference. The red vertical line in (d) indicates the feedback threshold ϕi for irreversibility. (e),(f) Schematic trajec- tories in the steady state bifurcation diagram for stimulus intensity σ, during transient perturbations that ﬁnish in (e) cell survival and (f) cell death. In both cases, the system is initially in the survival steady state y = 0 (1), and a stimulus of intensity σ0 is applied (2). While the stimulus lasts, active caspase concentration y increases (3). The stimulus is removed after the time τ (4). If the value of y has exceeded the value of the unstable ﬁxed point for σ = 0 (y∗, open blue dot), y keeps increasing up to the apoptotic steady state value (5) in (f). If not, it falls back to the survival steady state y = 0 in (e). (g),(h) Cell fate map for the stimulus intensity σ0 and duration τ, for a feedback strength ϕ = 6, in (g) linear and (h) log-log scales. The white line indicates the critical curve that separates the survival state (violet) from the death state (green). The black line indicates the critical curve obtained analytically for short stimuli [Eq. (17)]. (b)–(d),(g),(h) Color indicates the stationary value of active caspase concentration y with scale as in Fig. 4. Parameter values are β = 4, κ = 5, and h = 5. region. In this regime, ﬁxed points representing death and survival states are given by Eq. (13) and Eq. (14). V. FATE MAPS Parameter values are β = 4, κ = 5, and h = 5. FIG. 5. Transient stimulation. (a) Time-dependent stimulus of intensity σ0 and duration τ. (b)–(d) Cell fate maps for the stimulus in- tensity σ0 and feedback strength ϕ, with increasing stimulus duration τ as indicated. The solid white line marks the critical lower boundary of the cell death region, σc(ϕ). The dashed white line indicates the bistable region for reference. The red vertical line in (d) indicates the feedback threshold ϕi for irreversibility. (e),(f) Schematic trajec- tories in the steady state bifurcation diagram for stimulus intensity σ, during transient perturbations that ﬁnish in (e) cell survival and (f) cell death. In both cases, the system is initially in the survival steady state y = 0 (1), and a stimulus of intensity σ0 is applied (2). While the stimulus lasts, active caspase concentration y increases (3). g We have shown that the system displays a robust bistable regime over a wide range of parameter values. Such bistability means that small stimuli will not signiﬁcantly change the concentration of active caspase, ensuring the robustness of the survival state. A large stimulus will push the system over the threshold and trigger apoptosis through the conversion to active caspase that elicits the cell death response. There is also a wide region in parameter space where the system is irreversible and cannot switch back from high to low ac- tive caspase concentrations when the stimulus is reduced. We argue that irreversibility is a natural requirement from the biological perspective since apoptosis involves processes that damage cell components permanently. Thus, once apoptosis is initiated, there should not be a way back since its arrest would leave the cell nonfunctional. come of apoptosis, the apoptotic network should distinguish these situations. V. FATE MAPS For x < κ, the conversion rate factor is smaller than 1/2, and a larger stimulus intensity σ is required to trigger cell death. For κ < x, the conversion rate factor approaches 1 and a smaller stimulus intensity σ is enough to trigger cell death. Thus, the relative values of β and κ determine how fast a stimulus will grow. 0 5 0 4 0 (g) 0 2 0 3 y (e) 0 2 (f) survival death 10 1 100 101 103 0 (h) 1 1 2 2 3 3 4 4 5 0 5 0 4 0 (g) 10 1 100 101 103 0 (h) 0 2 0 3 y (e) 0 2 (f) survival death 1 1 2 2 3 3 4 4 5 (h) 3 (e) (f) FIG. 5. Transient stimulation. (a) Time-dependent stimulus of intensity σ0 and duration τ. (b)–(d) Cell fate maps for the stimulus in- tensity σ0 and feedback strength ϕ, with increasing stimulus duration τ as indicated. The solid white line marks the critical lower boundary of the cell death region, σc(ϕ). The dashed white line indicates the bistable region for reference. The red vertical line in (d) indicates the feedback threshold ϕi for irreversibility. (e),(f) Schematic trajec- tories in the steady state bifurcation diagram for stimulus intensity σ, during transient perturbations that ﬁnish in (e) cell survival and (f) cell death. In both cases, the system is initially in the survival steady state y = 0 (1), and a stimulus of intensity σ0 is applied (2). While the stimulus lasts, active caspase concentration y increases (3). The stimulus is removed after the time τ (4). If the value of y has exceeded the value of the unstable ﬁxed point for σ = 0 (y∗, open blue dot), y keeps increasing up to the apoptotic steady state value (5) in (f). If not, it falls back to the survival steady state y = 0 in (e). (g),(h) Cell fate map for the stimulus intensity σ0 and duration τ, for a feedback strength ϕ = 6, in (g) linear and (h) log-log scales. The white line indicates the critical curve that separates the survival state (violet) from the death state (green). The black line indicates the critical curve obtained analytically for short stimuli [Eq. (17)]. (b)–(d),(g),(h) Color indicates the stationary value of active caspase concentration y with scale as in Fig. 4. V. FATE MAPS Bifurcation curves mark the boundaries in the fate map where these ﬁxed points cease to exist, at y = 1. Thus, setting yD = 1 and yS = 1 in Eqs. (13) and (14), we obtain 0 time 0 0 (t) (a) 1 7 0 3 0 (b) = 1 1 7 (c) = 3 1 7 (d) = 20 σ = κ β −1 + 1 −ϕ and σ = κ β −1 + 1; (15) (15) see thick gray lines in Fig. 4(d). Increasing the dimensionless synthesis rate β shifts the bistable region towards smaller values of σ and ϕ; see Fig. 4(e). The total amount of caspase-3 in the steady state is z = β. In the absence of stimuli, all caspase-3 is in its inactive form, x = β and y = 0. Thus, in the absence of stimuli, inactive caspase levels are larger for larger β, constituting a large pool of enzyme waiting to be converted into the active form. Increasing the relative inhibitor concentration κ shifts the bistable region towards larger values of σ and ϕ; see Fig. 4(f). The factor x/(κ + x) modulates the stimulus of intensity σ for converting caspase into the active form [Eq. (8)]. For x < κ, the conversion rate factor is smaller than 1/2, and a larger stimulus intensity σ is required to trigger cell death. For κ < x, the conversion rate factor approaches 1 and a smaller stimulus intensity σ is enough to trigger cell death. Thus, the relative values of β and κ determine how fast a stimulus will grow. see thick gray lines in Fig. 4(d). Increasing the dimensionless synthesis rate β shifts the bistable region towards smaller values of σ and ϕ; see Fig. 4(e). The total amount of caspase-3 in the steady state is z = β. In the absence of stimuli, all caspase-3 is in its inactive form, x = β and y = 0. Thus, in the absence of stimuli, inactive caspase levels are larger for larger β, constituting a large pool of enzyme waiting to be converted into the active form. Increasing the relative inhibitor concentration κ shifts the bistable region towards larger values of σ and ϕ; see Fig. 4(f). The factor x/(κ + x) modulates the stimulus of intensity σ for converting caspase into the active form [Eq. (8)]. VI. TIME-DEPENDENT STIMULI We have considered the role of the feedback in generating a bistable regime and irreversibility under constant stimulation conditions. As discussed in Sec. I, apoptosis can be triggered both by external stimuli such as the binding of ligands to death receptors, or by internal stimuli produced by cell states that result from damage to cell components. In both cases, the stimulus can be time dependent rather than permanent. Motivated by these physiologically relevant situations, we are interested in analyzing the response of the system upon transient stimulation regimes. To incorporate such transient stimulation in the theory, we introduce a time dependence in the parameter controlling stimuli intensity. Thus, we ﬁrst examine the conditions for a ﬁnite stimulus to be lethal for the cell. We consider a situation where a cell is in the survival steady state under no stimulation and receives a transient stimulus. For simplicity, here we consider a square-pulse stimulus, that is a transient stimulus σ (t) of ﬁnite duration τ and intensity σ0; see Fig. 5(a). We evaluate the outcome of such a ﬁnite stimulus for different feedback strengths and stimulus intensities, and for varying stimulus duration. We ﬁnd that there is a threshold in the feedback strength to enter the apoptotic fate region; see Figs. 5(b)–5(d). This threshold value coincides with the minimum feedback strength ϕi that renders the system irreversible. Because the stimulus is eventually withdrawn (σ = 0 for t > τ), active caspase concentration will only remain in the apoptotic state in the cases where it is impossible to return to the cell sur- vival state; see Fig. 3(i). For feedback strength ϕ larger than this critical value, apoptosis still depends on the intensity Cells may receive transient strong signals that are meant to induce apoptosis, for example delivered from the surround- ing tissue during development. However, weaker confounding signals may also reach the cell transiently, and in these cases the cell should avoid entering apoptosis. Given the drastic out- 064410-6 PHYSICAL REVIEW E 104, 064410 (2021) EFFECTIVE DESCRIPTION OF BISTABILITY AND … 3.25 3 2.5 2 5.0 1.5 5.12 5.25 5.38 5.5 0 10 2.5 2 1.5 1 1 2 3 4 5 n T/ T/ 0 time 0 o (t) T n = 3 (a) 1 7 0 3 0 (b) (c) 1 19 n 1 10 (d) (e) FIG. 6. Pulsed stimulation. VI. TIME-DEPENDENT STIMULI Above the cusp, there is a critical line σc(ϕ) setting a minimum stimulus intensity to enter apoptosis, depending on the feedback strength. This critical line depends on stim- ulus duration τ; see Figs. 5(b)–5(d). This is because to enter apoptosis, active caspase concentration is required to surpass the value y∗of the unstable ﬁxed point for σ = 0, such that when the stimulus is removed the system is on the death state attraction basin; see Figs. 5(e) and 5(f). Thus, there is a critical curve that separates the survival state from the death state in the fate map for stimulus intensity and duration; see Figs. 5(g) and 5(h). We can approximate the dynamics for small y to obtain a curve σ (τ ) for short stimuli. In the absence of stimuli σ = 0, the steady state is x = β with y = 0 and the feedback term is off in Eq. (12). When a stimulus of intensity σ is applied, active caspase dynamics for short times can be approximated by (c) (a) 10 2.5 2 1.5 1 1 2 3 4 5 n T/ 3 (c) ( ) 0 1 7 0 3 0 (b) 3 (b) (e) 3.25 3 2.5 2 5.0 1.5 5.12 5.25 5.38 5.5 0 (e) T/ 1 19 n 1 10 (d) ˙y ≈−y + σ β β + κ . (16) (16) We can readily integrate Eq. (16) and impose the condition y(τ ) = y∗, the value of the unstable ﬁxed point for σ = 0, to obtain an approximate critical curve, σ (τ ) = y∗β + κ β 1 1 −e−τ ; (17) (17) FIG. 6. Pulsed stimulation. (a) Time-dependent stimulus is a train of n pulses of intensity σ0 and duration τ, separated a time T . (b) Cell fate map for the stimulus intensity σ0 and feedback strength ϕ, following stimuli of n = 3 pulses with T/τ = 2. (c) Cell fate maps as in (b) are arranged in a grid for different number of pulses n and T/τ values. (d) Cell fate map for T/τ and the number of pulses n, following stimuli of intensity σ0 = 2.5 with a feedback strength ϕ = 5.25. (e) Cell fate maps as in (d) are arranged in a grid for different values of stimulus intensity σ0 and feedback strength ϕ. VI. TIME-DEPENDENT STIMULI Both increasing feedback strength and stimulus intensity results in an increase of this critical time separation; see Fig. 6(e). Taken together, these fate maps reﬂect the activation dynamics during a pulse train: if the time interval between pulses is larger than the time it takes to return to the survival state, there is no cumulative effect due to consecutive pulses and the cell may survive. Activation dynamics depends on stimuli intensity and feedback strength and, for that reason, so does the critical time interval; see Fig. 6(e). Fig. 6(c). The changes observed in the minimum stimulus intensity required for apoptosis prompted us to map cell fate in terms of the pulse separation and their number for constant feedback strength and stimulus intensity; see Figs. 6(d) and 6(e). In these maps, we observe that for pulse separation larger than a critical value, the cell survival fate prevails indepen- dently of the number of stimuli pulses; see Fig. 6(d). Both increasing feedback strength and stimulus intensity results in an increase of this critical time separation; see Fig. 6(e). Taken together, these fate maps reﬂect the activation dynamics during a pulse train: if the time interval between pulses is larger than the time it takes to return to the survival state, there is no cumulative effect due to consecutive pulses and the cell may survive. Activation dynamics depends on stimuli intensity and feedback strength and, for that reason, so does the critical time interval; see Fig. 6(e). VI. TIME-DEPENDENT STIMULI In all cases, τ = 1, β = 4, κ = 5, h = 5. Color scale as in Fig. 4. see Appendix B and Fig. 5(h). This analytical insight indicates that there is no critical duration since for a very short stimulus there is always a value of stimulus intensity that triggers apoptosis. Beyond single pulses, cells may receive signals with more complex temporal patterns. Here we consider pulse trains, which provide a systematic way to probe the response to frequency and duration, and could be realized experimentally. We introduce a time-dependent stimulus with a number n of pulses of intensity σ0 and duration τ, that are separated by a time interval T ; see Fig. 6(a). We explore the conditions on the number of pulses n and pulse separation T to trig- ger apoptosis. We ﬁrst map cell fate in terms of feedback strength and stimulus intensity, for pulse trains with different pulse separation and number of pulses; see Figs. 6(b) and 6(c). As we observed for a single pulse, there is a minimum feedback strength required to enter apoptosis, coinciding with the minimum feedback strength needed for the system to be irreversible. When the separation interval between pulses equals the pulse duration, increasing the number of pulses is equivalent to a single pulse of increasing duration; see lower row in Fig. 6(c). For a ﬁxed number of pulses, increasing the pulse separation results in an increasing minimum stimulus intensity for entering apoptosis; see columns in Fig. 6(c). This is because in between pulses, the active caspase concentration drops transiently before the next pulse arrives. For a ﬁxed pulse separation, increasing the number of pulses in the train reduces the minimum stimulus intensity required to enter apoptosis; see rows in Fig. 6(c). As pulse separation keeps increasing, the pulses become effectively isolated and the fate maps are identical to that of a single pulse; see top row in Fig. 6(c). The changes observed in the minimum stimulus intensity required for apoptosis prompted us to map cell fate in terms of the pulse separation and their number for constant feedback strength and stimulus intensity; see Figs. 6(d) and 6(e). In these maps, we observe that for pulse separation larger than a critical value, the cell survival fate prevails indepen- dently of the number of stimuli pulses; see Fig. 6(d). VI. TIME-DEPENDENT STIMULI (a) Time-dependent stimulus is a train of n pulses of intensity σ0 and duration τ, separated a time T . (b) Cell fate map for the stimulus intensity σ0 and feedback strength ϕ, following stimuli of n = 3 pulses with T/τ = 2. (c) Cell fate maps as in (b) are arranged in a grid for different number of pulses n and T/τ values. (d) Cell fate map for T/τ and the number of pulses n, following stimuli of intensity σ0 = 2.5 with a feedback strength ϕ = 5.25. (e) Cell fate maps as in (d) are arranged in a grid for different values of stimulus intensity σ0 and feedback strength ϕ. In all cases, τ = 1, β = 4, κ = 5, h = 5. Color scale as in Fig. 4. and duration of the stimulus. If the intensity of the stimulus falls below the upper boundary of the cusp, active caspase concentration never overcomes the value of the lower stable ﬁxed point and returns to y = 0 when the stimulus is re- moved. Above the cusp, there is a critical line σc(ϕ) setting a minimum stimulus intensity to enter apoptosis, depending on the feedback strength. This critical line depends on stim- ulus duration τ; see Figs. 5(b)–5(d). This is because to enter apoptosis, active caspase concentration is required to surpass the value y∗of the unstable ﬁxed point for σ = 0, such that when the stimulus is removed the system is on the death state attraction basin; see Figs. 5(e) and 5(f). Thus, there is a critical curve that separates the survival state from the death state in the fate map for stimulus intensity and duration; see Figs. 5(g) and 5(h). We can approximate the dynamics for small y to obtain a curve σ (τ ) for short stimuli. In the absence of stimuli σ = 0, the steady state is x = β with y = 0 and the feedback term is off in Eq. (12). When a stimulus of intensity σ is applied, active caspase dynamics for short times can be approximated by and duration of the stimulus. If the intensity of the stimulus falls below the upper boundary of the cusp, active caspase concentration never overcomes the value of the lower stable ﬁxed point and returns to y = 0 when the stimulus is re- moved. VII. DISCUSSION Individual dynamics revealed a snap-action, switchlike be- havior, as well as a variable delay in activation time after stimuli [18,26]. The effective theory we have presented here captures this qualitative dynamics of caspase activation as well as the effect of stimulus intensity on the activation delay. In experiments, a reporter for active caspase concentration grows gradually upon stimulation until a sharper increase occurs. The slow increase followed by an inﬂection point to sharp activation that we describe here is reminiscent of this snap-action behavior reported in experiments. The timing of activation in extrinsic apoptosis and the outcome of cell fate was variable in a population of genetically identical cells [32]. This variability was described in theory by random sampling the initial conditions from experimentally determined proba- bility distributions [20]. While random initial conditions may account for extrinsic noise, the effects of intrinsic noise have also been considered in stochastic simulations [33]. Here we showed that feedback strength can also affect the timing of the inﬂection in activation dynamics. Thus, one possibility is that the variability observed in experiments could result from variability in the state of the network underlying positive feedback. Future work could address the effects of both extrin- sic ﬂuctuations—-by introducing distributions for parameter values, and intrinsic ﬂuctuations—-as an additional dynamic noise term. Previous theoretical descriptions have focused on dif- ferent aspects of the signaling network controlling apopto- sis [17,19,21–24]. The effective mutual activation between caspase-3 and -8 was explicitly described together with in- hibitors Bar and XIAP [21]. This work suggested that control at the level of initiator caspases was key to generate a bistable regime compatible with physiological parameter val- ues. Active caspase-3 is tagged by XIAP for ubiquitination, adding an additional degradation mechanism that might result in an effective degradation rate larger than that of inactive caspase-3 [15]. While this differential degradation has been contemplated in some works [21], here we consider the same degradation rate for active and inactive caspase for simplicity. In our theory, such differential degradation would couple to- tal caspase concentration to the active caspase and therefore change with perturbations from stimuli. It will be interesting to study how this affects steady states and cell responses to stimuli. Bistability has also been addressed in the context of the mitochondrial sector of the network [17,22–24]. VII. DISCUSSION In this work, we present a low-dimensional description of the apoptotic switch focused on caspase-3 dynamics. This description integrates intermediate events in the network into 064410-7 PHYSICAL REVIEW E 104, 064410 (2021) ARANCIBIA, GRECCO, AND MORELLI a nonlinear positive feedback function acting on effector caspase-3, a readout of cell fate. We integrate stimuli that trigger apoptosis and inhibitory elements into a saturating activation function. We explore caspase-3 activation dynamics and ﬁnd that increasing stimulus intensity causes an inﬂec- tion to sharp activation. The timing of this inﬂection depends on the feedback strength. The theory has a robust bistable regime where both survival and death are possible cell fates. In this regime, the system tolerates caspase-3 concentration perturbations, and has a sharp transition to cell death when a threshold is exceeded. For a range of parameter values, this switch to the death fate is irreversible. This may describe com- mitment to cell death once a critical level of cell damage has been traversed. We use this theory to explore the response to time-dependent stimuli. For a single pulse stimulus, we ﬁnd a critical feedback strength that allows for such transient stimuli to permanently trigger cell death. While very short stimuli can trigger apoptosis if their intensity is large enough, there is a critical stimulus intensity required to trigger apoptosis, even for long stimuli duration. Finally, we consider stimulation with pulse trains and ﬁnd that multiple consecutive pulses can trigger apoptosis if the separation between pulses is short enough. ampliﬁer described MOMP and the release of pro-apoptotic components, and the executioner module covered apopto- some formation, and activation of caspase-9 and -3. This work suggested that positive feedback in the initiator module may cause bistability and a sharp activation, while positive feedback in the executioner module may be responsible for irreversible activation. In our effective description, we con- sider a single feedback loop acting on caspase-3, mediated by caspase-6. Feedback is responsible for bistability, and further increasing feedback strength may also cause irreversibility. Alternatively, irreversibility may be induced both by increas- ing the total caspase concentration, here parametrized by β, or by decreasing the relative concentration of inhibitory proteins κ = K/C∗ 30, such as Bar and XIAP. 30 In recent years, experimental methods have allowed the study of apoptosis at the singe cell level, imaging reporters for caspase-3 and caspase-8 substrates, and for MOMP [26]. VII. DISCUSSION Mitochon- drial outer membrane permeabilization (MOMP) is tightly regulated by the balance of pro- and anti-apoptotic pro- teins in the Bcl-2 family [7,29]. Downstream of MOMP, caspase-9 molecules assemble into the apoptosome, where they self-activate [30]. One study proposed that coopera- tive apoptosome assembly may be a robust mechanism for inducing bistability [17]. Another study focused on mutual activation between caspase-9 and -3, and their inhibition by XIAP, which establishes an implicit positive feedback that promotes bistability together with irreversibility [22]. Further works proposed that bistability and irreversibility may arise upstream from mitochondria, as a consequence of interactions among proteins that regulate MOMP [23,24,31]. In subse- quent work, these different aspects of the complex molecular network controlling apoptosis were organized into three mod- ules: initiator, ampliﬁer, and executioner [19]. The initiator included interactions among the Bcl-2 family proteins, the Apoptosis is an attractive model system to study fate deci- sions, with a neatly deﬁned outcome and a solid background understanding of its regulatory network [1,34]. However, some key questions remain open [35], and it has been suggested that the ﬁeld is poised to integrate theory and exper- iment to bring insight into the mechanisms that regulate and trigger apoptosis [36]. The effective description we propose here has the advantage to provide analytic insights into these mechanisms, and makes some distinct predictions. One of the outstanding questions in the ﬁeld is how commitment to cell death is regulated and to what extent a cell may refrain from death once engaged in apoptotic events [35]. Here we showed that in the framework of the theory, there is a critical feedback strength necessary for a cell to irreversibly adopt a death fate. This critical feedback strength may be regulated both by the relative concentration of inhibitory proteins and by total caspase-3 concentration. In the theory, feedback strength is an effective quantity resulting from intermediate network 064410-8 PHYSICAL REVIEW E 104, 064410 (2021) EFFECTIVE DESCRIPTION OF BISTABILITY AND … state is state is components, such as caspase-6. Thus, interfering with caspase-6 function could be one way to bring the cell below the critical feedback strength required for irreversibility. This could be tested using transient stimulation and observing the nature of cell response. For transient stimuli, we showed that there is a critical stimulus intensity to trigger apoptosis. ACKNOWLEDGMENTS We thank F. Fabris, D. M. Arribas, and M. Wappner for providing valuable feedback on the manuscript. This research was sponsored by Agencia Nacional de Promoción Cientí- ﬁca y Tecnológica (ANPCyT) PICT 2013 1301 (H.E.G. and L.G.M.), PICT 2017 3753 (L.G.M.), and FOCEM-Mercosur (COF 03/11) to the IBioBA. yS = 1 2[β + κ + σ − (β + κ + σ )2 −4βσ]. (A9) (A9) VII. DISCUSSION The model predicts the shape of the critical curve between the duration and intensity of transient stimuli that trigger apopto- sis. For pulsed stimulation, the model predicts a characteristic relation between the number of pulses and the time interval between them to trigger apoptosis. These predictions may be tested imaging single cells during apoptosis [26,27,32]. We expect that our model may be a useful tool to interrogate the dynamics of apoptosis and its underpinnings in constant and variable experimental conditions. yD = 1 2[β + κ + σ + ϕ − (β + κ + σ + ϕ)2 −4β(σ + ϕ)]. (A6) (A6) Similarly, the ﬁxed point for y < 1 is the solution to 0 = −y + σ β −y κ + β −y. (A7) (A7) Multiplying again by (κ + β −y) and solving the resulting quadratic equation, we obtain solutions Multiplying again by (κ + β −y) and solving the resulting quadratic equation, we obtain solutions y± = 1 2[β + κ + σ ± (β + κ + σ )2 −4βσ]. (A8) (A8) We discard the unphysical solution y+ because it implies y > β and x < 0, assuming positive values for all parameters. Finally, the ﬁxed point for y < 1, which we interpret as the survival state, is APPENDIX A: ANALYTICAL SOLUTION FOR THE SURVIVAL AND DEATH STATES IN THE STRONGLY NONLINEAR REGIME In the absence of stimuli, the steady state of the system is x = β and y = 0. At this point, the feedback term is turned off and, if a stimulus of strength σ is applied, for very short times active caspase dynamics Eq. (12) can be approximated by Active caspase-3 dynamics is described by the single equa- tion dy dt ≈−y + σ β β + κ , (B1) (B1) ˙y = −y + σ β −y κ + β −y + ϕ yh 1 + yh β −y κ + β −y. (A1) (A1) where we have neglected the feedback term and considered y ≪β in the stimulus term. We separate variables to perform the integration The ﬁxed points of the dynamics are the solutions to ˙y = 0. In the strongly nonlinear regime h ≫1, it is possible to ﬁnd analytical expressions for the ﬁxed points. In this regime, for h ≫1, the sigmoidal Hill function approximates a discontin- uous step function which has a constant value of 1 for y > 1, and 0 for y < 1. y 0 dy′ −y′ + σ β β+κ ≈ t 0 dt′, (B2) (B2) The ﬁxed point for y > 1 is then the solution to resulting in the short term dynamics resulting in the short term dynamics 0 = −y + σ β −y κ + β −y + ϕ β −y κ + β −y. (A2) (A2) y(t) ≈σ β β + κ (1 −e−t ). (B3) (B3) Multiplying by (κ + β −y), we obtain Multiplying by (κ + β −y), we obtain Multiplying by (κ + β −y), we obtain From this short term dynamics, we can approximate a critical curve in a fate map for σ0 and τ, separating the survival and death states after a short transient stimulus. For a given stimulus duration τ, this critical curve is determined by the stimulus intensity that brings y to the value of the unstable ﬁxed point in the absence of stimuli, that we termed y∗. If this value is surpassed, y continues to increase up to the death state after the stimulus is removed. Thus, to obtain an expression for this critical curve, we set y(τ ) = y∗and solve for σ, (A3) 0 = −y(κ + β −y) + σ (β −y) + ϕ(β −y). APPENDIX A: ANALYTICAL SOLUTION FOR THE SURVIVAL AND DEATH STATES IN THE STRONGLY NONLINEAR REGIME (A3) Rearranging terms results in the quadratic expression 0 = y2 −(β + κ + σ + ϕ)y + β(σ + ϕ), (A4) (A4) (A4) 0 = y2 with solutions with solutions with solutions y± = 1 2[β + κ + σ + ϕ ± (β + κ + σ + ϕ)2 −4β(σ + ϕ)]. (A5) (A5) σ (τ ) = y∗β + κ β 1 1 −e−τ . (B4) (B4) We discard the unphysical solution y+ because it implies y > β and x < 0, assuming positive values for all parameters. Finally, the ﬁxed point for y > 1 that we interpret as the death We discard the unphysical solution y+ because it implies y > β and x < 0, assuming positive values for all parameters. Finally, the ﬁxed point for y > 1 that we interpret as the death The value of y∗can be obtained numerically. 064410-9 064410-9 ARANCIBIA, GRECCO, AND MORELLI PHYSICAL REVIEW E 104, 064410 (2021) [1] S. Bedoui, M. J. Herold, and A. Strasser, Nat. Rev. Mol. Cell Biol. 21, 678 (2020). [18] J. G. Albeck, J. M. Burke, S. L. Spencer, D. A. Lauffenburger, and P. K. Sorger, PLoS Biol. 6, e299 (2008). [19] T. Zhang, P. Brazhnik, and J. J. Tyson, Biophys. J. 97, 415 (2009). [2] B. Alberts, D. Bray, J. Lewis, M. Raff, K. Roberts, and J. Watson, Molecular Biology of the Cell, 4th ed. (Garland, New York, 2002). [20] S. Gaudet, S. L. Spencer, W. W. Chen, and P. K. Sorger, PLoS Comput. Biol. 8, e1002482 (2012). [3] P. Meier, A. Finch, and G. Evan, Nature (London) 407, 796 (2000). [21] T. Eissing, H. Conzelmann, E. D. Gilles, F. Allgöwer, E. Bullinger, and P. Scheurich, J. Biol. Chem. 279, 36892 (2004). [4] A. Ambrosini, M. Gracia, A. Proag, M. Rayer, B. Monier, and M. Suzanne, Mechan. Develop. 144, 33 (2017). [22] S. Legewie, N. Blüthgen, and H. Herzel, PLoS Comput. Biol. [5] P. M. Henson, D. L. Bratton, and V. A. Fadok, Curr. Biol. 11, 795 (2001). 2, e120 (2006). [23] C. Chen, J. Cui, H. Lu, R. Wang, S. Zhang, and P. Shen, Biophys. J. 92, 4304 (2007). [6] B. Favaloro, N. Allocati, V. Graziano, C. D. Ilio, and V. D. Laurenzi, Aging 4, 330 (2012). [24] J. Cui, C. Chen, H. Lu, T. Sun, and P. Shen, PLoS One 3(1), e1469 (2008). [7] R. APPENDIX A: ANALYTICAL SOLUTION FOR THE SURVIVAL AND DEATH STATES IN THE STRONGLY NONLINEAR REGIME Singh, A. Letai, and K. Sarosiek, Nat. Rev. Mol. Cell Biol. 20, 175 (2019). [25] P. Virtanen et al., Nat. Methods 17, 261 (2020). [8] N. A. Thornberry, T. A. Rano, E. P. Peterson, D. M. Rasper, T. [26] J. G. Albeck, J. M. Burke, B. B. Aldridge, M. Zhang, D. a. Lauffenburger, and P. K. Sorger, Mol. Cell 30, 11 (2008). [8] N. A. Thornberry, T. A. Rano, E. P. Peterson, D. M. Rasper, T. Timkey, M. Garcia-Calvo, V. M. Houtzager, P. A. Nordstrom, S. Roy, J. P. Vaillancourt, K. T. Chapman, and D. W. Nicholson, Timkey, M. Garcia-Calvo, V. M. Houtzager, P. A. Nordstrom, S. Roy, J. P. Vaillancourt, K. T. Chapman, and D. W. Nicholson, [27] A. A. Corbat, K. C. Schuermann, P. Liguzinski, Y. Radon, P. I. Bastiaens, P. J. Verveer, and H. E. Grecco, Redox Biol. 19, 210 (2018). J. Biol. Chem. 272, 17907 (1997). J. Biol. Chem. 272, 17907 (1997). [9] H. R. Stennicke, M. Renatus, M. Meldal, and G. S. Salvesen, Biochem. J. 350, 563 (2000). [28] J. J. Tyson, K. C. Chen, and B. Novak, Curr. Opin. Cell Biol. 15, 221 (2003). [10] O. Julien and J. A. Wells, Cell Death Differ. 24, 1380 (2017). [29] A. Peña\textrm-Blanco and A. J. García\textrm-Sáez, FEBS J. 285, 416 (2017). [11] S. H. Kaufmann and W. C. Earnshaw, Expt. Cell Res. 256, 42 (2000). ( ) [12] F. Gonzalvez and A. Ashkenazi, Oncogene 29, 4752 (2010). [30] S. M. Srinivasula, M. Ahmad, T. Fernandes-Alnemri, and E. S. Alnemri, Mol. Cell 1, 949 (1998). F. Gonzalvez and A. Ashkenazi, Oncogene 29, 4752 (201 [13] G. S. Salvesen and A. Ashkenazi, Cell 147, 476 (2011). [31] T. Sun, X. Lin, Y. Wei, Y. Xu, and P. Shen, FEBS Lett. 584, 954 (2010). [14] H. Zhang, Q. Xu, S. Krajewski, M. Krajewska, Z. Xie, S. Fuess, S. Kitada, K. Pawłowski, A. Godzik, and J. C. Reed, Proc. Natl. Acad. Sci. USA 97, 2597 (2000). [32] S. L. Spencer, S. Gaudet, J. G. Albeck, J. M. Burke, and P. K. Sorger, Nature (London) 459, 428 (2009). [15] M. Holcik and R. G. Korneluk, Nat. Rev. Mol. Cell Biol. 2, 550 (2001). [33] H. Ooi and L. Ma, BMC Syst. Biol. 7, 65 (2013). [16] V. Cowling and J. Downward, Cell Death Differen. 9, 1046 (2002). [34] R. C. Taylor, S. P. Cullen, and S. J. Martin, Nat. Rev. Mol. APPENDIX A: ANALYTICAL SOLUTION FOR THE SURVIVAL AND DEATH STATES IN THE STRONGLY NONLINEAR REGIME Cell Biol. 9, 231 (2008). [17] E. Z. Bagci, Y. Vodovotz, T. R. Billiar, G. B. Ermentrout, and I. Bahar, Biophys. J. 90, 1546 (2006). [35] D. R. Green, Cell 177, 1094 (2019). [36] J. J. Tyson, Nature (London) 445, 823 (2007). 064410-10 064410-10
https://openalex.org/W3158855825	https://www.researchsquare.com/article/rs-161781/latest.pdf	English	null	Structural, magnetic and electrical properties of RFeO3 (R= Dy, Ho, Yb &amp; Lu) compounds	Research Square (Research Square)	2,021	cc-by	5,952	Structural, magnetic and electrical properties of RFeO3 (R= Dy, Ho, Yb & Lu) compounds Mehrnoush Nakhaei Semnan University Davoud Sanavi Khoshnoud (  d_khoshnood@yahoo.com ) Semnan University https://orcid.org/0000-0001-5423-7351 Mehrnoush Nakhaei1, and Davoud Sanavi Khoshnoud2,* 1,2 Faculty of Physics, Semnan University P. O. Box 35195-363, Semnan, Iran * Corresponding author: dskhoshnoud@semnan.ac.ir 1,2 Faculty of Physics, Semnan University P. O. Box 35195-363, Semnan, Iran 1,2 Faculty of Physics, Semnan University P. O. Box 35195-363, Semnan, Iran * Corresponding author: dskhoshnoud@semnan.ac.ir * Corresponding author: dskhoshnoud@semnan.ac.ir * Tel: +982331533254 * Tel: +982331533254 * fax: +982333654081 Abstract The nanosized rare earth orthoferrites (RFeO3) with R= Dy, Ho, Yb & Lu were synthesis by sol-gel combustion method. The effects of varying rare earth ion on structural, magnetic and electrical properties of RFeO3 nanoparticles (NPs) have been studied. X-ray diffraction patterns of the prepared samples were in single phase with orthorhombic structure with space group Pbnm. The magnetic hysteresis measurements indicated the decrease in the antiferromagnetic nature of RFeO3 from Dy to Lu that can be ascribed to increasing noncollinear Fe moments because of the structural distortion. The electric hysteresis of these samples exhibited narrow electric field-polarization loops. The frequency response of dielectric constant of RFeO3 explained based on dipole relaxation process and Maxwell- Wagner model. In addition, the temperature dependence of dielectric constant illustrated an anomaly in the vicinity of the Neel temperature of RFeO3 compounds. This occurrence indicate that a coupling exist between magnetic and electric orders. RFeO3; Rietveld refinement; M-H loops; P-E measurements; Dielectric constant; Neel temperature. 1. Introduction RFeO3 compounds (R= rare earth elements) have shown both of magnetic and electrical ordering. Because of these features, RFeO3 orthoferrites have been used in many technological and fundamental applications such as their potential applications in spintronic, information storage and sensor technology [1]. There are two magnetic subsystems in RFeO3 compounds: one of them is due to rare earth ion (R3+) moment and the other one is due to iron ion (Fe3+) moment [2,3]. There are three types of magnetic interactions in this family: R3+-R3+, R3+-Fe3+ and Fe3+-Fe3+, each of them generally consists of the isotropic, the antisymmetric and the anisotropic-symmetric superexchange interactions [4]. RFeO3 compounds have perovskite structure with Fe3+ ions were surrounded by six oxygen ions which forming FeO6 octahedral. Beside, these materials exhibit noncollinear antiferromagnetic behavior [4]. The orthorhombic distortions and the super exchange interactions may originate ferroelectricity (FE) in this family similar to the other orthoferrites [5]. In addition, the ferroelectricity in RFeO3 perovskite structure is driven by rare earth ions in R-side [6]. 3d cations in Fe-side induce magnetic ordering in these materials and the spin orders induce multiferroicity because of their large magnetoelectric (ME) coupling [7]. The useful properties of RFeO3 orthoferrites related to the quality of their crystallites. In recent years, synthesis of nano-scaled materials with no impurities attracted a lot of interest because the ultrafine powders usually show better properties [8,9]. Therefore, RFeO3 compounds were obtained by different methods such as: solid state reaction, sonochemical, hydrothermal, combustion methods, polyol, sol-gel [10,11]. In this wok, sol-gel method was used because of its easier, inexpensive and more controllable process [12]. Thereafter, the influence of rare earth elements (R= Dy, Ho, Yb & Lu) on structural, magnetic and electrical properties of RFeO3 nanoparticles (NPs) had been investigated. Keywords RFeO3; Rietveld refinement; M-H loops; P-E measurements; Dielectric constant; Neel temperature. 1 1 2. Experimental procedure In this manuscript, RFeO3 (R= Dy, Ho, Yb & Lu) orthoferrites were synthesized with sol-gel combustion method. For that purpose, rare earth nitrate hydrates (R(NO3)3.xH2O) , iron nitrate hydrate (Fe(NO3)3.9H2O) and citric acid (C6H8O7) were used as prematerials in the stoichiometric ratio. All these used materials were in 99.9% purities from Merck (Darmstadt, Germany). First, 6 mmol citric acid was added into 20 ml deionized water under magnetic stirring at room temperature. Thereafter, the rare earth nitrate hydrates and the iron nitrate hydrate were put into the solution. Afterthat, the ammonia solution was used to set the pH value on 7. Then, for making the solution more homogenous, 2 it was stirred for about 2 h on magnetic stiller. At last, the solution was kept at T= 80°C for about 1 h under stirring to form the gel sightly with olivaceous green color. Afterthat, the achieved gel had been dried at T=120°C for ~12 h. At the end of this process, they were calcined at T=1000°C for 2 h to prepare nanosized RFeO3 orthoferrites. First of all, X-ray diffraction (XRD) pattern was used for the structural characterization and the x-pert and Fullprof softwares were used for investigating these results. With fitting the Williamson-Hall (W-H) curves of the prepared compounds, the crystallite sizes and lattice strains were calculated. Field emission scanning electron microscopy (FE- SEM, LEO 912 AB, Carl Zeiss SMT, Germany, 120 kV) technique was used for investigating the morphology and the size of RFeO3 (R= Dy, Ho, Yb & Lu) particles. For determining the elemental compositions of the samples energy dispersive spectroscopy (EDS) curves were used. The magnetic behavior of the prepared RFeO3 NPs were characterizing using vibration sample magnetometer (VSM) measurement under applied magnetic fields up to 20 kOe at T=300 K. Polarization (P) hysteresis curves in the presence of applied electric field (E) of these samples were measured in this research. The dielectric constant (ε) and dielectric loss (tanδ) of RFeO3 NPs were carried out using LCR meter (GW Instek 8110G). For dielectric measurement, RFeO3 pellets were made in 10.2 mm diameter and 1.6 mm thickness. The prepared pellets had been sintered at T= 1000oC for 2 h to become stable. At last, the pellets were coated by gold. 3.1. Structural studies Fig. 1a illustrates the XRD curves of RFeO3 (R= Dy, Ho, Yb & Lu) compounds. XRD patterns of the samples were comprised with the sample diffraction data (JCPDS files) codes: 00-046-0135, 00-046-0115, 00-047-0070 and 00-047-0071, respectively. These accordances show that the prepared RFeO3 NPs are in single phase with orthorhombic structure with Pbnm space group. Fig. 1b shows the shift of XRD peaks. These shifts were moving to right so according to Bragg’s law, the interplanar distances (d) of RFeO3 compounds should become smaller with decreasing the ionic radius of rare-earth elements [13]. The interplanar distances (d) of the prepared RFeO3 were reported in Table 1. For confirming the structural configuration the Fullprof software was used. Rietveld refinement of XRD patterns had been done according to Pseudo-voigt function and the final curves were shown in Fig. 2a. These refinements confirmed that the prepared samples are in single phase. The a, b & c (unit cell parameters) of RFeO3 were shown in Fig. 2b. As can be seen in Fig. 2b. the unit cell parameters became less in RFeO3 compounds from 3 R=Dy to Lu. Similar results had been seen before in other researches [14-16]. As O2- and Fe3+ ions are constant in all RFeO3 structures of this research, the variations of unit cell parameters can be related to the R3+ cations [15]. The schematic structure and lattice distortions were illustrated in Fig. 1c based on the Rietveld refinement output data of all prepared samples. According to the Rietveld refinement results, the ionic positions (x, y, z), unit cell volume (V), goodness of fit (𝐺𝑂𝐹= !!" !#$"), important bond lengths (Fe-O1 & Fe-O2) and bond angles (Fe-O1-Fe & Fe-O2-Fe) of RFeO3 compounds were shown in Table 1. Tolerance factors (τ) of the samples were shown in Table 1. τ value is using for showing the crystalline structures according to equation 1 [17]: Tolerance factors (τ) of the samples were shown in Table 1. τ value is using for showing the crystalline structures according to equation 1 [17]: 𝜏= (𝑅−𝑂) √2(𝐹𝑒−𝑂) ⁄ (1) 𝜏= (𝑅−𝑂) √2(𝐹𝑒−𝑂) ⁄ (1) In equation 1, R-O and Fe-O are average bond lengths that were reported in Table 1. As the tolerance factor (τ) has to be between 0.71 to 0.99 for orthorhombic structure therefore orthorhombic phase with Pbnm space group had been confirmed for these prepared samples [18]. 3.1. Structural studies For determining the crystallite sizes of the prepared samples the Debye-Scherrer equation was used as follow [19]: 𝐷"#$%& = 𝐾𝜆𝛽𝑐𝑜𝑠𝜃 ⁄ (2) For determining the crystallite sizes of the prepared samples the Debye-Scherrer equation was used as follow [19]: 𝐷"#$%& = 𝐾𝜆𝛽𝑐𝑜𝑠𝜃 ⁄ (2) 𝐷"#$%& = 𝐾𝜆𝛽𝑐𝑜𝑠𝜃 ⁄ (2) here λ is the x-ray wavelength (λCukα = 1.542 Å), β is the full width at half maximum (FWHM), θ is the diffraction angle and K is Scherrer’s constant (K=0.94). In this method, the highest peak was used for calculating the crystallite sizes of the samples. The crystallite sizes of prepared RFeO3 had been shown in Table 2. According to these calculations, it is clear that crystallite sizes of RFeO3 NPs decrease from R= Dy to Lu. With the Stokes and Wilson equation, the lattice strain (g) can be calculated (see Eq. 3) which induced broadening of the Bragg’s diffractions peaks [20]: With the Stokes and Wilson equation, the lattice strain (g) can be calculated (see Eq. 3) which induced broadening of the Bragg’s diffractions peaks [20]: 𝑔= 𝛽4𝑡𝑎𝑛𝜃 ⁄ (3) 𝑔= 𝛽4𝑡𝑎𝑛𝜃 ⁄ The g values of the samples were reported in Table 2. The Williamson-Hall (W-H) method (Eq. 4) can be used for calculating the crystallite sizes (DW-H) and the lattice strains (ε) of the compounds in addition of Scherrer and Stokes and Wilson equations, as follow [21]: 4 4 𝛽cos 𝜃= 𝑘𝜆𝐷 ⁄ + 4𝜀𝑠𝑖𝑛𝜃 (4) Fig. 3. Shows the Williamson-Hall (W-H) plot of the samples. At first, the W-H curves have to be fitted in linear type. Then, the y-intercept determines the crystallite sizes, and the slope of the linear fitting estimates the lattice strains (ε) of the samples. The amounts of the crystallite sizes (DW-H) and the lattice strains (ε) were shown in Table 2. Based on these results, the crystallite sizes of the samples decreased with reducing R3+ ionic radius. Beside the lattice strain (ε) became more with decreasing the crystallite sizes in these compounds. As a result, increase in the lattice strain (ɛ) will decrease the lattice distortions of RFeO3 NPs. FE-SEM images of RFeO3 (R= Dy, Ho, Yb & Lu) compounds were shown in Fig. 4. As can be seen in Fig. 4, the nanoparticles stick to each other. 3.2. Magnetic properties The magnetic hysteresis loops of RFeO3 (R= Dy, Ho, Yb & Lu) compounds were shown in Fig. 6. These measurements were taken under the magnetic field of 20 kOe at T= 300 K. While DyFeO3 and HoFeO3 NPs show narrow hysteresis loops, YbFeO3 and LuFeO3 NPs displayed typical banana and S-like hysteresis loops, respectively. The M-H loops had been magnified around 0 kOe magnetic field and then the values of net maximum magnetization (Mnet), remanent magnetisation (Mr) and coercivity field (Hc) of RFeO3 compounds were given in Table 4. In RFeO3 compounds, it is known that Fe and R sublattices are coupled in an antiferromagnetic order [25]. On the other hand, the R sublattice anisotropy decrease strongly with increasing temperature [26]. Consequently, at room temperature the magnetic behavior of RFeO3 orthoferrites (except for SmFeO3) were determined with Fe magnetic moments [27,28]. Therefore, the existence of hysteresis loop in these compounds can be attributed to noncollinear Fe moments because of Dzyaloshinsky-Moria (DM) interaction [29-31]. According to different values of Fe-O-Fe band angle for RFeO3 samples (see Table. 1) that is due to different distortions of FeO6 octahedral because of various R ions doping, can be expected that ferromagnetic feature increase from R=Dy to Lu. In samples with nano scale particles, another factor that affect the magnetic properties is size effect. In previous report on NdFeO3 NPs were indicated that the values of magnetization enhance with decreasing particles size [12]. This result can be applied based on core-shell model in compounds with antiferromagnetic nature with narrow hysteresis loop as depicted in the M-H loop in sample with R= Dy and Ho [4,32, 33]. Therefore, according to this model with decreasing nano particles size, the magnetization of HoFeO3 sample increase compared to that of DyFeO3 sample. Based on Table 4, the coercivity field (Hc) of the samples became more from R= Dy to Lu. This behavior can be attributed to the enhancement of ferromagnetic nature in these compounds that is related to decrease of Fe-O-Fe bond angles from DyFeO3 to LuFeO3. 3.1. Structural studies The homogeneity of RFeO3 NPs became more from DyFeO3 to LuFeO3 and the size of particles became smaller and the grain boundaries were more distinct, which were consistent with their XRD results. Digimizer software was used for investigating the morphology of the RFeO3 (R= Dy, Ho, Yb & Lu) NPs. Mean particle diameter (<DSEM>) was calculated by fitting the relations of 5, 6 & 7, as follows [17]: 𝑃(𝐷) = (1 √2𝜋𝜎𝐷 ⁄ )exp [−𝑙𝑛'(𝐷𝐷( ⁄ )/2𝜎'] (5) < 𝐷>)+= 𝐷(exp (𝜎' 2 ⁄ ) (6) 𝜎, =< 𝐷>)+ [exp(𝜎') −1]-/' (7) (5) (6) (7) In these equations, D0 and σ show the median diameter and the distribution width, respectively. In Table 2, the mean particle diameters (˂DSEM>) and standard deviations (σD) of the RFeO3 compounds were shown. Icry=(<DSEM>)/(<DXRD>) was used for calculating the crystallinity index which ˂DXRD>= DW-H is average crystallite size that was determined from XRD analysis [22]. Icry is 1 for single crystal structures and it will be bigger for the polycrystalline structures [23]. Crystallinity indexes (Icry) of RFeO3 compounds were reported in Table 2. Based on these values, DyFeO3 (Icry=1.41) is more polycrystalline compared to the other samples. It can be due to the bigger particle size of this compound. In this research, the energy dispersive spectroscopy (EDS) analysis was used for showing the elemental compositions of the compounds. EDS spectrums of the RFeO3 (R= Dy, Ho, Yb & Lu) compounds were shown in Fig. 5. The EDS analysis results supported the elemental compositions of RFeO3 orthoferrites. In this figure, each peak shows the separated element (R & Fe) but the not assigned peaks are of Au element which coated for measuring 5 process. The atomic percent ratio (Fe/R) and atomic weight percent (%) were shown in Table 3. As can be seen in this table, with increasing the atomic weight of these compounds, the atomic percent ratio (Fe/R) became less. The atomic weight percent (%) of R became less with increasing atomic weight approximately because the rare earth elements (R) became heavier From R= Dy to Lu. The oxygen atomic weight percentage (%) was the most one between all elements (Fe3+, R3+ & O2-) in all prepared samples because of the most atomic number and the less atomic weight [24]. 3.4. Dielectric studies 3.4. Dielectric studies 3.3. P-E measurements 6 Fig. 7. shows the polarization-electric field hysteresis loops (P-E) of RFeO3 (R= Dy, Ho, Yb & Lu) compounds under an applied electric field up to 40 kV/cm at 50 Hz at room temperature. The values of remanent polarization (Pr) and electric coercivity (Ec) for RFeO3 compounds were obtained with magnifying the P-E curves around origin (P = 0 and E = 0) and were listed in Table 5. Based on Fig. 7, not any saturation polarization has been found in P-E hysteresis curves which can be explain by Maxwell-Wagner (MW) theorem [34]. The MW effect is due to grains boundary effect at room temperature [35]. In addition, all four hysteresis loops are unclosed. Praveena et al. and Rajeswaran et al. showed that this behavior can be due to the existence of leakage current [36,37]. According to Table 5, both of Pr and Ec became more from R= Dy to Lu. The spontaneous weak ferroelectric origin of RFeO3 materials with magnetic R3+ions (here R=Dy, Ho & Yb) can be due to the exchange striction mechanism and the poling electric field which can displace the magnetic R-ions from their symmetric position [38,39]. Exchange field on R-ion site has stabilized them because of the ferromagnetism of Fe-sublattice [39]. Wang et al. found that the electric polarization of heavy rare earth orthoferrites can depend on the ionic radius of rare earth ions [40]. In addition, in RFeO3 compounds the lattice centerosymetmric will become less with increasing distortions so the net polarization will be grown [40]. Based on P-E curves of the prepared samples, LuFeO3 NPs showed the maximum amounts of Pr and Ec. In addition, LuFeO3 showed the maximum polarization about 0.257 μC/cm2 at the electric field of 49 kV/cm. This behavior had been seen in other previous researches [41]. In orthoferrites with nonmagnetic R ion (here Lu3+), the weak ferroelectric polarization can have different origin; one reason would be the high resistivity of this compound [42]. 3.4.1. Frequency dependent of dielectric behavior Frequency response of dielectric constant (ε) and dielectric loss (tanδ) of RFeO3 (R= Dy, Ho, Yb & Lu) compounds were shown in Fig. 8. These measurements were done at room temperature in 500 Hz to 10 MHz range of frequency. According to Fig. 8, dielectric constant (ɛ) and dielectric loss (tanδ) decreased obviously with increasing the frequency for all samples. It illustrates that the dielectric behavior of these compounds can be explained by dipole relaxation process and Maxwell-Wagner model [17,32]. This behavior is a common feature in the ferrite compounds family [12,43]. The reason of this behavior can be related to the distortion of FeO6 octahedron [44]. Beside, not only the dielectric constant became less in RFeO3 NPs from R= Dy to Yb but also HoFeO3 and YbFeO3 have almost shown 7 the same dielectric values. In addition, LuFeO3 showed the maximum amount of ε (about 127). The reason may be due to very small particle size of this compound because in particles with small size, an increase in the number of conduction electrons can be seen at the surface of particles. Therefore, the dielectric constant will increase due to the interactions among these conduction electrons [45-47]. 3.4.2. Temperature dependence of dielectric constant Temperature dependence of dielectric constant (ε) curves of RFeO3 (R= Dy, Ho, Yb & Lu) compounds were shown in Fig. 9. The insets show the variations of dielectric loss (tanδ) versus temperature in these samples. Both of dielectric constant (ε) and dielectric loss (tanδ) increase with increasing temperature up to observed peaks. This increase of dielectric constant at high temperature is due to the existence of additional thermal energy. This energy comes from increasing the mobility of charge carriers in high temperatures but in low temperatures there is not any charge carriers due to the thermal energy. This process provides higher polarization at high temperatures, therefore the dielectric constant will increase [43]. On the other hand, the presence of anomaly in ε and tanδ in this temperature range can related to magnetic phase transition from ordered antiferromagnetic state to disordered paramagnetic state [48, 49]. Therefore, the observed peaks can be considered the Neel temperature (TN) of these compounds. In addition, reduction in dielectric constant after magnetic transition at TN can be due to scatter of charge carriers which leads to depletion in space charge [50]. TN of RFeO3 (R= Dy, Ho, Yb & Lu) are seen at 638, 631, 625 & 619 K in Fig. 9. In addition the amounts of TN can calculate based on the following equations [46]: T/ = 01(13-)5 6% (8) (8) In this equation S is Fe3+ spins and J is: J = 789&(: ' ⁄ ) <' (9) And (9) And w = 180 −< Fe −O −Fe >=> (10) d =< Fe −O >=> (11) w = 180 −< Fe −O −Fe >=> d =< Fe −O >=> w = 180 −< Fe −O −Fe >=> (11) d =< Fe −O >=> That <Fe-O-Fe>av and <Fe-O>av are the average of Fe-O-Fe bond angles and Fe-O bond lengths, respectively. By using the experimental data of <Fe-O-Fe>av and <Fe-O>av from Table.1 and by inserting them in formulas 8-11, the 8 8 decrease of TN for RFeO3 compounds from R=Dy to Lu can be confirmed. This behavior had been seen before in other researches [33,51,52]. decrease of TN for RFeO3 compounds from R=Dy to Lu can be confirmed. This behavior had been seen before in other researches [33,51,52]. 4. Conclusions In this research, RFeO3 (R= Dy, Ho, Yb & Lu) nanoparticles had been synthesized by sol-gel combustion method. According to XRD analysis all samples were single phase with Pbnm space group. The nano-crystallite sizes of RFeO3 were calculated using the Scherrer and Williamson-Hall methods and these values were compared with the average particle sizes which were determined from FE-SEM images. All prepared samples showed weak ferromagnetic behavior. LuFeO3 showed the highest amount of Hc because of more lattice distortion. P-E curves of the samples had shown weak ferroelectric behavior. Neel temperature of these compounds had been determined at 638, 631, 625 and 619 K using temperature dependence of dielectric constant curves. References 34, 189 (2010). 23. M. Farbod, V. K. dehbidi, M. Z, Shoushtari, Ceram. Int. 43, 13670 (2017). 24. X. Yuan, Y. Tang, Y. Sun, M. Xu, J. Appl. Phys. 111, 053911 (2012). 24. X. Yuan, Y. Tang, Y. Sun, M. Xu, J. Appl. Phys. 111, 053911 (2012). 25. Z. Zhou, L. Guo, H. Yang, Q. Liu, F. Ye, J. Alloys. Compd. 583, 21 (2014). 25. Z. Zhou, L. Guo, H. Yang, Q. Liu, F. Ye, J. Alloys. Compd. 583, 21 (2014). 26. W. Zhu, L. Pi, S. Tan, Y. Zhang, Appl. Phys. Lett. 100, 052407 (2012). 26. W. Zhu, L. Pi, S. Tan, Y. Zhang, Appl. Phys. Lett. 100, 052407 (2012). 27. S. Yuan, Y. Wang, M. Shao, F. Chang, B. Kang, Y. Isikawa, S. Cao, J. Appl. Phys. 109, 07E141 (2011). 28. S. Cao, H. Zhao, B. Kang, J. Zhang. W. Ren, Scientific Reports, 4, 1 (2014). 28. S. Cao, H. Zhao, B. Kang, J. Zhang. W. Ren, Scientific Reports, 4, 1 (2014). 29. T. Yamaguchi, J. Phys. Chem. Solids. 35, 479 (1974). 30. H. Shen, J. Xu, M. Jin, G. Jiang, Ceram. Int. 38, 1473 (2012). 31. H. Shen, J. Y. Xu, A. H. Wu, J. Rare Earths. 28, 416 (2010). 31. H. Shen, J. Y. Xu, A. H. Wu, J. Rare Earths. 28, 416 (2010). 32. H. Anwar, Y. Jamil, G. Mustafa, M. U. Islam, A. Shakoor, M. I. Arshad, H. Akhtar, M. R. S. Aleem, J. von. Research. 13, 241 (2017). 33. Z. Wang, Y. Lan, Z. Zeng, X. Chen, Q. Chen, Sol. State. Com. 288, 10 (2019). 33. Z. Wang, Y. Lan, Z. Zeng, X. Chen, Q. Chen, Sol. State. Com. 288, 10 (2019). 34. Y. Tokura, S. Seki, N. Nagaosa, Rep. Prog. Phys. 77, 076501 (2014). 34. Y. Tokura, S. Seki, N. Nagaosa, Rep. Prog. Phys. 77, 076501 (2014). 35. F. Hong, Zh. Cheng, X. Wang, J. Appl. Phys. 112, 013920 (2012). 35. F. Hong, Zh. Cheng, X. Wang, J. Appl. Phys. 112, 013920 (2012). 36. K. Praveena, P. Bharathi, H. L. Liu, K. B. R. Varma, Ceram. Int. 42, 13572 (2016). 37. B. Rajeswaran, D. I. Khomskii, A. K. Zvezdin, C. N. R. Rao, Phys. Rev. B 86, 214409 (2012). 38. H. J. Zhao, Y. Yang, W. Ren, A. J. Mao, X. M. Chen, L. Bellaiche, J. Phys.: Condens. Matter. 26 39. B. Rajeswaran, D. Sanyal, M. Chakrabarti, Y. Sundarayya, A. Sundaresan, C. References 1. J. F. Scott, Multiferroic memories, Nat. Mater. 6, 256 (2007). 2 Y. B. Bazailiy, L. T. Tsymbal, G. N. Nakazei, P. E. Wigen, J. Appl. Phys. 95, 6622 (2004). 3. B. Ita, P. Murugavel, V. Ponnambalam, A. R. Raju, Proc. Indian Acad. Sci. 115, 519 (2003). 4 S. Shravan Kumar Reddy, N. Raju, Ch. Gopal Reddy, P. Yadagiri Reddy, K. Rama Reddy, V. Raghavendra Reddy, J. Magn. Magn. Mater. 396, 214 (2015). y j y g y y g y J. Magn. Magn. Mater. 396, 214 (2015). 5. G. Catalan, J. F. Scott, Adv. Mater. 21, 2463 (2009). 5. G. Catalan, J. F. Scott, Adv. Mater. 21, 2463 (2009). 6. H. Katsura, N. Nagaosa, A. V. Balatsky, Phys. Rev. Lett. 95, 057205 (2005). 7. A. Fossdal, M. A. Einarsrud, T. J. Grande, Europ. Ceram. Soc. 25, 927 (2005). 8. E. Bousquet, N. A. Spaldin, Phys. Rev. Letters. 107, 197603 (2011). 9. S. Mathur, H. Shen, N. Lecerf, A. Kjekshus, H. Fjellvag, G. F. Goya, Adv. Mater. 14, 1405 (2002). 10. M. Rajendran, A. K. Bhattacharya, J. Eur. Ceram. Soc. 26, 3675 (2006). 10. M. Rajendran, A. K. Bhattacharya, J. Eur. Ceram. Soc. 26, 3675 (2006). 11. X. Qi, J. Zhou, Z. Yue, Z. Gui, L. Li, Ceram. Int. 29, 347 (2003). 12. M. Nakhaei, D. Sanavi Khoshnoud, Phys. B Condens. Matter. 553, 53 (2019). 13. C. Kittel, Introduction to Solid State Physics, 8nd edn (University of California, Berkeley, 2005) pp. 25-27. 14. Z. Q. Wang, Y. Mu, Z. Y. Zeng, X. R. Chen, Q. F. Chen, Mater. Res. Express. 6, 055605 (2019) 9 15. R. D. Shannon, Acta. Cryst. 32, 751 (1976). 16. M. Rajendran, A.K. Bhattacharya, J. Euro. Ceram. Soc. 24, 111 (2004). 17. M. Rahimkhani, D. Sanavi Khoshnoud, M. H. Ehsani, J. Magn. Magn. Mat. 449, 538 (2018). 18. L. X. Chun, H. Rongzi, T. Chang sheng, J. Mater. Sci: Materials in Electronics. 20, 323 (2008). 19. M. R. A. Bhuiyan, M. M. Alam, M. A. Momin, M. J. Uddin, M. Islam, Int. J. Mater. Mech. Eng. 1, 21 (2012). 20. A. R. Stokes, A. J. C. Wilson, Proc. Phys. Soc. 56, 174 (1944). 20. A. R. Stokes, A. J. C. Wilson, Proc. Phys. Soc. 56, 174 (1944). 21. V. D. Mote, Y. Purushotham, B. N. Dole, J. Theoric. Appl. Phys. 6, 7235 (2012). 22. M. A. Bhuiyan, S. M. Hoque, S. Choudhaury, J. Bangladesh, Aca. Sci. 40. W. Wang, W. Sun, G. zhang, Z. Cheng, Y. Wang, J. Mater. Chem. C 7, 1 (2019). References N. R. Rao, EPL (Europhysics Letters) 101, 1 (2013). 40. W. Wang, W. Sun, G. zhang, Z. Cheng, Y. Wang, J. Mater. Chem. C 7, 1 (2019). 10 41. U. Chowdhury, S. Goswami, A. Roy, Sh. Rajput, A. K. Mall, R. Gupta, S. D. Kaushik, V. Siruguri, S. Saravanakumar, S. Israel, R. Saravanan, A. Senyshyn, T. Chatterji, J. F. Scott, A. Garg, D. Bhattacharya, Phys. Rev. B 100, 195116 (2019). 42. S. Acharya, J. Mondal, S. Ghosh, S. K. Roy, P. K. Chakrabarti, Mater. Lett. 64, 415 (2010). 43. K. Sultan, R. Samad, S. A. U. Islam, M. Z. Habib, M. Ikram, J. Elec. Mater. 48, 6003 (2019). 44. A. Berenov, E. Angeles, J. Rossiny, E. Raj, J. Kilner, A. Atkinson, Sol. State. Ionic. 179, 1090 (2008). 45. L. Zhu, N. Sakai, T. Yanoh, S. Yano, N. Wada, H. Takeuchi, A. Kurokawa, Y. Ichiyanagi, J. Phys. 352, 012021 (2012). 46. J. S. Zhou, J. A. Alonso, V. Pomjakushin, J. B. Goodenough, Y. Ren, J. Q. Yan, J. G. Cheng, Phys. Rev. B 81, 214115 (2010). 47. H. Ghayour, M. Abdellahi, M. Bahmanpour, A. Khanda, J. Comput. Electron. 4, 1 (2016). 48. F. Yen, B. Lorenz, Y. Y. Sun, C. W. Chu, L. N. Bezmaternykh, A. N. Vasiliev, Physic. Rev. B 73, 054435 (2006). 49. I. Ahmad, M. J. Akhtar, M. Younas, M. Siddique, M. M. Hasan, J. Appl. Phys. 112, 074105 (2012) 074105- 074113. 50. S. Manzoor, S. Husain, A. Somvanshi, M. Fatema, J. Mater. Sci: Materials in Electronics. 31, 7811 (2020). 51. M. C. Weber, M. Guennou, H. J. Zhao, J. Iniguez, R. Vilarinho, A. Almeida, J. A. Moreira, J. Kreisel, Phys. Rev. B 94, 214103 (2016). 52. M. Eibschutz, S. Shtrikman, D. Treves, Phys. Rev. 156, 562 (1967). 52. M. Eibschutz, S. Shtrikman, D. Treves, Phys. Rev. 156, 562 (1967). 11 11 Caption tables and Figures Table 1. Details of Rietveld refinement of XRD analysis in RFeO3 compounds. Table 1. Details of Rietveld refinement of XRD analysis in RFeO3 compounds. Table 2. Crystallite size according to Scherrer equation (DScher) and Williamson-Hall method (DW-H), lattice strain (g & ɛ), mean particle diameter (˂D>SEM), standard deviation (σD) and crystallinity index (Icry) of RFeO3 compounds. Table 3. Atomic percent ratio (Fe/R) and atomic weight percent (%) of RFeO3 compounds according to EDS spectrums. Table 2. Crystallite size according to Scherrer equation (DScher) and Williamson-Hall method (DW-H), lattice strain (g & ɛ), mean particle diameter (˂D>SEM), standard deviation (σD) and crystallinity index (Icry) of RFeO3 compounds. Table 2. Crystallite size according to Scherrer equation (DScher) and Williamson-Hall method (DW-H), lattice strain (g & ɛ), mean particle diameter (˂D>SEM), standard deviation (σD) and crystallinity index (Icry) of RFeO3 compounds. ), p ( SEM), ( D) y y ( cry) 3 p Table 3. Atomic percent ratio (Fe/R) and atomic weight percent (%) of RFeO3 compounds according to EDS spectrums. Table 3. Atomic percent ratio (Fe/R) and atomic weight percent (%) of RFeO3 compounds according to EDS spectrums. Table 4. Mr, Hc and Mnet of RFeO3 compounds. Fig. 1. a) X-ray diffraction patterns and b) the shifts of XRD peaks of RFeO3 compounds. Fig. 1. a) X-ray diffraction patterns and b) the shifts of XRD peaks of RFeO3 compounds. Fig. 2. a) The Rietveld refinement of XRD patterns for YbFeO3 compounds; Observed (circles), calculated (solid lines) and differences (at the bottom) and vertical marks show the positions of the allowed Bragg reflections; b) Variations of cell parameters (a, b & c) versus R3+ ionic radius of RFeO3 compounds (R= Dy, Ho, Yb, & Lu and c) schematic view of the lattice distortions of RFeO3: pink, purple, brown and blue spheres show O12-, O22-, Fe3+ and R3+ions, respectively. Fig. 3. The Williamson-Hall curves of RFeO3 compounds. Fig. 3. The Williamson-Hall curves of RFeO3 compounds. Fig. 4. The FE-SEM images of RFeO3 compounds. Fig. 4. The FE-SEM images of RFeO3 compounds. Fig. 4. The FE-SEM images of RFeO3 compounds. Fig. 5. EDS spectrums of RFeO3 compounds. Fig. 6. The magnetic hysteresis loops of RFeO3 compounds at room temperature. Fig. 6. The magnetic hysteresis loops of RFeO3 compounds at room temperature. Fig. 8. Frequency response of (a) dielectric constant (ε); (b) dielectric loss (tanδ) in RFeO3 compounds. Fig. 9. Caption tables and Figures Temperature dependence of dielectric constant (ԑ) at 1000 Hz frequency in RFeO3 compounds. The inset of figures shows dielectric loss (tanδ) versus temperature of RFeO3 compounds. 12 Table 1 Sample Position GOF Volume ((Ȧ)3) Bond length (Ȧ) Bond angle (ᵒ) τ d (Ȧ) DyFeO3 Lable X Y Z Dy 0.9837 0.0632 0.2500 O1 0.1013 0.4610 0.2500 O2 0.6984 0.3040 0.0521 1.21 228.080 (Fe-O1)avg 1.9986 (Fe-O2)avg 2.0151 Fe-O1-Fe 145.983 Fe-O2-Fe 147.097 0.88 2.72 HoFeO3 Ho 0.9822 0.0677 0.2500 O1 0.1037 0.4569 0.2500 O2 0.6935 0.3081 0.0601 1.28 224.724 (Fe-O1)avg 2.0021 (Fe-O2)avg 2.0242 Fe-O1-Fe 143.632 Fe-O2-Fe 143.617 0.86 2.70 YbFeO3 Yb 0.9807 0.0705 0.2500 O1 0.1287 0.4460 0.2500 O2 0.6837 0.3030 0.0644 1.29 220.158 (Fe-O1)avg 2.0407 (Fe-O2)avg 2.0222 Fe-O1-Fe 136.023 Fe-O2-Fe 141.470 0.81 2.68 LuFeO3 Lu 0.9799 0.0710 0.2500 O1 0.1268 0.4555 0.2500 O2 0.6917 0.3045 0.0614 1.29 219.128 (Fe-O1)avg 2.1644 (Fe-O2)avg 1.9525 Fe-O1-Fe 135.281 Fe-O2-Fe 140.040 0.80 2.67 13 Table 2 Sample DScher (nm) DW-H (nm) ɛ ˂D>SEM(nm) σD (nm) g Icry DyFeO3 56±0.25 63±0.02 0.0056±0.05 89±0.42 40.22±0.009 0.218±0.06 1.41±0.08 HoFeO3 53±0.13 61±0.06 0.019±0.01 86±0.36 34.16±0.007 0.228±0.04 1.40±0.15 YbFeO3 49±0.22 54±0.02 0.0395±0.03 69±0.21 22.09±0.008 0.235±0.04 1.28±0.09 LuFeO3 46±0.15 49±0.03 0.0412±0.02 60±0.22 20.59±0.003 0.257±0.08 1.23±0.06 14 Table 3 Sample Atomic Percent Ratio (Fe/R) Atomic Weight Percent (%) DyFeO3 0.89 Fe 21.12 Dy 23.24 O 55.64 HoFeO3 0.74 Fe 25.80 Ho 36.32 O 37.88 YbFeO3 0.71 Fe 13.00 Yb 32.74 O 54.26 LuFeO3 0.88 Fe 18.32 Lu 20.58 O 61.10 Table 3 Sample Atomic Percent Ratio (Fe/R) Atomic Weight Percent (%) DyFeO3 0.89 Fe 21.12 Dy 23.24 O 55.64 HoFeO3 0.74 Fe 25.80 Ho 36.32 O 37.88 YbFeO3 0.71 Fe 13.00 Yb 32.74 O 54.26 LuFeO3 0.88 Fe 18.32 Lu 20.58 O 61.10 15 Table 4 Sample Mr (10-3 emu/g) Hc(10-3 Tesla) Mnet(*10-3 emu/g) DyFeO3 192 73 2935 HoFeO3 270 77 3484 YbFeO3 135 381 773 LuFeO3 224 643 316 Table 4 16 Table 5 Sample Pr (μC/cm2) Ec (kV/cm) DyFeO3 0.006 2.262 HoFeO3 0.007 2.709 YbFeO3 0.009 3.374 LuFeO3 0.022 4.350 Table 5 Table 5 Table 5 Table 5 Table 5 Sample Pr (μC/cm2) Ec (kV/cm) DyFeO3 0.006 2.262 HoFeO3 0.007 2.709 YbFeO3 0.009 3.374 LuFeO3 0.022 4.350 17 17 Fig. 1. 18 Fig. 2. Fig. 2. Fig. 2. 19 Fig. 3. 20 Fig. 4. Fig. 4. Fig. 4. 21 Fig. 5. 22 Fig. 5. Fig. 5. 22 Fig. 6. Fig. 6. 23 Fig. 7. Fig. 7. Fig. 7. 24 Fig. 8. Fig. 8. 25 25 Fig. 9. 26 Figures Figures Figure 1 a) X-ray diffraction patterns and b) the shifts of XRD peaks of RFeO3 compounds. Figure 1 Figure 1 a) X-ray diffraction patterns and b) the shifts of XRD peaks of RFeO3 compounds. igure 2 ) The Rietveld re¦nement of XRD patterns for YbFeO3 compounds; Observed (circles), calculated (solid nes) and differences (at the bottom) and vertical marks show the positions of the allowed Bragg e§ections; b) Variations of cell parameters (a, b & c) versus R3+ ionic radius of RFeO3 compounds (R= Dy, Ho, Yb, & Lu and c) schematic view of the lattice distortions of RFeO3: pink, purple, brown and blue pheres show O12 O22 Fe3+ and R3+ions respectively Figure 2 a) The Rietveld re¦nement of XRD patterns for YbFeO3 compounds; Observed (circles), calculated (solid lines) and differences (at the bottom) and vertical marks show the positions of the allowed Bragg re§ections; b) Variations of cell parameters (a, b & c) versus R3+ ionic radius of RFeO3 compounds (R= Dy, Ho, Yb, & Lu and c) schematic view of the lattice distortions of RFeO3: pink, purple, brown and blue spheres show O12-, O22-, Fe3+ and R3+ions, respectively. Figure 2 Figure 2 a) The Rietveld re¦nement of XRD patterns for YbFeO3 compounds; Observed (circles), calculated (solid lines) and differences (at the bottom) and vertical marks show the positions of the allowed Bragg re§ections; b) Variations of cell parameters (a, b & c) versus R3+ ionic radius of RFeO3 compounds (R= Dy, Ho, Yb, & Lu and c) schematic view of the lattice distortions of RFeO3: pink, purple, brown and blue spheres show O12-, O22-, Fe3+ and R3+ions, respectively. a) The Rietveld re¦nement of XRD patterns for YbFeO3 compounds; Observed (circles), calculated (solid lines) and differences (at the bottom) and vertical marks show the positions of the allowed Bragg re§ections; b) Variations of cell parameters (a, b & c) versus R3+ ionic radius of RFeO3 compounds (R= Dy, Ho, Yb, & Lu and c) schematic view of the lattice distortions of RFeO3: pink, purple, brown and blue spheres show O12-, O22-, Fe3+ and R3+ions, respectively. ure 3 e Williamson-Hall curves of RFeO3 compounds. The Williamson-Hall curves of RFeO3 compounds. Figure 4 Figure 4 The FE-SEM images of RFeO3 compounds. The FE-SEM images of RFeO3 compounds. Figure 5 EDS spectrums of RFeO3 compounds. Figure 5 EDS spectrums of RFeO3 compounds. Figure 6 The magnetic hysteresis loops of RFeO3 compounds at room temperature. Figure 6 The magnetic hysteresis loops of RFeO3 compounds at room temperature. The magnetic hysteresis loops of RFeO3 compounds at room temperature Figure 7 Figure 7 P-E curves of RFeO3 compounds P-E curves of RFeO3 compounds P-E curves of RFeO3 compounds Figure 8 Frequency response of (a) dielectric constant (ε); (b) dielectric loss (tanδ) in RFeO3 compounds. Figure 8 Frequency response of (a) dielectric constant (ε); (b) dielectric loss Figure 8 Frequency response of (a) dielectric constant (ε); (b) dielectric loss (tanδ) in RFeO3 compounds. Figure 8 Frequency response of (a) dielectric constant (ε); (b) dielectric loss (tanδ) in RFeO3 compounds. Figure 9 Temperature dependence of dielectric constant (ԑ) at 1000 Hz frequency in RFeO3 compounds. The inset of ¦gures shows dielectric loss (tanδ) versus temperature of RFeO3 compounds. Figure 9 Figure 9 Temperature dependence of dielectric constant (ԑ) at 1000 Hz frequency in RFeO3 compounds. The inset of ¦gures shows dielectric loss (tanδ) versus temperature of RFeO3 compounds.
https://openalex.org/W2995293477	https://www.kerwa.ucr.ac.cr/bitstream/10669/87322/1/2019%20ELANS%20Socio-demographic%20patterning%20selfreported%20PA.pdf	English	null	Original research Socio-demographic patterning of self-reported physical activity and sitting time in Latin American countries: findings from ELANS	BMC public health	2,019	cc-by	10,540	Ferrari et al. BMC Public Health (2019) 19:1723 https://doi.org/10.1186/s12889-019-8048-7 Ferrari et al. BMC Public Health (2019) 19:1723 https://doi.org/10.1186/s12889-019-8048-7 Open Access Original research Socio-demographic patterning of self-reported physical activity and sitting time in Latin American countries: findings from ELANS Gerson Luis de Moraes Ferrari1,2* , Irina Kovalskys3, Mauro Fisberg2,4, Georgina Gómez5, Attilio Rigotti6, Lilia Yadira Cortés Sanabria7, Martha Cecilia Yépez García8, Rossina Gabriella Pareja Torres9, Marianella Herrera-Cuenca10, Ioná Zalcman Zimberg11, Viviana Guajardo3, Michael Pratt12, Abby C. King13, Dirceu Solé2 and on behalf of the ELANS Study Group * Correspondence: gersonferrari08@yahoo.com.br * Correspondence: gersonferrari08@yahoo.com.br 1Centro de Investigación en Fisiologia del Ejercicio – CIFE, Universidad Mayor, José Toribio Medina, 29. Estacion Central, Santiago, Chile 2Departamento de Pediatria da Universidade Federal de São Paulo, São Paulo, Brazil Full list of author information is available at the end of the article Abstract 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. * Correspondence: gersonferrari08@yahoo.com.br 1Centro de Investigación en Fisiologia del Ejercicio – CIFE, Universidad Mayor, José Toribio Medina, 29. Estacion Central, Santiago, Chile 2Departamento de Pediatria da Universidade Federal de São Paulo, São Paulo, Brazil Full list of author information is available at the end of the article Abstract Background: Low levels of physical activity (PA) and prolonged sitting time (ST) increase the risk of non- communicable diseases and mortality, and can be influenced by socio-demographic characteristics. The aim of this study was to use self-report data to characterise socio-demographic patterns of PA and ST in eight Latin American countries. Methods: Data were obtained from the Latin American Study of Nutrition and Health (ELANS), a household population-based, multi-national, cross-sectional survey (n = 9218, aged 15–65 years), collected from September 2014 to February 2015. Transport and leisure PA and ST were assessed using the International Physical Activity Questionnaire–long version. Overall and country-specific mean and median levels of time spent in transport and leisure PA and ST were compared by sex, age, socioeconomic and education level. Results: Mean levels of transport and leisure PA were 220.3 min/week (ranging from 177.6 min/week in Venezuela to 275.3 min/week in Costa Rica) and 316.4 min/week (ranging from 272.1 min/week in Peru to 401.4 min/week in Ecuador). Transport and leisure PA were higher (p < 0.005) in men than women with mean differences of 58.0 and 34.0 min/week. The mean and median for transport PA were similar across age groups (15–29 years: mean 215.5 and median 120 min/week; 30–59 years: mean 225.0 and median 120 min/week; ≥60 years: mean 212.0 and median 120 min/week). The median time spent in transport and leisure PA between three strata of socioeconomic and education levels were similar. The prevalence of not meeting PA recommendations were 69.9% (95% CI: 68.9–70.8) for transport and 72.8% (95% CI: 72.0–73.7) for leisure. Men, younger people (15–29 years), individuals with higher socioeconomic and education levels spent significantly (p < 0.001) more time sitting than women, older people (30–59 years and ≥60 years) and those in the middle and low socioeconomic and education groups, respectively. Conclusions: Transport and leisure PA and ST range widely by country, sex, and age group in Latin America. Programs for promoting leisure and transport PA and reducing ST in Latin America should consider these differences by age and gender and between countries. (Continued on next page) © The Author(s). Methods Latin American Study of Nutrition and Health and sample The Latin American Study of Nutrition and Health (Estudio Latinoamericano de Nutrición y Salud; ELANS) is a cross-sectional, multinational representative sample conducted in 8 of the 33 Latin America countries (Argentina, Brazil, Chile, Colombia, Costa Rica, Ecuador, Peru and Venezuela). Only urban populations were in- cluded to enhance comparability across countries and for reasons of feasibility [20]. The present study used large-scale urban population samples, and these preva- lence estimates may reasonably be generalized to the country level given the high degree of urbanization. Data were collected from September 2014 to February 2015. The overarching ELANS protocol was approved by the Western Institutional Review Board (#20140605) and is registered at ClinicalTrials.gov (#NCT02226627). Add- itional site-specific protocols were also approved by the ethical review boards of the participating institutions. All of the participants provided informed consent/assent for participation in their country-level study. The eight par- ticipating countries followed a common protocol, in- cluding training for all research professionals. A balance number of participants were stratified by sex, age group and socioeconomic level. In total, 9218 (4409 [47.8%] men) participants aged 15–65 years were included in the study. Sample size and exclusion criteria can be found elsewhere [21]. On the basis of this evidence the World Health Organization (WHO) has developed global recommen- dations for physical activity and an action plan for in- creasing physical activity and decreasing time spent in sedentary behavior [4]. Policy development and evalu- ation in this area depend on consistent and valid assess- ment of prevalence and trends in physical activity, adherence to physical activity recommendations, and time spent sitting. Continued improvements in monitor- ing physical activity and sitting time are needed to guide policy making and programs for increasing physical ac- tivity and reducing sitting time [4]. In recent decades, Latin America has experienced accel- erated demographic and epidemiological transitions, and many countries are facing the double burden of commu- nicable and non-communicable diseases [9]. Introduction difficult [16]. Latin America is the most urbanized region in the world, with nearly 80 % of people living in cities [17]. At the same time, the urban environment in Latin America differs considerably from those in the high- income countries [18] and has the largest percentage of the population living in slums [17], and high rates crime and violence [19]. Physical activity and sitting patterns vary by sociodemographic characteristics such as country, sex, age, level of income, and education [15] and these fac- tors must be taken into account as public health programs are developed. The aim of this study was to use self-report data to characterise socio-demographic patterns of phys- ical activity and sitting time in eight Latin American coun- tries in order to better inform public health policy and programs in the region. The incidence of cardiovascular diseases (CVD) is in- creasing throughout the developing world; causing more than 16 million deaths each year, 80% of which occur in low and middle-income countries [1]. Regular physical activity reduces the risk of cardiovascular mortality [2]. Physical inactivity also [3] accounts for 1–3% of health care costs, excluding costs associated with mental health and musculoskeletal conditions [4] and contributes 6% of the mortality burden of coronary heart disease and 10% of breast and colon cancer [5]. Sedentary behavior, as distinct from physical activity, encompasses a broad range of behaviors that involve a sit- ting or reclining posture and do not increase energy ex- penditure above 1.5 metabolic equivalents during waking time [6]. Sedentary behaviours are associated with CVD, cancer and all-cause mortality, independent of physical ac- tivity [7]. Current physical activity guidelines do not pre- scribe a quantitative guideline for sitting time [4, 8]. Methods In Latin America physical inactivity levels are high, national health care expenditures due to inactivity are more than USD 3 billion [10], and inactivity has been identified as a critical public health challenge [9, 11], Evidence from a compre- hensive review suggests that populations with higher levels of transport physical activity have higher overall levels of physical activity than those populations who rely more on private transportation [12, 13], and that individuals who engage in transport physical activity have lower risk of CVD and all-cause mortality [13, 14]. However, data on physical activity by mode (e.g., active transportation, leis- ure activity, sitting time) and country in Latin America re- main scarce [15]. As such, international comparisons are © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Page 2 of 12 Ferrari et al. BMC Public Health (2019) 19:1723 (Continued from previous page) Trial registration: ClinicalTrials.Gov NCT02226627. Retrospectively registered on August 27, 2014. Keywords: Epidemiology, Physical activity, Sitting time, Self-report, Public health Sociodemographic characteristics Participants reported age by year (15–65 years), and three age groups (15–29, 30–59, and ≥60 years) were de- fined to ensure adequate sample sizes. Sex, socioeco- nomic and educational level were categorized using standard questionnaires. Socioeconomic level was evalu- ated by questionnaire using country-specific definitions based on national norms, laws, and the questionnaires used on national surveys in each country [26–31]. Given the variability in categorizing socioeconomic strata, a standard three level system (low, medium, high) was de- veloped [21]. A similar process was used to standardize level of education in three strata (basic or lower [low], high school [medium], and university degree [high]) in the eight countries. y The levels of transport physical activity were higher for men (mean: 251.2 min/week; 95% CI: 238.9–262.2; median: 125.0 min/week) than for women (mean: 193.2 min/week; 95% CI: 185.2–210.0; median: 105.0 min/ week) overall, with a mean difference of 58 min/week. The largest sex difference was in Colombia (88.9 mean min/week), followed by Chile (88.6 mean min/week) and the smallest sex difference was in Venezuela (2.4 mean min/week). For leisure physical activity, the largest sex difference was in Venezuela (148.9 mean min/week) and the smallest was in Argentina (5 mean min/week). Men (mean: 479.1 min/day; 95% CI: 470.4–488.0; median: 420.0 min/day) spent more time sitting than women Physical activity and sitting time assessment Physical activity and sitting time assessment Physical activity and sitting time were assessed using a Spanish language long-form “last 7 days” self-administered Ferrari et al. BMC Public Health (2019) 19:1723 Page 3 of 12 evaluate the data distribution. Differences between groups were analyzed using Wilcoxon or Kruskal-Wallis tests. version of the International Physical Activity Questionnaire (IPAQ) [22]. The IPAQ contains questions about the amount of walking, moderate physical activity, and vigorous physical activity occurring as part of active transport and in leisure-time [22]. The transport and leisure-time physical activity sections were included, due to the greater relevance of these domains for guiding public health policies and pro- grams [23], and the poor validity of the IPAQ occupational and home-based physical activity questions in Latin Ameri- can urban settings. The main outcomes were the mean and median time, in min/week, spent in the two modes of physical activity (transport: walking + bicycle; leisure: walking + moder- ate + vigorous) and sitting time (min/day). Results were stratified by sex, age group, socioeconomic level, and educational level. We also reported the proportion of each group meeting the WHO physical activity guide- lines (e.g. > 60 min/day for adolescents and > 150 min/ week of moderate-to-vigorous physical activity for adults) in transport and leisure. Data analyses were per- formed with IBM SPSS, version 22 for Windows [32]. The samples were weighted to adjust for sociodemo- graphic characteristics, sex, and income [21]. Data were analyzed in accordance with the IPAQ scor- ing protocol (www.ipaq.ki.se). IPAQ assesses walking separately. Thus, IPAQ physical activity data are re- ported as min/day of walking, moderate, and vigorous physical activity. Total time (min/week) and time spent in each of the physical activity modes (i.e., transport and leisure-time) were estimated and used as analysis vari- ables. We analyzed transport physical activity (walking + bicycle) and leisure physical activity (walking + moderate + vigorous) separately. In addition, the IPAQ contained two items capturing sitting time. Sitting time was assessed from the questions in the IPAQ long-form [22, 24]. Participants were asked to report time spent sitting over the past 7 days, with separate amounts reported for weekdays and weekends. We calculated average sitting time per day (min/day) as follows: (weekday time5 + weekend time2)/7 [25]. Results The proportion of women (52.2%; 95% CI: 51.2–53.2) was higher than men (47.8%; 95% CI: 46.8–48.8). In terms of age, 39.4% (95% CI: 38.5–40.5) of participants were aged < 30 years, 53.9% (95% CI: 52.9–55.0) aged 30–59 years, and 6.7% (95% CI: 6.1–7.2) aged ≥60 years. About half were classified as having a low socioeco- nomic level (52.0%; 95% CI: 51.0–53.0) and/or low edu- cational level (61.2%; 95% CI: 60.3–62.3) (Table 1). Overall, the response rate for IPAQ was 99.4%. For transport physical activity, Venezuela had the lowest values (mean: 177.6 min/week; 95% CI: 160.7–194.6; me- dian: 100.0 min/week), and the highest average was in Costa Rica (mean: 275.3 min/week; 95% CI: 249.6–301.8; median: 147.0 min/week). The difference between these two countries was 97.7 min/week. For leisure physical ac- tivity, the highest values were in Ecuador (mean: 401.4 min/week; 95% CI: 370.6–435.3; median: 240.0 min/week) and the lowest was in Peru (mean: 272.1 min/week; 95% CI: 248.1–297.3; median: 150.0 min/week), with a mean difference of 129.3 min/week between these two countries. For sitting time, the mean difference between Argentina (highest sitting time) and Ecuador (lowest sitting time) was 196.3 min/day (Table 2). Statistical analysis Overall, the mean and median transport physical activ- ity were similar across age groups (15–29 years: mean 215.5 min/week, 95% CI: 205.0–226.7 and median 120 min/week; 30–59 years: mean 225.0 min/week, 95% CI: 215.7–234.3 and median 120 min/week; ≥60 years: 212.0 min/week, 95% CI: 187.1–238.0 and median 120 min/ week). Significant difference (p = 0.027) between age group for transport physical activity was found only in the Ecuador (Table 3). Overall, the prevalence of insufficient physical activity in the transport and leisure domains were 69.9% (95% CI: 68.9–70.8) and 72.8 (95% CI: 72.0–73.7); ranging from 59.8% (Chile) to 81.0% (Venezuela) for transport and 46.1% (Ecuador) to 83.8% (Venezuela) for leisure (Additional file 1: Figure S1). In each country and over- all, women (76.9; 95% CI: 75.8–78.1) were more likely to be insufficient physical activity than men (68.4%; 95% CI: 67.0–69.8) for leisure (Additional file 2: Figure S2). The time spent in leisure physical activity among the 15–29 years old group was significantly higher in Brazil (p = 0.011), Venezuela (p = 0.001) and in overall (p < 0.001) than for those 30–59 and ≥60 years old, and the higher difference between 15 and 29 and ≥60 years was in Venezuela (217.7 mean min/week). For sitting time, the time spent was significantly higher in 15–29 years old than for those 30–59 and ≥60 years in overall and in five countries (Brazil, Chile, Colombia, Costa Rica, and Peru (Table 3). Overall, insufficient transport physical activity preva- lence was lower among those aged 30–59 years (68.7%; 95% CI: 67.4–73.0) compared to those aged <30 years (71.5%; 95% CI: 69.9–73.0) and those aged ≥60 years (69.1%; 95% CI: 65.5–72.5). Insufficient leisure physical activity prevalence was lower among those aged < 30 years compared to those aged 30–59 and ≥60 years (Additional file 3: Figure S3). In each country and overall, the median times spent in transport physical activity were similar between the three socioeconomic strata (p > 0.05). Statistical analysis Descriptive analyzes are presented as arithmetic mean, median, frequency, percentage, and 95% confidence in- tervals (95% CI) for physical activity (transport and leis- ure time) and sitting time for each country and for the entire sample (sum of the eight countries). Since mi- nutes of physical activity (transport and leisure time) and sitting time were not normally distributed, values for the 25th, 50th and 75th percentile were also ob- tained. A Kolmogorov-Smirnov test was applied to Ferrari et al. BMC Public Health (2019) 19:1723 Page 4 of 12 Table 1 Sample distribution (%) according to sex, age group, socioeconomic level, and educational level from ELANS study Country N Sex Age group (years) Socioeconomic level a Educational level Men Women 15–29 30–59 ≥60 Low Medium High Low Medium High Argentina 1266 45.3 54.7 35.9 56.4 7.7 48.7 46.2 5.1 75.4 20.3 4.3 Brazil 2000 47.1 52.9 35.6 57.2 7.2 45.8 45.8 8.4 48.4 43.2 8.4 Chile 879 48.4 51.6 38.3 54.9 6.8 46.8 44.1 9.1 65.1 23.7 11.2 Colombia 1230 49.0 51.0 39.3 52.7 8.0 63.3 31.2 5.4 65.0 23.9 11.1 Costa Rica 798 49.4 50.6 41.4 53.6 5.0 32.8 53.6 13.6 81.6 12.6 5.8 Ecuador 800 49.6 50.4 43.0 51.1 5.9 49.9 37.1 13.0 83.0 10.5 6.5 Peru 1113 47.0 53.0 44.2 50.5 5.3 47.9 31.9 20.2 23.1 67.1 9.8 Venezuela 1132 48.8 51.2 42.4 51.7 5.9 77.7 16.8 5.5 68.6 12.6 18.8 Total 9218 47.8 52.2 39.4 53.9 6.7 52.0 38.4 9.6 61.2 29.3 9.5 aEstimate distribution of sample (n) according to socioeconomic level association between socioeconomic level and transport or leisure physical activity. There are no significant dif- ferences between socioeconomic levels for transport and leisure physical activity. Overall, individuals with higher socioeconomic (p < 0.001) and education levels (p < 0.001) spent more time sitting than those in the middle and low socioeconomic and education groups (Tables 4 and 5). (442.7 min/day; 95% CI: 434.6–450.8; median: 418.0 min/ day). The mean difference between sex was 36.4 mean min/day. In overall, men had a significantly higher values than women for transport (p < 0.001), leisure (p = 0.011) physical activity and sitting time (p < 0.001). Statistical analysis In leisure physical activity, we found significant differences in Brazil (low: mean 293.0 min/week, 95% CI: 252.4–336.3 and median 140.0 min/week; medium: mean 332.8 min/week, 95% CI: 295.4–371.9 and median 230 min/week; high: mean 307.6 min/week, 95% CI: 239.1–381.9 and median 180 min/week) and in Peru (low: mean 344.7 min/week, 95% CI: 299.5–391.2 and median 139.0 min/week; medium: mean 268.7 min/week, 95% CI: 226.6–312.3 and median 180.0 min/week; high: mean 281.6 min/week, 95% CI: 234.5–333.1 and median 120 min/week) between strata socioeconomic level. The results do not show a clear Participants with low socioeconomic level and low education level had a slightly higher percentage of in- sufficient physical activity. However, these trends were not observed in all countries for transport physical activity. Persons of high socioeconomic level (65.8%; 95% CI: 62.6–68.6) had a lower prevalence of insuffi- cient leisure physical activity overall compared with those of middle socioeconomic level (75.2%; 95% CI: 73.9–76.4) or low SES (71.4%; 95% CI: 69.9–73.0) (Additional file 4: Figure S4). Individuals with low (74.1%; 95% CI: 73.0–75.3) or middle (72.0%; 95% CI: 70.2–73.6) education levels had a higher prevalence of insufficient leisure physical activity compared with Ferrari et al. Statistical analysis those with a high education level (67.4%; 95% CI: 64.2–70.6) (Additional file 5: Figure S5). physical activity were similar across age groups, but leis- ure physical activity was higher in the 15–29 group than for those 30–59 and ≥60 years old in Brazil (p = 0.011), Venezuela (p = 0.001) and overall (p < 0.001). Sitting time was highest among those with higher socioeconomic and education levels. In contrast, the relationships be- tween physical activity and socioeconomic and education levels were more variable across countries. Statistical analysis BMC Public Health (2019) 19:1723 Page 5 of 12 Table 2 Characteristics of participants by transport, leisure physical activity and sitting time by sex from ELANS study Country Total Men Women P value Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Transport physical activity (min/week) Argentina 251.9 (231.7–271.7) 140.0 (60.0–300.0) 288.2 (247.6–328.7) 130.0 (60.0–350.0) 227.4 (204.4–205.0) 140.0 (60.0–280.0) 0.702 Brazil 206.2 (192.1–220.8) 105.0 (60.0–210.0) 245.9 (220.5–271.9) 120.0 (60.0–245.0) 170.1 (155.6–185.6) 105.0 (60.0–210.0) 0.008 Chile 216.7 (194.3–238.9) 105.0 (60.0–105.0) 262.4 (222.8–301.8) 137.5 (70.0–280.0) 173.8 (152.3–194.3) 100.0 (50.0–210.0) <0.001 Colombia 231.0 (212.4–251.3) 120.0 (25.0–240.0) 277.4 (243.5–307.9) 140.0 (71.5–295.5) 188.5 (166.5–210.2) 105.0 (50.0–210.0) <0.001 Costa Rica 275.3 (249.6–301.8) 147.0 (63.0–350.0) 308.4 (268.6–351.9) 180.0 (70.0–400.0) 244.8 (217.2–277.4) 140.0 (60.0–283.7) 0.087 Ecuador 221.3 (201.1–242.9) 140.0 (75.0–210.0) 250.5 (221.6–284.7) 140.0 (75.0–300.0) 191.5 (167.6–214.1) 120.0 (70.0–210.0) 0.404 Peru 194.4 (179.8–209.4) 120.0 (60.0–210.0) 204.9 (182.5–228.2) 120.0 (60.0–210.0) 185.3 (167.7–203.8) 120.0 (60.0–210.0) 0.618 Venezuela 177.6 (160.7–194.6) 100.0 (50.0–180.0) 176.3 (148.9–205.3) 90.0 (50.0–175.0) 178.7 (156.5–201.2) 100.0 (50.0–200.0) 0.599 Overall 220.3 (213.5–227.0) 120.0 (60.0–240.0) 251.2 (238.9–262.2) 125.0 (60.0–280.0) 193.2 (185.2–200.5) 105.0 (60.0–210.0) <0.001 Leisure physical activity (min/week) Argentina 298.1 (271.7–327.4) 180.0 (90.0–390.0) 295.7 (259.7–335.4) 180.0 (90.0–390.0) 300.7 (262.3–341.0) 205.0 (90.0–378.7) 0.505 Brazil 312.6 (285.7–339.8) 180.0 (70.0–420.0) 345.9 (308.1–384.9) 180.0 (90.0–486.0) 273.8 (240.9–310.9) 140.0 (60.0–360.0) 0.018 Chile 341.4 (306.7–378.5) 210.0 (90.0–420.0) 360.5 (310.5–416.1) 210.0 (85.0–445.0) 322.3 (280.3–369.1) 210.0 (90.0–408.7) 0.817 Colombia 290.6 (265.2–316.6) 150.0 (60.0–360.0) 312.4 (276.8–348.4) 180.0 (90.0–403.7) 263.5 (226.1–307.7) 130.0 (60.0–300.0) 0.025 Costa Rica 293.2 (259.0–329.6) 180.0 (70.0–360.0) 330.4 (282.5–386.3) 180.0 (90.0–420.0) 249.6 (207.2–290.7) 150.0 (60.0–300.0) 0.025 Ecuador 401.4 (370.6–435.3) 240.0 (131.2–480.0) 358.9 (322.7–399.1) 240.0 (120.0–480.0) 447.6 (394.5–502.5) 280.0 (140.0–577.5) 0.064 Peru 272.1 (248.1–297.3) 150.0 (60.0–355.0) 293.9 (259.9–327.9) 180.0 (80.0–370.0) 246.4 (210.6–284.8) 120.0 (60.0–300.0) 0.013 Venezuela 333.5 (296.9–376.3) 227.5 (105.0–438.5) 393.6 (333.4–448.9) 255.0 (120.0–540.0) 244.7 (202.1–289.3) 180.0 (81.7–300.0) 0.014 Overall 316.4 (306.1–327.1) 180.0 (90.0–420.0) 332.3 (317.5–347.5) 198.0 (90.0–440.0) 298.3 (281.7–314.6) 180.0 (60.0–375.0) 0.011 Sitting time (min/day) Argentina 540.7 (526.0–557.9) 480.0 (327.5–720.0) 548.4 (524.3–572.7) 480.0 (330.0–720.0) 534.4 (514.4–553.1) 480.0 (300.0–720.0) 0.971 Brazil 433.9 (420.3–448.1) 360.0 (210.0–600.0) 462.3 (441.2–484.4) 420.0 (240.0–636.0) 408.8 (391.3–427.6) 360.0 (180.0–598.0) 0.003 Chile 485.8 (466.2–499.1) 420.0 (300.0–660.0) 497.5 (470.4–527.4) 450.0 (300.0–660.0) 474.6 (449.1–500.8) 420.0 (300.0–600.0) 0.387 Colombia 481.9 (466.2–499.1) 420.0 (240.0–660.0) 507.3 (482.7–532.9) 480.0 (270.0–720.0) 457.5 (435.3–482.7) 420.0 (239.0–658.0) 0.079 Costa Rica 452.3 (431.3–474.3) 389.0 (240.0–600.0) 486.9 (454.5–520.8) 420.0 (240.0–660.0) 418.1 (394.0–444.7) 360.0 (210.0–570.0) 0.369 Ecuador 344.3 (328.1–362.1) 300.0 (180.0–480.0) 486.8 (454.6–520.8) 300.0 (180.0–500.0) 321.5 (299.6–344.7) 270.0 (180.0–430.0) 0.076 Peru 525.9 (508.3–544.0) 480.0 (328.0–690.0) 548.4 (522.5–572.8) 510.0 (360.0–720.0) 506.1 (482.7–529.4) 480.0 (300.0–660.0) 0.098 Venezuela 383.3 (367.1–398.5) 360.0 (180.0–540.0) 389.4 (367.2–413.3) 360.0 (180.0–540.0) 377.4 (355.3–396.9) 360.0 (180.0–500.0) 0.079 Overall 460.2 (453.8–466.5) 420.0 (240.0–600.0) 479.1 (470.4–488.0) 420.0 (240.0–660.0) 442.7 (434.6–450.8) 418.0 (240.0–600.0) <0.001 95% CI: 95% confidence interval those with a high education level (67.4%; 95% CI: 64.2–70.6) (Additional file 5: Figure S5). Discussion The aim of this study was to quantify and characterise socio-demographic patterns of physical activity and sit- ting time in eight Latin American countries. On average, participants spent 220.3 min/week (median: 120.0 min/ week) in transport physical activity, 316.4 min/week (me- dian: 180.0 min/week) in leisure physical activity and 460.2 min/day (median: 420.0 min/day) in sitting time. When all countries were analyzed together, transport and leisure physical activity and sitting time were higher in men than women. The mean and median of transport The present study reports population-level prevalence estimates and patterns of physical activity and sitting time in urban samples from eight countries using a com- parable, reliable, and validated survey instrument [22]. Previous similar studies in Latin America have generally assessed physical activity at the sub-national level [33– Ferrari et al. Discussion BMC Public Health (2019) 19:1723 Page 6 of 12 Table 3 Characteristics of participants by transport, leisure physical activity and sitting time by age group from ELANS study Country 15–29 years 30–59 years ≥60 years P value Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Transport physical activity (min/week) Argentina 237.2 (205.4–270.2) 120.0 (60.0–290.0) 254.9 (227.2–286.3) 140.0 (60.0–300.0) 304.9 (217.9–407.9) 132.5 (60.0–330.0) 0.652 Brazil 213.6 (189.2–236.9) 105.0 (60.0–210.0) 207.4 (190.2–227.7) 120.0 (60.0–210.0) 156.6 (117.5–209.3) 90.0 (60.0–210.0) 0.247 Chile 216.1 (186.9–249.7) 120.0 (70.0–240.0) 210.8 (179.5–242.2) 105.0 (60.0–210.0) 269.0 (149.3–407.4) 105.0 (41.2–210.0) 0.609 Colombia 218.0 (190.7–247.2) 105.0 (60.0–234.7) 246.3 (216.6–275.8) 120.0 (60.0–270.0) 196.0 (147.8–249.8) 105.0 (60.0–228.0) 0.796 Costa Rica 288.5 (250.1–327.9) 160.0 (75.0–360.0) 271.9 (238.1–307.2) 140.0 (60.0–345.0) 204.0 (123.4–301.3) 82.5 (45.0–240.0) 0.180 Ecuador 195.4 (168.3–224.3) 120.0 (70.0–210.0) 241.6 (210.2–272.4) 140.0 (75.0–256.2) 238.1 (178.7–312.6) 165.0 (101.2–295.0) 0.027 Peru 184.9 (162.4–209.0) 105.0 (60.0–210.0) 199.1 (179.8–220.8) 120.0 (60.0–210.0) 226.7 (166.1–308.6) 150.0 (70.0–227.5) 0.088 Venezuela 182.2 (155.4–212.3) 100.0 (60.0–180.0) 180.1 (154.8–208.1) 90.0 (50.0–205.0) 124.2 (94.2–158.4) 95.0 (47.2–150.0) 0.533 Overall 215.5 (205.0–226.7) 120.0 (60.0–210.0) 225.0 (215.7–234.3) 120.0 (60.0–240.0) 212.0 (187.1–238.0) 120.0 (60.0–210.0) 0.384 Leisure physical activity (min/week) Argentina 329.2 (288.2–376.2) 240.0 (90.0–450.0) 274.7 (238.3–314.6) 180.0 (80.0–360.0) 262.6 (182.4–371.3) 180.0 (90.0–326.2) 0.041 Brazil 358.8 (314.5–405.2) 232.5 (90.0–480.0) 270.1 (240.2–300.3) 146.0 (60.0–360.0) 363.5 (230.4–518.0) 180.0 (80.0–420.0) 0.011 Chile 382.2 (331.5–438.5) 240.0 (120.0–480.0) 297.0 (252.4–342.3) 180.0 (60.0–384.0) 402.5 (247.2–578.4) 195.0 (60.0–645.0) 0.057 Colombia 313.3 (270.2–360.5) 170.0 (60.0–420.0) 273.6 (240.6–310.6) 150.0 (60.0–322.5) 267.9 (174.6–379.8) 169.5 (60.0–310.0) 0.711 Costa Rica 307.1 (255.1–364.5) 160.0 (60.0–360.0) 280.3 (237.7–322.2) 180.0 (90.0–350.0) 285.2 (137.4–455.2) 160.0 (70.0–270.0) 0.969 Ecuador 380.6 (339.7–428.1) 255.0 (123.7–480.0) 412.2 (362.9–462.8) 240.0 (121.2–507.5) 470.8 (334.1–604.1) 370.0 (150.0–622.5) 0.578 Peru 296.2 (260.9–332.9) 165.0 (70.0–360.0) 252.6 (220.9–288.3) 140.0 (60.0–302.5) 191.6 (127.3–274.8) 135.0 (65.0–240.0) 0.138 Venezuela 395.8 (339.2–458.2) 300.0 (120.0–520.0) 266.1 (213.9–322.4) 180.0 (72.5–357.5) 178.1 (94.5–296.8) 147.0 (60.0–180.0) 0.001 Overall 342.1 (326.0–359.9) 210.0 (90.0–450.0) 292.2 (278.0–305.9) 180.0 (70.0–360.0) 315.4 (269.8–358.1) 180.0 (80.0–390.0) <0.001 Sitting time (min/day) Argentina 544.1 (516.9–569.3) 480.0 (360.0–720.0) 534.9 (514.6–555.1) 480.0 (300.0–720.0) 569.4 (507.8–633.5) 570.0 (307.0–720.0) 0.745 Brazil 493.0 (467.6–519.7) 420.0 (240.0–660.0) 403.8 (386.7–420.7) 360.0 (180.0–540.0) 385.4 (342.5–429.8) 360.0 (185.0–500.0) <0.001 Chile 539.2 (509.5–569.9) 480.0 (360.0–720.0) 452.2 (428.0–479.1) 420.0 (270.0–570.0) 453.3 (390.7–526.6) 420.0 (245.0–600.0) <0.001 Colombia 536.9 (508.3–566.7) 480.0 (300.0–720.0) 442.8 (419.8–468.7) 390.0 (238.0–600.0) 473.3 (415.6–540.2) 392.5 (240.0–615.0) <0.001 Costa Rica 503.7 (468.6–538.9) 465.0 (240.0–690.0) 420.1 (393.2–448.7) 360.0 (210.0–558.7) 365.9 (292.5–440.7) 300.0 (195.0–480.0) <0.001 Ecuador 362.2 (336.9–391.3) 300.0 (180.0–510.0) 333.7 (310.8–360.6) 300.0 (165.0–450.0) 303.0 (240.5–363.0) 240.0 (170.0–450.0) 0.293 Peru 556.8 (530.3–585.2) 540.0 (360.0–720.0) 503.9 (480.9–528.8) 480.0 (300.0–660.0) 477.1 (411.8–541.6) 465.0 (318.7–600.0) 0.008 Venezuela 408.9 (383.9–433.6) 360.0 (240.0–540.0) 366.2 (344.9–388.1) 360.0 (180.0–490.0) 349.2 (295.7–412.7) 300.0 (135.0–480.0) 0.099 Overall 496.8 (486.5–507.3) 475.0 (270.0–660.0) 436.8 (428.7–444.6) 390.0 (240.0–600.0) 432.6 (410.6–455.3) 360.0 (240.0–600.0) <0.001 95% CI: 95% confidence interval Compared with the rest of the world, Latin American countries had high prevalences of insufficient physical activity (i.e. Discussion not meeting WHO guidelines) [41]. Our analyses show that the prevalence of insufficient trans- port and leisure physical activity varies greatly across the eight Latin American countries (Fig. S1-S5); insufficient physical activity was lower in Costa Rica (59.8%) and higher in Venezuela (81.0%) for transport and Ecuador (46.1%) and Venezuela (83.8%) for leisure. Werneck et al. [34] compiled self-reported data from six surveys across South American countries (116.982 participants) showed that the highest levels of leisure physical inactiv- ity (< 150 min/week) were in Peru (91.4%), Ecuador (84.7%), Brazil (79.7%), Chile (79.2%) and Argentina 37] and have not used standard surveys, timelines, and methods in representative national samples [37]. In con- trast, ELANS was conducted simultaneously in the urban populations of the most populous cities of eight countries in Latin America. Despite the many manu- scripts describing the global impacts of physical inactiv- ity [4, 5, 10, 38] and global calls for action to reverse the physical inactivity pandemic, few physical activity inter- ventions have occurred in Latin America. While cross- country interventions may be challenging given varying cultural, geographical, social, and economic milieus in different countries, the current results suggest some similarities that may set the stage for further exploration and intervention development [39, 40]. Ferrari et al. Discussion BMC Public Health (2019) 19:1723 Page 7 of 12 Table 4 Characteristics of participants by transport, leisure physical activity and sitting time by socioeconomic level from ELANS study Country Low Medium High P value Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Transport physical activity (min/week) Argentina 255.3 (224.9–287.6) 140.0 (60.0–300.0) 240.1 (210.5–268.5) 140.0 (60.0–280.0) 324.8 (214.7–442.3) 155.0 (70.0–420.0) 0.779 Brazil 221.4 (199.7–244.2) 120.0 (60.0–240.0) 193.6 (174.1–215.1) 105.0 (60.0–210.0) 180.3 (130.2–238.0) 100.0 (45.0–210.0) 0.149 Chile 206.4 (177.6–235.7) 105.0 (60.0–228.0) 221.7 (187.3–259.2) 120.0 (60.0–210.0) 246.6 (167.2–337.5) 105.0 (60.0–210.0) 0.763 Colombia 219.7 (197.8–242.0) 111.5 (60.0–210.0) 250.2 (213.2–290.6) 105.0 (60.0–280.0) 258.1 (181.7–355.6) 140.0 (87.0–329.0) 0.356 Costa Rica 269.2 (227.2–317.2) 140.0 (60.0–300.0) 284.7 (249.3–319.6) 168.5 (70.0–360.0) 253.1 (187.8–326.0) 140.0 (70.0–300.0) 0.338 Ecuador 227.5 (200.8–255.5) 140.0 (75.0–243.7) 216.6 (185.6–249.0) 120.0 (75.0–210.0) 210.3 (155.6–267.3) 140.0 (75.0–243.7) 0.300 Peru 201.5 (180.9–225.9) 120.0 (60.0–210.0) 169.7 (149.7–193.5) 105.0 (60.0–210.0) 213.3 (182.5–247.5) 140.0 (70.0–240.0) 0.169 Venezuela 174.2 (155.2–195.4) 100.0 (50.0–180.0) 195.7 (147.8–251.4) 105.0 (60.0–190.0) 170.7 (114.2–240.8) 95.0 (60.0–210.0) 0.618 Overall 218.1 (208.7–227.4) 120.0 (60.0–240.0) 222.2 (211.2–233.2) 120.0 (60.0–240.0) 225.2 (203.5–246.9) 120.0 (60.0–238.0) 0.471 Leisure physical activity (min/week) Argentina 312.9 (270.2–358.9) 200.0 (90.0–390.0) 281.8 (249.1–317.3) 180.0 (80.0–365.0) 316.5 (217.4–438.8) 225.0 (112.5–444.0) 0.608 Brazil 293.0 (252.4–336.3) 140.0 (60.0–360.0) 332.8 (295.4–371.9) 230.0 (87.5–477.5) 307.6 (239.1–381.9) 180.0 (90.0–442.5) 0.003 Chile 323.0 (277.4–371.5) 190.0 (90.0–420.0) 340.9 (291.7–393.3) 210.0 (76.2–420.0) 411.3 (304.7–512.4) 300.0 (120.0–438.7) 0.346 Colombia 286.2 (252.9–325.5) 143.5 (60.0–360.0) 310.9 (264.5–358.3) 180.0 (80.0–406.0) 235.8 (178.3–303.1) 195.0 (40.0–303.7) 0.252 Costa Rica 302.7 (243.4–366.3) 180.0 (62.5–360.0) 294.0 (248.2–342.3) 180.0 (80.0–360.0) 270.6 (202.2–346.7) 138.0 (60.0–375.0) 0.360 Ecuador 418.9 (371.8–468.6) 270.0 (140.0–540.0) 375.7 (327.5–430.0) 240.0 (120.0–440.0) 408.9 (320.0–499.3) 240.0 (150.0–507.5) 0.314 Peru 270.2 (235.0–312.2) 139.0 (60.0–330.0) 268.7 (226.6–312.3) 155.0 (60.0–360.0) 281.6 (234.5–333.1) 200.0 (75.0–420.0) 0.035 Venezuela 344.7 (299.5–391.2) 240.0 (120.0–420.0) 313.9 (220.2–405.6) 180.0 (60.0–472.5) 246.2 (142.8–363.9) 120.0 (56.0–480.0) 0.466 Overall 316.8 (300.7–316.9) 180.0 (80.0–405.0) 315.1 (298.6–331.7) 180.0 (90.0–416.0) 317.9 (288.4–347.4) 210.0 (90.0–420.0) 0.081 Sitting time (min/day) Argentina 539.5 (515.6–563.4) 480.0 (330.0–720.0) 544.0 (519.3–568.4) 510.0 (352.5–720.0) 522.6 (452.6–589.8) 495.0 (240.0–750) 0.920 Brazil 401.5 (382.6–420.8) 360.0 (180.0–558.0) 453.9 (433.9–474.5) 390.0 (240.0–600.0) 518.9 (467.3–572.5) 360.0 (180.0–558.0) <0.001 Chile 457.7 (430.3–484.2) 420.0 (240.0–600.0) 506.5 (477.6–534.3) 480.0 (300.0–660.0) 532.2 (474.5–592.9) 480.0 (360.0–660.0) 0.026 Colombia 474.1 (451.2–498.9) 420.0 (240.0–660.0) 497.8 (466.4–528.4) 440.0 (290.0–660.0) 479.5 (417.9–549.1) 480.0 (300.0–600.0) 0.710 Costa Rica 405.1 (371.6–438.8) 360.0 (180.0–540.0) 480.1 (451.1–512.0) 420.0 (240.0–660.0) 457.5 (394.1–523.3) 360.0 (180.0–660.0) 0.140 Ecuador 331.1 (302.6–358.3) 265.0 (150.0–480.0) 355.8 (329.2–381.4) 300.0 (180.0–480.0) 362.6 (316.7–414.1) 360.0 (180.0–480.0) 0.185 Peru 515.9 (488.6–540.2) 480.0 (300.0–690.0) 515.6 (483.9–546.6) 480.0 (330.0–660.0) 565.0 (529.8–599.2) 525.0 (360.0–712.5) 0.139 Venezuela 377.9 (360.4–394.9) 360.0 (180.0–540.0) 412.8 (371.4–451.2) 390.0 (195.0–600.0) 367.3 (295.5–444.5) 315.0 (180.0–480.0) 0.293 Overall 428.1 (415.9–440.3) 370.0 (210.0–600.0) 471.5 (457.7–485.2) 420.0 (240.0–660.0) 490.9 (466.0–515.7) 477.5 (300.0–660.0) <0.001 95% CI: 95% confidence interval The presence of “ciclovías” improved the participation of adults in active transportation by walking [45]. Discussion Results from such programs are important because they can sup- port the actions described in the new urban plans of sev- eral countries from Latin America (i.e., Peru, Santiago, Colombia, and Brazil). These plans include efforts aimed at increasing accessibility to public parks [46]. For ex- ample, The “Ciclovía-Recreativa” programs from Bogotá have shown that users of “ciclovía programmes” contrib- ute substantially to meeting physical activity guidelines and have better quality of life [47]. Such public health campaigns can inspire populations to travel by walk and cycle more [48]. “Ciclovías”, which temporarily close (70.8%). Besides the leisure physical activity, active trans- portation has beneficial effects on all-cause mortality, cardiovascular disease and cancer and can increase the physical activity levels of entire populations [12, 42, 43]. Weekly transport physical activity time was highest in Costa Rica (mean: 275.3 min/week; median: 147.0 min/ week) and lowest in Venezuela (mean: 177.6 min/week; median: 100.0 min/week). Active transportation (walking and cycling) is potentially an important contributor to health, particularly in highly urbanized regions like Latin America [39, 44], in which it may improve people’s men- tal and physical health, prevent road traffic-related in- jury, and decrease environmental pollution [43]. Ferrari et al. BMC Public Health (2019) 19:1723 Page 8 of 12 Table 5 Characteristics of participants by transport, leisure physical activity and sitting time by education level from ELANS study Country Low Medium High P value Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Mean (95% CI) Median (Q1-Q3) Transport physical activity (min/week) Argentina 265.5 (238.6–292.4) 140.0 (60.0–300.0) 221.9 (185.4–258.5) 120.0 (60.0–280.0) 175.1 (120.8–229.4) 140.0 (75.0–210.0) 0.595 Brazil 218.2 (195.3–241.1) 107.5 (60.0–213.7) 190.2 (171.8–208.6) 105.0 (60.0–210.0) 216.8 (158.9–274.7) 135.0 (50.0–217.5) 0.644 Chile 205.5 (179.9–231.0) 105.0 (60.0–212.5) 218.1 (171.6–264.7) 120.0 (60.0–213.7) 281.2 (186.3–376.1) 140.0 (60.0–315.0) 0.267 Colombia 242.8 (217.9–267.7) 120.0 (60.0–276.0) 199.4 (163.9–234.8) 105.0 (49.7–210.0) 228.0 (173.9–282.1) 120.0 (60.0–240.0) 0.058 Costa Rica 276.5 (248.5–304.5) 157.5 (60.0–346.2) 260.7 (189.4–331.9) 120.0 (70.0–290.0) 289.6 (184.2–394.9) 140.0 (75.0–390.0) 0.451 Ecuador 226.7 (204.9–248.5) 140.0 (75.0–240.0) 174.2 (113.3–235.2) 100.0 (60.0–150.0) 231.6 (148.0–315.1) 120.0 (70.0–210.0) 0.009 Peru 193.3 (163.1–223.6) 107.5 (60.0–210.0) 195.9 (177.3–214.6) 120.0 (60.0–210.0) 186.9 (150.7–223.3) 120.0 (73.7–210.0) 0.590 Venezuela 175.9 (154.3–197.6) 90.0 (50.0–180.0) 196.0 (144.9–247.2) 105.0 (60.0–210.0) 170.3 (126.4–214.3) 100.0 (60.0–168.7) 0.523 Overall 230.5 (221.4–239.6) 120.0. Discussion (60.0–250.0) 200.7 (189.8–211.6) 110.0 (60.0–210.0) 215.9 (193.8–238.1) 120.0 (60.0–210.0) 0.103 Leisure physical activity (min/week) Argentina 297.0 (263.3–330.7) 180.0 (90.0–360.0) 309.4 (264.9–353.9) 240.0 (120.0–405.0) 265.9 (147.7–384.3) 180.0 (90.0–360.0) 0.019 Brazil 321.6 (281.3–361.9) 170.0 (72.5–420.0) 302.8 (262.3–343.3) 180.0 (60.0–427.5) 311.5 (247.2–375.8) 240.0 (90.0–455.0) 0.391 Chile 332.1 (289.4–374.8) 205.0 (90.0–420.0) 326.9 (254.8–399.0) 210.0 (60.0–420.0) 413.0 (307.2–518.8) 300.0 (120.0–420.0) 0.159 Colombia 283.3 (247.5–319.1) 150.0 (60.0–300.0) 305.8 (257.5–354.2) 180.0 (60.0–420.0) 298.3 (219.9–376.6) 150.0 (60.0–335.0) 0.273 Costa Rica 288.8 (250.8–326.7) 170.0 (60.0–360.0) 309.5 (212.7–406.2) 180.0 (90.0–360.0) 311.5 (201.1–421.9) 207.5 (108.7–363.7) 0.147 Ecuador 406.2 (371.4–441.1) 255.0 (127.5–527.5) 358.6 (263.9–453.1) 217.5 (120.0–425.0) 410.6 (244.8–576.4) 255.0 (145.0–447.5) 0.473 Peru 289.1 (229.7–348.5) 150.0 (60.0–340.0) 274.6 (245.3–303.9) 160.0 (61.0–360.0) 218.5 (150.8–286.2) 140.0 (60.0–240.0) 0.643 Venezuela 348.6 (294.2–403.0) 210.0 (105.0–435.0) 242.2 (185.4–298.9) 240.0 (67.5–345.0) 350.1 (268.7–431.5) 240.0 (67.5–345.0) 0.427 Overall 325.2 (310.8–339.5) 180.0 (90.0–420.0) 297.7 (279.8–315.5) 180.0 (75.0–391.2) 320.8 (287.9–353.6) 202.5 (90.0–403.7) 0.353 Sitting time (min/day) Argentina 532.6 (513.9–551.2) 480.0 (300.0–720.0) 563.9 (529.8–597.9) 540.0 (360.0–720.0) 574.8 (485.6–663.9) 540.0 (300.0–735.0) 0.243 Brazil 398.1 (379.2–416.9) 360.0 (180.0–540.0) 458.2 (436.6–479-8) 420.0 (240.0–640.0) 528.1 (480.8–575.4) 480.0 (360.0–720.0) <0.001 Chile 466.4 (443.2–489.6) 420.0 (270.0–600.0) 532.6 (491.573.6) 480.0 (340.0–720.0) 499.8 (449.5–550.0) 420.0 (360.0–600.0) 0.082 Colombia 455.4 (433.3–477.6) 420.0 (220.0–655.0) 530.6 (494.7–566.5) 480.0 (300.0–720.0.) 529.7 (477.7–581.8) 510.0 (300.0–720.0) <0.001 Costa Rica 432.8 (409.9–455.7) 360.0 (240.0–600.0) 518.6 (450.2–586.9) 450.0 (260.0–720.0) 579.5 (479.2–679.8) 540.0 (360.0–780.0) 0.005 Ecuador 333.1 (314.2–352.0) 270.0 (160.0–480.0) 416.0 (356.9–475.1) 360.0. (240.0–510.0) 370.6 (310.2–430.9) 360.0 (213.7–540.0) 0.005 Peru 529.6 (491.5–567.7) 480.0 (300.0–710.0) 519.5 (498.7–540.4) 480.0 (315.0–660.0) 560.6 (514.6–606.7) 540.0 (382.5–712.5) 0.080 Venezuela 376.1 (356.9–395.3) 360.0 (180.0–500.0) 355.1 (312.5–397.8) 315.0 (180.0–485.0) 429.9 (394.0–465.8) 420.0 (240.0–600.0) 0.001 Overall 436.5 (428.7–444.3) 370.0 (225.0–600.0) 495.7 (484.2–507.3) 456.0 (270.0–660.0) 503.7 (484.5–522.8) 480.0 (300.0–660.0) <0.001 95% CI: 95% confidence interval 95% CI: 95% confidence interval streets to private transport to create a safe place for people to cycle, walk, run, and participate in social health promo- tion and cultural events, have been shown to be very ef- fective “programmes” in the Latin American region [39]. environment, has been identified as a major correlate of physical activity levels worldwide [53]. A better under- standing of sex differences can also occur through meas- uring their participation in diverse domains of activity (i.e., transport and leisure time activities). Discussion More oppor- tunities for safe and available leisure activities for women, as well as the impacts of cultural norms, trad- itional roles, and lack of social and community support all can lead to reduced participation in physical activities among women [52]. Understanding and addressing these barriers are necessary to plan and deliver socially sensi- tive programs to support behavior change. Another way to improve leisure-time physical activity may be to pro- mote women’s involvement in sport, as women do not Sex differences in transport and leisure physical activ- ity have been reported in studies from countries with different income levels [41]. Overall, our investigation found, similarly to other regions (Mexico, Europe, and the United States) [49–51], significantly lower physical activity in women than men. Such results argue for in- terventions targeting women to help close the sex differ- ence and reach the global physical activity goals [41, 52]. Of note, how women, vs. men respond to their local built environments, including the walkability of their Ferrari et al. BMC Public Health (2019) 19:1723 Page 9 of 12 Ferrari et al. BMC Public Health (2019) 19:1723 Page 9 of 12 have the same opportunity to engage in sport. Collabor- ation from the government, sports institutions and health professionals can help to increase women’s par- ticipation in physical activity. Our study showed that in most countries, younger partic- ipants (<30 years) spent more time sitting time than older participants (≥30 years). This behaviour indicates more fre- quent sedentary occupations in the use of passive transport among younger adults. It is possible that this pattern could signal a future increase in for poor health outcomes [61]. While the need to monitor sedentary behaviours in national health surveys seems clear, few efforts exist to study seden- tary behaviour worldwide [34], and no policies to decrease sedentary behaviours exist in Latin America. Overall, physical activity (transport and leisure) and sitting time showed high variability across countries. There was, however, no clear pattern in the time spent in physical activity and sitting in relation to dif- ferences at the socioeconomic and education levels. The current patterns of physical activity and sitting time by socioeconomic strata are closely related to urban development in Latin America characterized by social and environmental inequalities, unplanned and disorganized growth, and underlying political and socioeconomic factors [54, 55]. Discussion Factors such as globalization and industrialization influenced the mi- gration from rural to urban areas. The fact that phys- ical activity and sitting time vary greatly across countries and cities suggests that the factors that in- fluence inactivity lie at national, sub-national, and community levels, and policies specific to these levels may be needed to increase physical activity [4]. y ELANS provides data allowing comparisons across eight countries from Latin American for the first time. Major inputs included the production of comparable physical activity values in eight countries using a com- mon protocol. Many manuscripts showed moderate cor- relations between IPAQ and accelerometers [62, 63]. Questionnaires remain the most practical method for measuring physical activity in populations due to the low price and high burden of respondents [64]. Com- pared with many current physical activity questionnaires, a strength of IPAQ is the ability to quantify both leisure and transport physical activity. IPAQ is widely used for measuring and tracking physical activity levels in Latin American populations [23, 36, 65]. Its use in Latin America has not been without challenges, and has re- quired several cultural and structural adaptations. IPAQ measurement results can be overestimated [66–68]. Total physical activity may have higher values than only leisure activities [69]. This between-country difference appeared even greater in the low- and middle- income countries [70]. Another limitation in this study was the complexity of socioeconomic strata classification that may have led to misclassification within the three socio- economic levels. Developing a feasible, realistic, stan- dardized socioeconomic strata categorization was more difficult than expected, requiring extensive and innova- tive work. Measurement and definitions of socioeco- nomic status and educational level across countries requires close attention to ensure comparability [33, 56]. The differences in physical activity with socioeconomic level are clearer when evaluated at the level of the trans- port and leisure domains [56]. We found differences be- tween socioeconomic level strata and leisure physical activity in Brazil and Peru. There have been reports of stronger relationship between socioeconomic strata of leisure physical activity in European countries [56]. A higher socioeconomic position is associated with better facilities and environments and more opportunities for leisure time physical activity [35]. Discussion Building more places appropriate for leisure time physical activity such as parks outdoor courts and bicycle paths [35] and improv- ing walkability of streets [33] may be important strat- egies for increasing opportunities for leisure physical activity among lower SES groups in Latin America. These actions are included in national physical activity policies in Latin America countries [57, 58]. Each country in our study had a mean level of sitting time higher than 7 hours/day. Van Dyck et al. [59] re- ported means of 7.9 and 7.8 h/day of sedentary behav- iour in Brazil and Colombia. Most countries in the current study showed socioeconomic (Brazil, Chile, Ecuador, Peru, overall) and educational (Argentina, Brazil, Colombia, Costa Rica, Peru, Venezuela and over- all) gradients in sitting time, with higher levels with higher socioeconomic and education level. Presumably adults with higher education and from higher income groups have more sedentary jobs, are more likely to use cars than active travel as a means of transport, and have more electronic entertainment and labor-saving devices at home. Cultural factors may also explain some pat- terns, through behavioral preferences [60]. Abbreviations CI95%: 95% confidence interval; CVD: Cardiovascular diseases; ELANS: Latin American Study of Nutrition and Health / Estudio Latinoamericano de Nutrición y Salud; IPAQ: International Physical Activity Questionnaire; min: minutes; PA: Physical activity; ST: Sitting time; WHO: World Health Organization CI95%: 95% confidence interval; CVD: Cardiovascular diseases; ELANS: Latin American Study of Nutrition and Health / Estudio Latinoamericano de Nutrición y Salud; IPAQ: International Physical Activity Questionnaire; min: minutes; PA: Physical activity; ST: Sitting time; WHO: World Health Organization Argentina: Comité de ética de la Asociación Médica Argentina; Brazil: Comité de ética do Instituto Pensi – Fundação José Luiz Setubal – Hospital Infantil Sabara; Chile: Comité de ético científico de la Facultad de Medicina de la Pontificia Universidad Católica de Chile; Colombia: Comité de Investigación y ética de la Faculdade de Ciencias de la Pontificia Universidad Javeriana; Costa Rica: Comité ético científico de la Vicerrectoría de Investigación de La Universidad de Costa Rica; Ecuador: Comité de Bioética Universidad de San Francisco de Quito; Peru: Comité Institucional de ética del Instituto de Investigación Nutricional; Venezuela: Comisión de Bioética de la Escuela de Antropología de la Universidad Central de Venezuela. Authors’ contributions GLMF, and ACK, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. The corresponding author had final responsibility for the decision to submit for publication. Study concept and design: GLMF, and ACK. Data collection: GLMF, IK, MF, GG, AR, LYCS, MCYG, RGPT, MH-C, IZZ, VG, MP, and DS. Statistical analysis: GLMF, and DS. Drafting of the manuscript: GLMF, and ACK. All authors have provided a critical revision and final approval of the manuscript. Competing interests All authors declare that they have no competing interests. Competing interests All authors declare that they have no competing interests. Author details 1 1Centro de Investigación en Fisiologia del Ejercicio – CIFE, Universidad Mayor, José Toribio Medina, 29. Estacion Central, Santiago, Chile. 2Departamento de Pediatria da Universidade Federal de São Paulo, São Paulo, Brazil. 3Commitee of Nutrition and Wellbeing, International Life Science Institute (ILSI-Argentina), Buenos Aires, Argentina. 4Instituto Pensi, Fundação José Luiz Egydio Setubal, Hospital Infantil Sabará, São Paulo, Brazil. 5Departamento de Bioquímica, Escuela de Medicina, Universidad de Costa Rica, San José, Costa Rica. 6Centro de Nutrición Molecular y Enfermedades Crónicas, Departamento de Nutrición, Diabetes y Metabolismo, Escuela de Medicina, Pontificia Universidad Católica, Santiago, Chile. 7Departamento de Nutrición y Bioquímica, Pontificia Universidad Javeriana, Bogotá, Colombia. 8Colégio de Ciencias de la Salud, Universidad San Francisco de Quito, Quito, Ecuador. 9Instituto de Investigación Nutricional, La Molina, Lima, Peru. 10Centro de Estudios del Desarrollo, Universidad Central de Venezuela (CENDES-UCV)/Fundación Bengoa, Caracas, Venezuela. 11Departamento de Psicobiologia, Universidade Federal de São Paulo, São Paulo, Brazil. 12Institute for Public Health, University of California San Diego, La Jolla, CA, USA. 13Health Research & Policy Department and the Stanford Prevention Research Center, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA. Ethics approval and consent to participate h d d d d h Additional file 5: Figure S5. Prevalence (% and 95 confidence interval) of insufficient physical activity by education level from eight Latin America countries. Additional file 5: Figure S5. Prevalence (% and 95 confidence interval) of insufficient physical activity by education level from eight Latin America countries. This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects/patients and each site specific protocol was also approved by the ethical review boards of the participating institutions. The overarching ELANS protocol was approved by the Western Institutional Review Board (#20140605) and is registered at Clinical Trials (#NCT02226627). Availability of data and materials h d d d l d d Availability of data and materials The dataset used and analysed during the current study are available from the corresponding author on reasonable request. Availability of data and materials The dataset used and analysed during the current study are available from the corresponding author on reasonable request. Additional file 4: Figure S4. Prevalence (% and 95 confidence interval) of insufficient physical activity by socioeconomic level from eight Latin America countries. Additional file 4: Figure S4. Prevalence (% and 95 confidence interval) of insufficient physical activity by socioeconomic level from eight Latin America countries. Acknowledgments We would like to thank the following individuals at each of the participating sites who made substantial contributions to the ELANS: Luis A. Moreno, Beate Lloyd, Brenda Lynch, Mariela Jauregui, Alejandra Guidi, Luis Costa, and Regina Mara Fisberg. An document provided a short description of the purpose of the survey, confidentiality practices, contact information, and a link to the survey. Participants were considered consented once they read the document and signed to the survey. Informed assent was obtained from every adolescents and all parents and/or legal guardians signed an informed consent. All participants signed a written informed consent/assent before commencement of the study. Participants’ confidentiality for the pooled data was maintained using numeric identification codes rather than names. All data transfer was done with a secure file sharing system. ‡The following are members of ELANS Study Group: Chairs: Mauro Fisberg and Irina Kovalskys; Co-chair: Georgina Gómez; Core Group members: Attilio Rigotti, Lilia Yadira Cortés Sanabria, Georgina Gómez, Martha Cecilia Yépez García, Rossina Gabriella Pareja Torres, and Marianella Herrera-Cuenca; Steer- ing Committee: Berthold Koletzko, Luis A. Moreno, Michael Pratt, and Regina Mara Fisberg; Project Managers: Viviana Guajardo and Ioná Zalcman Zimberg; International Life Sciences Institute-Argentina: Irina Kovalskys, Viviana Gua- jardo, María Paz Amigo, Ximena Janezic; Instituto Pensi-Hospital Infantil Sabara-Brazil: Mauro Fisberg, Ioná Zalcman Zimberg, and Natasha Aparecida Grande de França; Pontificia Universidad Católica de Chile: Attilio Rigotti, Guadalupe Echeverría, Leslie Landaeta, and Óscar Castillo; Pontificia Universi- dad Javeriana-Colombia: Lilia Yadira Cortés Sanabria, Luz Nayibe Vargas, Luisa Fernanda Tobar, and Yuri Milena Castillo; Universidad de Costa Rica Georgina Gómez and Anne Chinnock, Instituto Costarricense de Enseñanza e Investiga- ción en Nutrición y Salud: Rafael Monge Rojas; Universidad San Francisco de Quito-Ecuador: Martha Cecilia Yépez García, Mónica Villar Cáceres, and María Belén Ocampo; Instituto de Investigación Nutricional-Perú: Rossina Pareja Torres, María Reyna Liria, Krysty Meza; Universidad Central de Venezuela: Mar- ianella Herrera-Cuenca, Maritza Landaeta, Betty Méndez, Maura Vasquez, Omaira Rivas, Carmen Meza, Servando Ruiz, Guillermo Ramirez, and Pablo Hernández; Accelerometry analysis: Priscila Bezerra Gonçalves and Claudia Alberico; Physical activity advisor: Gerson Luis de Moraes Ferrari. Supplementary information Supplementary information Supplementary information accompanies this paper at https://doi.org/10. 1186/s12889-019-8048-7. Additional file 1: Figure S1. Prevalence (% and 95 confidence interval) of insufficient physical activity from eight Latin America countries. Additional file 2: Figure S2. Prevalence (% and 95 confidence interval) of insufficient physical activity by sex from eight Latin America countries. Additional file 3: Figure S3. Prevalence (% and 95 confidence interval) of insufficient physical activity by age group from eight Latin America countries. Additional file 4: Figure S4. Prevalence (% and 95 confidence interval) of insufficient physical activity by socioeconomic level from eight Latin America countries. Additional file 5: Figure S5. Prevalence (% and 95 confidence interval) of insufficient physical activity by education level from eight Latin America countries. (CENDES-UCV)/Fundación Bengoa, Universidad San Francisco de Quito, and Instituto de Investigación Nutricional de Peru. The founding sponsors had no role in study design, in the collection, analyses, or interpretation of data, in the writing of the manuscript, and in the decision to publish the results. Dr. King received partial funding support from the U.S. National Institutes of Health National Cancer Institute grant P20CA217199 (King, PI). This study is registered at www.clinicaltrials.gov (No. NCT02226627). Supplementary information Supplementary information accompanies this paper at https://doi.org/10. 1186/s12889-019-8048-7. Additional file 3: Figure S3. Prevalence (% and 95 confidence interval) of insufficient physical activity by age group from eight Latin America countries. Consent for publication Not applicable. Consent for publication Not applicable. Conclusions The study findings show wide variation in transport and leisure physical activity and sitting time by sex and age group in eight Latin America countries. The results do not show significant difference in transport and leisure physical activity by socioeconomic and education levels. The ob- served variability across countries sets the stage for future investigations to inform interventions at the national and regional levels. Future studies should seek to better under- stand the challenges of promoting transport and leisure physical activity and reducing sitting time in urban regions. Such studies may help in gaining a deeper understanding of the factors that can be targeted to increase physical activity in Latin America. Page 10 of 12 Page 10 of 12 Ferrari et al. BMC Public Health (2019) 19:1723 Page 10 of 12 References 2012;380(9838): 219–29. 5. Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT, et al. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet. 2012;380(9838): 219–29. 29. Instituto Nacional de Estadística y Censos de Ecuador. Encuesta de Estratificación de Nivel Socioeconómico; 2011. 6. Stamatakis E, Ekelund U, Ding D, Hamer M, Bauman AE, Lee IM. Is the time right for quantitative public health guidelines on sitting? A narrative review of sedentary behaviour research paradigms and findings. Br J Sports Med. 2019;53(6):377–82. 30. Asociacion Investigadores de Mercado. Grupos Socioeconómicos Chile. Chile: Asociacion Investigadores de Mercado; 2012. 31. Instituto Nacional de Estadística. República Bolivariana de Venezuela, Síntesis Estadística de Pobreza e Indicadores de Desigualdad; 2011. 7. Schmid D, Ricci C, Leitzmann MF. Associations of objectively assessed physical activity and sedentary time with all-cause mortality in US adults: the NHANES study. PLoS One. 2015;10(3):e0119591. 32. IBM Corp. Released. IBM SPSS Statistics for Windows, Version 22.0. Armonk: IBM Corp; 2013. 33. Poggio R, Seron P, Calandrelli M, Ponzo J, Mores N, Matta MG, et al. Prevalence, patterns, and correlates of physical activity among the adult population in Latin America: cross-sectional results from the CESCAS I Study. Glob Heart. 2016;11(1):81–8 e81. 8. 2018 Physical Activity Guidelines Advisory Committee. 2018 Physical Activity Guidelines Advisory Committee Scientific Report. Washington, DC: US. Department of Health and Human Services; 2018. 9. Barreto SM, Miranda JJ, Figueroa JP, Schmidt MI, Munoz S, Kuri-Morales PP, et al. Epidemiology in Latin America and the Caribbean: current situation and challenges. Int J Epidemiol. 2012;41(2):557–71. 34. Werneck AO, Baldew SS, Miranda JJ, Diaz Arnesto O, Stubbs B, Silva DR, et al. Physical activity and sedentary behavior patterns and sociodemographic correlates in 116,982 adults from six South American countries: the South American physical activity and sedentary behavior network (SAPASEN). Int J Behav Nutr Phys Act. 2019;16(1):68. 10. Ding D, Lawson KD, Kolbe-Alexander TL, Finkelstein EA, Katzmarzyk PT, van Mechelen W, et al. The economic burden of physical inactivity: a global analysis of major non-communicable diseases. Lancet. 2016;388(10051): 1311–24. y 35. Salvo D, Sarmiento OL, Reis RS, Hino AAF, Bolivar MA, Lemoine PD, et al. Where Latin Americans are physically active, and why does it matter? Findings from the IPEN-adult study in Bogota, Colombia; Cuernavaca, Mexico; and Curitiba, Brazil. Prev Med. 2017;103S:S27–33. 11. Funding h The ELANS was supported by a scientific grant from the Coca Cola Company, and support from the Ferrero, Instituto Pensi / Hospital Infantil Sabara, International Life Science Institute of Argentina, Universidad de Costa Rica, Pontificia Universidad Católica de Chile, Pontificia Universidad Javeriana, Universidad Central de Venezuela Page 11 of 12 Page 11 of 12 Ferrari et al. BMC Public Health (2019) 19:1723 Ferrari et al. BMC Public Health (2019) 19:1723 Received: 20 June 2019 Accepted: 8 December 2019 References Rivera JA, Barquera S, Campirano F, Campos I, Safdie M, Tovar V. Epidemiological and nutritional transition in Mexico: rapid increase of non- communicable chronic diseases and obesity. Public Health Nutr. 2002;5(1A): 113–22. 36. Pitanga FJG, Matos SMA, Almeida MDC, Barreto SM, Aquino EML. Leisure- time physical activity, but not commuting physical activity, is associated with cardiovascular risk among ELSA-Brasil participants. Arq Bras Cardiol. 2018;110(1):36–43. 12. Celis-Morales CA, Lyall DM, Welsh P, Anderson J, Steell L, Guo Y, et al. Association between active commuting and incident cardiovascular disease, cancer, and mortality: prospective cohort study. BMJ. 2017;357:j1456. 37. Mindell JS, Moody A, Vecino-Ortiz AI, Alfaro T, Frenz P, Scholes S. et all. Comparison of health examination survey methods in Brazil, Chile, Colombia, Mexico, England, Scotland, and the United States. Am J Epidemiol. 2017;186(6):648–58. 13. Saunders LE, Green JM, Petticrew MP, Steinbach R, Roberts H. What are the health benefits of active travel? A systematic review of trials and cohort studies. PLoS One. 2013;8(8):e69912. 14. Kelly P, Kahlmeier S, Gotschi T, Orsini N, Richards J, Roberts N, et al. Systematic review and meta-analysis of reduction in all-cause mortality from walking and cycling and shape of dose response relationship. Int J Behav Nutr Phys Act. 2014;11:132. 38. Saint-Maurice PF, Troiano RP, Matthews CE, Kraus WE. Moderate-to-vigorous physical activity and all-cause mortality: do bouts matter? J Am Heart Assoc. 2018;7(6):e007678. 39. Gomez LF, Sarmiento R, Ordonez MF, Pardo CF, de Sa TH, Mallarino CH, et al. Urban environment interventions linked to the promotion of physical activity: a mixed methods study applied to the urban context of Latin America. Soc Sci Med. 2015;131:18–30. 15. Knuth AG, Hallal PC. Temporal trends in physical activity: a systematic review. J Phys Act Health. 2009;6(5):548–59. 16. Koyanagi A, Stubbs B, Vancampfort D. Correlates of low physical activity across 46 low- and middle-income countries: a cross-sectional analysis of community-based data. Prev Med. 2018;106:107–13. 40. Hoehner CM, Ribeiro IC, Parra DC, Reis RS, Azevedo MR, Hino AA, et al. Physical activity interventions in Latin America: expanding and classifying the evidence. Am J Prev Med. 2013;44(3):e31–40. 17. United Nations, 2012. World urbanization Prospects: The 2011 Revision: Data tables and highlights. 2011 rev ed. 2012. 41. Guthold R, Stevens GA, Riley LM, Bull FC. Worldwide trends in insufficient physical activity from 2001 to 2016: a pooled analysis of 358 population- based surveys with 1.9 million participants. Lancet Glob Health. 2018;6(10): e1077–86. 18. Received: 20 June 2019 Accepted: 8 December 2019 23. Hallal PC, Gomez LF, Parra DC, Lobelo F, Mosquera J, Florindo AA, et al. Lessons learned after 10 years of IPAQ use in Brazil and Colombia. J Phys Act Health. 2010;7(Suppl 2):S259–64. 24. Sedentary Behaviour Research Network. Standardized use of the terms “sedentary” and “sedentary behaviours”. Appl Physiol Nutr Metab. 2012;7: 540–2. References 1. Gersh BJ, Sliwa K, Mayosi BM, Yusuf S. Novel therapeutic concepts: the epidemic of cardiovascular disease in the developing world: global implications. Eur Heart J. 2010;31(6):642–8. 1. Gersh BJ, Sliwa K, Mayosi BM, Yusuf S. Novel therapeutic concepts: the epidemic of cardiovascular disease in the developing world: global implications. Eur Heart J. 2010;31(6):642–8. 1. Gersh BJ, Sliwa K, Mayosi BM, Yusuf S. Novel therapeutic concepts: the epidemic of cardiovascular disease in the developing world: global implications. Eur Heart J. 2010;31(6):642–8. 25. Bauman A, Ainsworth BE, Sallis JF, Hagstromer M, Craig CL, Bull FC, et al. The descriptive epidemiology of sitting. A 20-country comparison using the International Physical Activity Questionnaire (IPAQ). Am J Prev Med. 2011; 41(2):228–35. 2. Lachman S, Boekholdt SM, Luben RN, Sharp SJ, Brage S, Khaw KT, et al. Impact of physical activity on the risk of cardiovascular disease in middle aged and older adults: EPIC Norfolk prospective population study. Eur J Prev Cardiol. 2018;25(2):200–8. 2. Lachman S, Boekholdt SM, Luben RN, Sharp SJ, Brage S, Khaw KT, et al. Impact of physical activity on the risk of cardiovascular disease in middle aged and older adults: EPIC Norfolk prospective population study. Eur J Prev Cardiol. 2018;25(2):200–8. 26. Associação Brasileira de Empresas de Pesquisa (ABEP). Critério padrão de classificação econômica Brasil; 2013. 3. World Health Organization (WHO). Global recommendations on physical activity for health. Geneva: World Health Organization; 2010. 3. World Health Organization (WHO). Global recommendations on physical activity for health. Geneva: World Health Organization; 2010. 27. Comisión de Enlace Institucional AAM-SAIMO-CEIM. Nivel Socioeconómic Antecedentes, marco conceptual, enfoque metodológico y fortalezas. Buenos Aires: Comisión de Enlace Institucional AAM-SAIMO-CEIM; 2006. activity for health. Geneva: World Health Organization; 2010. 4. Global action plan on physical activity 2018–2030. More active people for a healthier world. Geneva: World Health Organization; 2018. 4. Global action plan on physical activity 2018–2030. More active people for a healthier world. Geneva: World Health Organization; 2018. 4. Global action plan on physical activity 2018–2030. More active people for a healthier world. Geneva: World Health Organization; 2018. healthier world. Geneva: World Health Organization; 2018. 28. Departamento Administrativo Nacional de Estadisticas de Colombia. Proyecciones nacionales y departamentales de poblacion 2005–2020, Estudios Postcensales No. 7; 2009. 5. Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT, et al. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet. References Salvo D, Reis RS, Sarmiento OL, Pratt M. Overcoming the challenges of conducting physical activity and built environment research in Latin America: IPEN Latin America. Prev Med. 2014;69(Suppl 1):S86–92. 19. Soares RR, Naritomi J. Understanding high crime rates in Latin America: the role of social and policy factors. Chicago: The economics of crime: Lessons for and from Latin America. University of Chicago Press; 2010. 42. Ostergaard L, Jensen MK, Overvad K, Tjonneland A, Grontved A. Associations between changes in cycling and all-cause mortality risk. Am J Prev Med. 2018;55(5):615–23. 20. World Bank. https://data.worldbank.org/indicator/SP.URB.TOTL.IN. ZS?locations=ZJ&name_desc=false. Accessed 11 Sept 2019. 43. Dinu M, Pagliai G, Macchi C, Sofi F. Active commuting and multiple health outcomes: a systematic review and meta-analysis. Sports Med. 2019;49(3): 437–52. 21. Fisberg M, Kovalskys I, Gomez G, Rigotti A, Cortes LY, Herrera-Cuenca M, et al. Latin American Study of Nutrition and Health (ELANS): rationale and study design. BMC Public Health. 2016;16(1):93. 44. Steell L, Garrido-Mendez A, Petermann F, Diaz-Martinez X, Martinez MA, Leiva AM, et al. Active commuting is associated with a lower risk of obesity, diabetes and metabolic syndrome in Chilean adults. J Public Health (Oxf). 2018;40(3):508–16. 22. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35(8):1381–95. Page 12 of 12 Page 12 of 12 Page 12 of 12 Ferrari et al. BMC Public Health (2019) 19:1723 45. Florindo AA, Barrozo LV, Cabral-Miranda W, Rodrigues EQ, Turrell G, Goldbaum M, et al. Public open spaces and leisure-time walking in brazilian adults. Int J Environ Res Public Health. 2017;14(6):E553. 45. Florindo AA, Barrozo LV, Cabral-Miranda W, Rodrigues EQ, Turrell G, Goldbaum M, et al. Public open spaces and leisure-time walking in brazilian adults. Int J Environ Res Public Health. 2017;14(6):E553. 66. Sallis JF, Saelens BE. Assessment of physical activity by self-report: status, limitations, and future directions. Res Q Exerc Sport. 2000;71(Suppl 2):1–14. 67. Rzewnicki R, Vanden Auweele Y, De Bourdeaudhuij I. Addressing overreporting on the International Physical Activity Questionnaire (IPAQ) telephone survey with a population sample. Public Health Nutr. 2003;6(3): 299–305. 46. Rubinstein A, Miranda JJ, Beratarrechea A, Diez-Canseco F, Kanter R, Gutierrez L, et al. Effectiveness of a health intervention to improve the cardiometabolic profile of people with prehypertension in low-resource urban settings in Latin America: a randomized controlled trial. Lancet Diabetes Endocrinol. 2016;4(1):52–63. 68. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 50. Dilley JR, Moore JB, Summers P, Price AA, Burczyk M, Byrd L, et al. A citizen science approach to determine physical activity patterns and demographics of greenway users in winston-salem, North Carolina. Int J Environ Res Public Health. 2019;16(17):E3150. 51. Medina C, Romero-Martinez M, Bautista-Arredondo S, Barquera S, Janssen I. Move on bikes program: a community-based physical activity strategy in Mexico city. Int J Environ Res Public Health. 2019;16(10):E1685. 52. Mielke GI, da Silva ICM, Kolbe-Alexander TL, Brown WJ. Shifting the physical inactivity curve worldwide by closing the gender gap. Sports Med. 2018; 48(2):481–9. 53. Althoff T, Sosic R, Hicks JL, King AC, Delp SL, Leskovec J. Large-scale physical activity data reveal worldwide activity inequality. Nature. 2017;547(7663): 336–9. 54. ONU-HABITAT. State of Latin American and Caribbean cities. 2012. 55. Simoes EJ, Hallal P, Pratt M, Ramos L, Munk M, Damascena W, et al. Effects of a community-based, professionally supervised intervention on physical activity levels among residents of Recife, Brazil. Am J Public Health. 2009; 99(1):68–75. 56. Stalsberg R, Pedersen AV. Are differences in physical activity across socioeconomic groups associated with choice of physical activity variables to report? Int J Environ Res Public Health. 2018;15(5):E922. 57. Brazilian ministry of Health. Strategic action plan for dealing with noncommunicable diseases (NCDs) in Brazil from 2011 to 2022. Brasília; 2011. 58. Gobierno de Chile. Estrategia Nacional de Salud Para el cumplimiento de los Objetivos Sanitarios de la Década 2011-2020. Santiago; 2011. 59. Van Dyck D, Cerin E, De Bourdeaudhuij I, Hinckson E, Reis RS, Davey R, et al. International study of objectively measured physical activity and sedentary time with body mass index and obesity: IPEN adult study. Int J Obes. 2015; 39(2):199–207. 60. Bauman A, Ainsworth BE, Bull F, Craig CL, Hagstromer M, Sallis JF, et al. Progress and pitfalls in the use of the International Physical Activity Questionnaire (IPAQ) for adult physical activity surveillance. J Phys Act Health. 2009;6(Suppl 1):S5–8. 61. Vancampfort D, Firth J, Schuch FB, Rosenbaum S, Mugisha J, Hallgren M, et al. Sedentary behavior and physical activity levels in people with schizophrenia, bipolar disorder and major depressive disorder: a global systematic review and meta-analysis. References Vandelanotte C, De Bourdeaudhuij I, Sallis JF, Spittaels H, Brug J. Efficacy of sequential or simultaneous interactive computer-tailored interventions for increasing physical activity and decreasing fat intake. Ann Behav Med. 2005; 29(2):138–46. 47. Sarmiento O, Torres A, Jacoby E, Pratt M, Schmid TL, Stierling G. The ciclovia-recreativa: a mass-recreational program with public health potential. J Phys Act Health. 2010;7(Suppl 2):S163–80. 47. Sarmiento O, Torres A, Jacoby E, Pratt M, Schmid TL, Stierling G. The ciclovia-recreativa: a mass-recreational program with public health potential. J Phys Act Health. 2010;7(Suppl 2):S163–80. 69. Pereira MA, FitzerGerald SJ, Gregg EW, Joswiak ML, Ryan WJ, Suminski RR, et al. A collection of Physical Activity Questionnaires for health-related research. Med Sci Sports Exerc. 1997;29(6 Suppl):S1–205. 48. Mansfield TJ, Gibson JM. Estimating active transportation behaviors to support health impact assessment in the United States. Front Public Health. 2016;4:63. 48. Mansfield TJ, Gibson JM. Estimating active transportation behaviors to support health impact assessment in the United States. Front Public Health. 2016;4:63. 70. Guthold R, Ono T, Strong KL, Chatterji S, Morabia A. Worldwide variability in physical inactivity a 51-country survey. Am J Prev Med. 2008;34(6):486–94. 49. Te Velde SJ, Haraldsen E, Vik FN, De Bourdeaudhuij I, Jan N, Kovacs E, et al. Associations of commuting to school and work with demographic variables and with weight status in eight European countries: The ENERGY-cross sectional study. Prev Med. 2017;99:305–12. Publisher’s Note World Psychiatry. 2017;16(3):308–15. 62. Scholes S, Coombs N, Pedisic Z, Mindell JS, Bauman A, Rowlands AV, et al. Age- and sex-specific criterion validity of the health survey for England Physical Activity and Sedentary Behavior Assessment Questionnaire as compared with accelerometry. Am J Epidemiol. 2014;179(12):1493–502. 63. Dyrstad SM, Hansen BH, Holme IM, Anderssen SA. Comparison of self- reported versus accelerometer-measured physical activity. Med Sci Sports Exerc. 2014;46(1):99–106. 64. Scholes S, Bridges S, Ng Fat L, Mindell JS. Comparison of the physical activity and sedentary behaviour assessment questionnaire and the short- form international physical activity questionnaire. An analysis of health survey for England data. PLoS One. 2016;11(3):e0151647. 65. Silva LJ, Matsudo VK, Andrade DR, Azevedo M, Ferrari GL, et al. The prevalence of physical activity and its associated effects among students in the Sao Paulo public school network, Brazil. Cien Saude Colet. 2016;21(4): 1095–103.
https://openalex.org/W4211156615	https://www.nature.com/articles/s41467-022-28452-z.pdf	English	null	Copy number amplification of ENSA promotes the progression of triple-negative breast cancer via cholesterol biosynthesis	Nature communications	2,022	cc-by	18,773	ARTICLE ARTICLE NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 1 Department of Breast Surgery, Shanghai Cancer Center and Cancer Institute, Shanghai Medical College, Fudan University, 200032 Shanghai, P. R. China. 2 Human Phenome Institute, Fudan University, 825 Zhangheng Road, 201203 Shanghai, P. R. China. ✉email: yukeda@fudan.edu.cn Results As a hallmark of cancer, CNAs are ubiquitous in cancers and are the most com- mon type of somatic genetic events10,11. Altered expression of oncogenes or tumor suppressors mediated by CNA is linked to the development, clinicopathological characteristics, and prog- noses of cancers12,13. According to a previously deﬁned onco- genic signature across human cancers, breast cancer falls into a class where recurrent CNAs are predominant over mutations14. Understanding the phenotypic effects and underlying mechanism of CNAs in breast cancer has brought advantages in targeted treatment, as highlighted by the application of trastuzumab-based therapy in HER2-overexpressing breast cancer15. In breast cancer, tumorigenesis driven by genomic instability is most prevalent in basal-like subtype breast cancers (most of which are TNBC), where tumors exhibit extensive CNAs16,17. Our previous study on a Chinese TNBC cohort revealed recurrent copy number gains in chromosomes 1q, 8q, and 10p and copy number deletions in chromosome 8p7. Although some reported oncogenes and tumor suppressors are affected by the most frequent CNAs, many new CNA-affected genes in TNBC remain unexplored. g p Cholesterol is a precursor of bile acids, steroid hormones, vitamin D, and a component of cell membranes. It plays a critical role in cell growth and differentiation. Upregulated cholesterol biosynthesis has been discovered in several cancers, where it supports the growth, metastasis, stemness, and therapeutic resistance of tumors18,19. In breast cancer, cholesterol and its metabolites have been found to promote tumor progression both preclinically and clinically20–22. Compared to the ER + subtype, TNBC displays elevated cholesterol biosynthesis, which could have profound biological functions and indicate potential ther- apeutic strategies23. However, there is a limited understanding of how CNAs activate cholesterol biosynthesis programs in TNBC at the genetic level. ENSA promotes the growth of TNBC cells. To determine the roles of ENSA in TNBC, we knocked down its expression using shRNA and restored the expression of its most common tran- script in BT549 and MDA-MB-231 cells, two TNBC cell lines with relatively high expression of ENSA (Fig. 2a and Supple- mentary Fig. 3a). ENSA downregulation markedly impaired the cell growth and colony formation of TNBC cells (Fig. 2b, c). However, the growth of ENSA-depleted luminal and HER2 cells was only slightly inhibited, less than the results shown in TNBC cells (Supplementary Fig. 3b). In addition, overexpression of ENSA in TNBC cells attenuated the suppression of cell growth and colony formation induced by ENSA silencing in vitro (Fig. Results Results Integrated DNA copy number and transcriptome proﬁling reveals ampliﬁed ENSA at the 1q21.3 region in TNBC. To identify dysregulated gene expression programs driven by CNAs in TNBC, we ﬁrst integrated gene-level CNA data and RNA-seq data of 302 female patients from the Fudan University Shanghai Cancer Center (FUSCC) TNBC cohort to screen CNA-affected oncogenes with the criteria listed in Fig. 1a. A total of 41 genes located in several recurrent CNA regions were found (Supple- mentary Fig. 1). Among them, ENSA and Golgi phosphoprotein 3 like (GOLPH3L), located at the 1q21.3 peak, were the most fre- quently ampliﬁed genes in TNBC (ampliﬁcation in 18.5% and copy number gain in 57.6% of patients), followed by genes at the 1q43 and 10p15.1 loci (Fig. 1a). In the TCGA database, the 1q21.3 segment was also frequently ampliﬁed in approximately 11% of all breast cancer patients (Fig. 1b) and predicted worse disease-free survival and disease-speciﬁc survival (Fig. 1c). However, ampliﬁcation of the 1q43 and 10p15.1 regions showed no prognostic signiﬁcance (Supplementary Fig. 2a). We further investigated the difference in 1q21.3 alteration between different subtypes of breast cancer in the TCGA cohort. We found a higher ampliﬁcation frequency in the TNBC (27.7%) and basal-like (32.6%) subtypes (Fig. 1d), which implies that this amplicon is more critical in TNBC than in other subtypes. To identify potential tumor-promoting genes in the 1q21.3 locus of interest, we performed survival analysis of ENSA and GOLPH3L in the Kaplan–Meier plotter database and observed that elevated expression of ENSA, not GOLPH3L, was linked with poor relapse-free survival in TNBC and basal-like breast cancer (Fig. 1e, f). However, ENSA expression did not predict the sur- vival outcomes of patients with other breast cancer subtypes, further suggesting its importance in the TNBC subtype (Sup- plementary Fig. 2b). ENSA expression was upregulated in several tumors and was markedly upregulated in breast cancer (Supple- mentary Fig. 2c). Its expression was higher in tumor tissues than in paired normal tissues and increased along with genetic ampliﬁcation in TNBC (Fig. 1g, h). These results suggest that ENSA is ampliﬁed at the 1q21.3 region and is highly expressed in TNBC with clinical prognostic value. Copy number alterations (CNAs) refer to somatic changes in chromosome structure, typically submicroscopic DNA alterations between 1 Kbp and 1 Mbp in length, characterized by either depletions or ampliﬁcations of DNA segments. Copy number ampliﬁcation of ENSA promotes the progression of triple-negative breast cancer via cholesterol biosynthesis Yi-Yu Chen1,2, Jing-Yu Ge1, Si-Yuan Zhu1, Zhi-Ming Shao1 & Ke-Da Yu 1✉ Copy number alterations (CNAs) are pivotal genetic events in triple-negative breast cancer (TNBC). Here, our integrated copy number and transcriptome analysis of 302 TNBC patients reveals that gene alpha-endosulﬁne (ENSA) exhibits recurrent ampliﬁcation at the 1q21.3 region and is highly expressed in TNBC. ENSA promotes tumor growth and indicates poor patient survival in TNBC. Mechanistically, we identify ENSA as an essential regulator of cholesterol biosynthesis in TNBC that upregulates the expression of sterol regulatory element-binding transcription factor 2 (SREBP2), a pivotal transcription factor in cholesterol biosynthesis. We conﬁrm that ENSA can increase the level of p-STAT3 (Tyr705) and acti- vated STAT3 binds to the promoter of SREBP2 to promote its transcription. Furthermore, we reveal the efﬁcacy of STAT3 inhibitor Stattic in TNBC with high ENSA expression. In con- clusion, the ampliﬁcation of ENSA at the 1q21.3 region promotes TNBC progression and indicates sensitivity to STAT3 inhibitors. 1 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z T riple-negative breast cancer (TNBC) accounts for approximately 10–20% of all breast cancer cases. It is characterized by negative estrogen receptor (ER) and progestogen receptor (PR) expression and the lack of over- expression of HER2 (also deﬁned by lack of ERBB2 ampliﬁcation)1. Compared to other forms of breast cancer, TNBCs exhibit more aggressive clinical characteristics, including younger age of onset, larger tumor size, higher tumor grade, and more signiﬁcant metastasis potential2,3. Patients with TNBC have an increased risk of distant recurrence and death within 5 years of diagnosis, and the peak of distant recurrence occurs at ~3 years2,4. Although chemotherapy remains the standard of care for TNBC, a subset of patients shows limited responsiveness and develops advanced diseases due to a lack of effective targeted therapy and predictive markers in this heterogeneous disease5,6. To solve this problem, investigators have made great efforts to elucidate the molecular nature of TNBC and seek options for molecularly targeted therapy7–9. ENSA expression. These ﬁndings uncovered the mechanism by which high ENSA ampliﬁcation might promote TNBC progres- sion and suggested potential therapeutic targets. T NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z 1q21.3 amp 1q21.3 alteration frequency % 1.00 0.75 0.50 0.25 0.00 1.00 0.75 0.50 0.25 0.00 Disease-free survival 0 50 100 150 200 Disease-specific survival P = 0.0088 P = 0.013 0 50 100 200 9.6 15.8 20.9 27.7 9.7 8.9 16.4 32.6 150 b c d e f Months Months h g FUSCC TNBC TCGA TNBC ENSA mRNA expression Log2(RSEM+1) 11 12 13 14 15 FUSCC TNBC 9 P < 0.0001 ENSA mRNA expression log2(FPKM+1) 6 7 8 TCGA-BRCA 1q21.3 not amp N=93 N=836 N=106 N=944 0 50 100 Down Norm Up Amp 0 % TCGA-BRCA 58 302 99 RNA-seq N=360 OncoScan N=401 Higher expression in amplified group (Pearson r>0.3, p<0.05) n = 103 Higher expression in tumor (p<0.05) n = 41 41 genes FUSCC Amplified genes n = 514 Rank by Amp and Gain frequency Amp N= Gain N= 56 46 48 22 41 29 35 …… 174 169 145 152 123 133 113 …… 1q21.3 1q43 10p15.1 2p25.3 3q26.31 6p22.3 13q34 …… a 200 200 0.0 0.2 0.4 0.6 0.8 1.0 Relapse- la viv r u s e e rf HR=1.9 (1.1-1.328) P =0.019 Expression — Low ENSA — High ENSA TNBC 0 50 100 150 0 50 100 150 Expression — Low ENSA — High ENSA HR=1.48 (1.07-2.05) P =0.016 Basal-like Months Months 0.0 0.2 0.4 0.6 0.8 1.0 50 100 150 200 Months 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.59 (0.38−0.92) P = 0.018 TNBC Relapse-free survival 50 100 150 200 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.84 (0.68−1.06) P = 0.14 Basal-like Months Amplification (11%) No amplification 1q21.3 TCGA-BRCA 6 7 8 9 n ois s e r p x e A N R m A S N log2(FPKM+1) P = 0.0459 P = 0.0017 P = 0.0068 P < 0.0001 P < 0.0001 P < 0.0001 P = 0.0031 P < 0.0001 P < 0.0001 1q21.3 amp 1.00 0.75 0.50 0.25 0.00 1.00 0.75 0.50 0.25 0.00 Disease-free survival 0 50 100 150 200 Disease-specific survival P = 0.0088 P = 0.013 0 50 100 200 150 b c Months Months TCGA-BRCA 1q21.3 not amp N=93 N=836 N=106 N=944 Amplification (11%) No amplification 1q21.3 TCGA-BRCA 58 302 99 RNA-seq N=360 OncoScan N=401 Higher expression in amplified group (Pearson r>0.3, p<0.05) n = 103 Higher expression in tumor (p<0.05) n = 41 41 genes FUSCC Amplified genes n = 514 Rank by Amp and Gain frequency Amp N= Gain N= 56 46 48 22 41 29 35 …… 174 169 145 152 123 133 113 …… 1q21.3 1q43 10p15.1 2p25.3 3q26.31 6p22.3 13q34 …… a b a c MDA-MB-231 ENSA knockdown cells and control cells. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Gene set enrichment analysis (GSEA) revealed that the cholesterol path- way was the most enriched downregulated pathway in ENSA- silenced cells (Fig. 3a, b). GO analysis also indicated that the cholesterol biosynthesis pathway was strongly suppressed by ENSA knockdown (Supplementary Fig. 4a). In addition, GSEA revealed an upregulated apoptosis pathway in ENSA-silence cells, in accordance with the proapoptotic phenotype induced b ENSA depletion (Fig. 3c). We further explored the correlatio between ENSA expression and cholesterol biosynthesis pathwa activity in the FUSCC TNBC cohort using gene set variatio analysis. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z We found a signiﬁcant positive association betwee 1q21.3 alteration frequency % 0.75 0.50 0.25 0.00 0.75 0.50 0.25 0.00 Disease-free surv 0 50 100 150 200 Disease-specific su P = 0.0088 P = 0.013 0 50 100 200 9.6 15.8 20.9 27.7 9.7 8.9 16.4 32.6 150 d e f Months Months h g FUSCC TNBC TCGA TNBC ENSA mRNA expression Log2(RSEM+1) Norm Gain Amp N= 29 84 40 10 11 12 13 14 15 FUSCC TNBC 9 P < 0.0001 ENSA mRNA expression log2(FPKM+1) Normal Tumor 5 6 7 8 N= 88 88 N=93 N=836 N=106 N=944 0 50 100 Down Norm Up Amp 0 % TCGA-BRCA Higher expression in amplified group (Pearson r>0.3, p<0.05) n = 103 Higher expression in tumor (p<0.05) n = 41 41 genes Rank by Amp and Gain frequency Amp N= Gain N= 56 46 48 22 41 29 35 …… 174 169 145 152 123 133 113 …… 1q21.3 1q43 10p15.1 2p25.3 3q26.31 6p22.3 13q34 …… 200 200 0.0 0.2 0.4 0.6 0.8 1.0 Relapse- la viv r u s e e rf HR=1.9 (1.1-1.328) P =0.019 Expression — Low ENSA — High ENSA TNBC 0 50 100 150 0 50 100 150 Expression — Low ENSA — High ENSA HR=1.48 (1.07-2.05) P =0.016 Basal-like Months Months 0.0 0.2 0.4 0.6 0.8 1.0 50 100 150 200 Months 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.59 (0.38−0.92) P = 0.018 TNBC Relapse-free survival 50 100 150 200 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.84 (0.68−1.06) P = 0.14 Basal-like Months Down Norm Gain Amp 5 6 7 8 9 n ois s e r p x e A N R m A S N E log2(FPKM+1) N= 9 64 173 56 P = 0.0459 P = 0.0017 P = 0.0068 P < 0.0001 P < 0.0001 P < 0.0001 P = 0.0031 P < 0.0001 P < 0.0001 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 1q21.3 alteration frequency % 9.6 15.8 20.9 27.7 9.7 8.9 16.4 32.6 d 0 50 100 Down Norm Up Amp 0 % TCGA-BRCA d 1q21.3 alteration fre e f h g FUSCC TNBC TCGA TNBC ENSA mRNA expression Log2(RSEM+1) 11 12 13 14 15 FUSCC TNBC 9 P < 0.0001 ENSA mRNA expression log2(FPKM+1) 6 7 8 0 50 Down Norm Up Rank by Amp and Gain frequency p N= N= 56 46 48 22 41 29 35 …… 174 169 145 152 123 133 113 …… 1q21.3 1q43 10p15.1 2p25.3 3q26.31 6p22.3 13q34 …… 200 200 0.0 0.2 0.4 0.6 0.8 1.0 Relapse- la viv r u s e e rf HR=1.9 (1.1-1.328) P =0.019 Expression — Low ENSA — High ENSA TNBC 0 50 100 150 0 50 100 150 Expression — Low ENSA — High ENSA HR=1.48 (1.07-2.05) P =0.016 Basal-like Months Months 0.0 0.2 0.4 0.6 0.8 1.0 50 100 150 200 Months 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.59 (0.38−0.92) P = 0.018 TNBC Relapse-free survival 50 100 150 200 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.84 (0.68−1.06) P = 0.14 Basal-like Months 6 7 8 9 n ois s e r p x e A N R m A S N log2(FPKM+1) P = 0.0459 P = 0.0017 P = 0.0068 P < 0.0001 P < 0.0001 P < 0.0001 P = 0.0031 P < 0.0001 P < 0.0001 e f 200 200 0.0 0.2 0.4 0.6 0.8 1.0 Relapse- la viv r u s e e rf HR=1.9 (1.1-1.328) P =0.019 Expression — Low ENSA — High ENSA TNBC 0 50 100 150 0 50 100 150 Expression — Low ENSA — High ENSA HR=1.48 (1.07-2.05) P =0.016 Basal-like Months Months 0.0 0.2 0.4 0.6 0.8 1.0 50 100 150 200 Months 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.59 (0.38−0.92) P = 0.018 TNBC Relapse-free survival 50 100 1 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.84 (0.68− P = 0.14 Basal-like Months e f 200 200 0.0 0.2 0.4 0.6 0.8 1.0 Relapse- la viv r u s e e rf HR=1.9 (1.1-1.328) P =0.019 Expression — Low ENSA — High ENSA TNBC 0 50 100 150 0 50 100 150 Expression — Low ENSA — High ENSA HR=1.48 (1.07-2.05) P =0.016 Basal-like Months Months f f 00 0.0 0.2 0.4 0.6 0.8 1.0 50 100 150 200 Months 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.59 (0.38−0.92) P = 0.018 TNBC Relapse-free survival 50 100 150 200 0 Expression — Low GOLPH3L — High GOLPH3L HR = 0.84 (0.68−1.06) P = 0.14 Basal-like Months e h FUSCC TNBC 9 P < 0.0001 ENSA mRNA expression log2(FPKM+1) Normal Tumor 5 6 7 8 N= 88 88 TCGA TNBC ENSA mRNA expression Log2(RSEM+1) Norm Gain Amp N= 29 84 40 10 11 12 13 14 15 P = 0.0031 P < 0.0001 P < 0.0001 g FUSCC TNBC Down Norm Gain Amp 5 6 7 8 9 n ois s e r p x e A N R m A S N E log2(FPKM+1) N= 9 64 173 56 P = 0.0459 P = 0.0017 P = 0.0068 P < 0.0001 P < 0.0001 P < 0.0001 h h g revealed an upregulated apoptosis pathway in ENSA-silenced cells, in accordance with the proapoptotic phenotype induced by ENSA depletion (Fig. Results 2d, e, and Supplementary Fig. 3c). ENSA knockdown also caused pronounced apoptosis among TNBC cells but had almost no effect on cell cycle progression (Fig. 2f and Supplementary Fig. 3d, e). These results suggest the critical role of ENSA in promoting the growth of TNBC cells. In the current study, we used integrated copy number and transcriptome analyses to discover CNA-affected oncogenes in TNBC. ENSA (alpha-endosulﬁne) was found to be ampliﬁed at the 1q21.3 locus in more than 18% of patients, and its ampliﬁ- cation was correlated with increased expression in TNBC. ENSA could promote tumor growth by promoting the cholesterol bio- synthesis program in TNBC. At the molecular level, the tran- scriptional activation of sterol regulatory element-binding transcription factor 2 (SREBP2) by phosphorylated STAT3 (p- STAT3) played a critical role in ENSA-induced cholesterol metabolism dysregulation. Consequently, hindering STAT3 phosphorylation resulted in tumor inhibition in TNBC with high ENSA plays a crucial role in cholesterol biosynthesis in TNBC. To explore the underlying molecular mechanisms of ENSA in TNBC cells, we performed RNA-sequencing analysis of both NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 2 ARTICLE NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 22) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Decreased expression of cholesterol biosynthesis enzymes and SREBP2, a crucial transcription factor that preferentially activates genes involved in cholesterol biosynthesis, was further validated at the mRNA and protein levels in TNBC cells expressing ENSA shRNA (Fig. 3e, f). Importantly, the protein levels of the full and cleaved forms of SREBP2 were both decreased, which was consistent with the decreased mRNA level of SREBP2 upon ENSA silencing (Fig. 3f). Overexpression of ENSA in ENSA knockdown TNBC cells restored the expression of cholesterol biosynthesis genes (Supplementary Fig. 4c). Intriguingly, the expression of SREBP2 and downstream enzymes was not altered by ENSA knockdown in non-TNBC cell lines, suggesting a critical role of the ENSA-regulated cholesterol pathway in TNBC (Supplemen- tary Fig. 4d). To determine whether the main products of the cholesterol biosynthesis pathway had changed, we detected the contents of cholesterol and intermediate metabolites by liquid chromatography-mass spectrometry. Consistent with the down- regulated mRNA and protein expression of genes involved in cholesterol biosynthesis, the concentration of total cholesterol and most intermediates decreased upon ENSA silencing in TNBC cells (Fig. 3g, h). In addition, ENSA knockdown caused a sig- niﬁcant decrease in free cholesterol content in TNBC cells (Fig. 3i). We next examined whether the ENSA knockdown- induced phenotype was attributable to cholesterol depletion. The addition of cholesterol (2.5 μg/ml) partially attenuated the inhi- bition of cell growth and increased apoptosis in ENSA-depleted TNBC cells (Fig. 3j and Supplementary Fig. 4e, f). Together, these ﬁndings support the notion that ENSA might promote the growth of TNBC cells by regulating cellular cholesterol biosynthesis. ENSA activates STAT3 to regulate cholesterol biosynthesis. We next questioned whether activated STAT3 participated in ENSA- induced alterations in the cholesterol biosynthesis pathway. The prediction from the JASPAR database showed that STAT3 was likely to bind to the promoter of SREBP2 (Fig. 5a). ChIP-seq data analysis from GSE152203 also showed a STAT3-binding peak in SREBP2 in MDA-MB-231 cells (Supplementary Fig. 6a). Using a ChIP assay, we conﬁrmed that STAT3 occupied a site in the promoter region of SREBP2 (Fig. 5b). Moreover, STAT3 silencing in TNBC cells suppressed the mRNA and protein expression of SREBP2 (Fig. 5c and Supplementary Fig. 6b). These results indicate that STAT3 could bind to the promoter of SREBP2 and alter the expression of SREBP2 at the transcriptional level in TNBC cells. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Fig. 1 ENSA is ampliﬁed at the 1.21.3 region, is highly expressed and predicts poor survival in TNBC. a Schematic diagram depicting the screening for copy number alteration-affected genes in TNBC. b Copy number alteration proﬁle of the 1q21.3 region in the TCGA breast cancer cohort. c Kaplan–Meier survival analysis of 1q21.3 copy number alterations in TCGA breast cancer patients. Log-rank test. d 1q21.3 alteration frequency in the TCGA cohort with different breast cancer subtypes. e, f Kaplan–Meier plots of ENSA and GOLPH3L in TNBC or basal-like BC (https://kmplot.com/analysis/). Log-rank test. g ENSA expression of samples with different ENSA copy number in FUSCC and TCGA TNBC cohorts. n = 302 in FUSCC TNBC cohort and n = 153 in TCGA TNBC cohort. Data are presented as mean ± SD. Two-tailed one-way ANOVA tests and adjustments were made for multiple comparisons. h ENSA expression of 88 paired tumor tissues versus adjacent normal tissues in FUSCC cohort. n = 88 paired samples. Data are presented as mean ± SD. Two- tailed paired Student’s t test. Source data are provided as a Source Data ﬁle. FUSCC Fudan University Shanghai Cancer Center, TCGA The Cancer Genome Atlas, TNBC triple-negative breast cancer, BRCA breast cancer, FPKM fragments per kilobase million, RSEM RNA-seq by expectation maximization, Amp ampliﬁcation, Norm normal. overexpression of ENSA and mostly rescued by overexpression of STAT3 (Fig. 4f, g, and Supplementary Fig. 5a). Immunohis- tochemical (IHC) staining of p-STAT3 (Tyr705) signiﬁcantly decreased upon ENSA silencing. However, in ENSA-depleted TNBC tumors, overexpression of either ENSA or STAT3 increased p-STAT3 (Tyr705) staining (Fig. 4h and Supplementary Fig. 5b). In accordance with the proapoptotic phenotype induced by ENSA depletion in vitro, IHC staining of cleaved caspase 3 displayed a corresponding change in each group of xenografts (Fig. 4h and Supplementary Fig. 5c). In addition, we also observed decreased lung metastasis incidence in the ENSA-depleted group (Supplementary Fig. 5d). Taken together, these data suggest the critical role of ENSA-STAT3 signaling in promoting the growth of triple-negative tumors. ENSA mRNA expression and the cholesterol biosynthesis pro- gram in TNBC (Fig. 3d), which was further validated in The Cancer Genome Atlas (TCGA), Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) and Korean breast cancer (SMC) cohorts (Supplementary Fig. 4b). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z 3c). We further explored the correlation between ENSA expression and cholesterol biosynthesis pathway activity in the FUSCC TNBC cohort using gene set variation analysis. We found a signiﬁcant positive association between MDA-MB-231 ENSA knockdown cells and control cells. Gene set enrichment analysis (GSEA) revealed that the cholesterol path- way was the most enriched downregulated pathway in ENSA- silenced cells (Fig. 3a, b). GO analysis also indicated that the cholesterol biosynthesis pathway was strongly suppressed by ENSA knockdown (Supplementary Fig. 4a). In addition, GSEA 3 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z 15 a b MDA-MB-231 BT549 shENSA - #1 #2 - #1 #2 ENSA TUBA1B 60 kDa MDA-MB-231 Relative growth rate BT549 1 2 3 4 5 6 0 2 4 6 8 shCtrl shENSA#1 shENSA#2 1 2 3 4 5 6 shCtrl shENSA#1 shENSA#2 0 2 4 6 8 10 P <0.0001 P <0.0001 P <0.0001 P <0.0001 b Days Relative growth rate BT549 1 2 3 4 5 6 0 2 4 6 8 shCtrl shENSA#1 shENSA#2 P <0.0001 P <0.0001 b Days Days MDA-MB-231 Relative growth rate BT549 1 2 3 4 5 6 0 2 4 6 8 shCtrl shENSA#1 shENSA#2 1 2 3 4 5 6 shCtrl shENSA#1 shENSA#2 0 2 4 6 8 10 P <0.0001 P <0.0001 P <0.0001 P <0.0001 15 a MDA-MB-231 BT549 shENSA - #1 #2 - #1 #2 ENSA TUBA1B 60 kDa Days MDA-MB-231 1 2 3 4 5 6 shCtrl shENSA#1 shENSA#2 0 2 4 6 8 10 P <0.0001 P <0.0001 b a a shENS Colony number P < 0.0001 P < 0.0001 0 50 100 150 200 - #1 #2 shENSA c BT549 - #1 #2 shENSA MDA-MB-231 - #1 #2 SA Colony number - #1 #2 shENSA P = 0.0004 P = 0.0006 0 100 200 300 Colony number - #1 #2 shENSA P = 0.0004 P = 0.0006 0 100 200 300 MDA-MB-231 - #1 #2 SA c d - + shENSA Vector ENSA BT549 d Colony number shENSA ENSA - - + + - + - + P < 0.0001 P < 0.0001 0 100 200 300 shENSA ENSA Colony number - - + + - + - + 0 50 100 150 200 P < 0.0001 P = 0.0002 MDA-MB-231 - + shENSA Vector ENSA e f MDA-MB-231 BT549 Percentage of Annexin V+ cells (%) - + shENSA - + 0 5 10 15 20 P = 0.0004 0 5 10 15 P < 0.0001 MDA-MB-231 BT549 e t a r h t w o r g e vit ale R 1 2 3 4 5 6 0 2 4 6 8 10 Days shCtrl+Vec shCtrl+ENSA shENSA+Vec shENSA+ENSA P < 0.0001 P < 0.0001 1 2 3 4 5 6 0 2 4 6 8 10 Days shCtrl+Vec shCtrl+ENSA shENSA+Vec shENSA+ENSA P < 0.0001 P < 0.0001 MDA-MB-231 BT549 Percentage of Annexin V+ cells (%) - + shENSA - + 0 5 10 15 20 P = 0.0004 0 5 10 15 P < 0.0001 MDA-MB-231 1 2 3 4 5 6 0 2 4 6 8 10 Days shCtrl+Vec shCtrl+ENSA shENSA+Vec shENSA+ENSA P < 0.0001 P < 0.0001 e BT549 e t a r h t w o r g e vit ale R 1 2 3 4 5 6 0 2 4 6 8 10 Days shCtrl+Vec shCtrl+ENSA shENSA+Vec shENSA+ENSA P < 0.0001 P < 0.0001 e Fig. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z We next validated that the promoter activity of SREBP2 was restrained by ENSA depletion and rescued by STAT3 overexpression based on the luciferase reporter assay (Fig. 5d). Consistently, the expression level of SREBP2 and enzymes involved in cholesterol biosynthesis decreased after ENSA knockdown, but this effect was reversed by the over- expression of STAT3 (Fig. 5e). In addition, the cellular free cholesterol levels, which were decreased with ENSA depletion, could also be restored by overexpression of STAT3 (Fig. 5f and Supplementary Fig. 6c). In mouse mammary fat pad xenograft models, the expression of SREBP2 and FDPS decreased after ENSA knockdown and was restored when exogenous STAT3 was overexpressed (Supplementary Fig. 6d). Activated STAT3 is involved in ENSA-induced tumor growth. To elucidate the molecular mechanism of the ENSA depletion- induced phenotype in TNBC cells, we performed transcription factor (TF) analysis of the RNA-seq data and hypothesized that STAT3 was a key TF responsible for the downstream alteration (Fig. 4a). In TNBC cells, p-STAT3 (Tyr705), but not p-STAT3 (Ser727) or total STAT3, was markedly decreased upon ENSA knockdown (Fig. 4b). Overexpression of ENSA restored the level of p-STAT3 (Tyr705) (Fig. 4c). Next, we conﬁrmed the involve- ment of STAT3 in ENSA knockdown-induced growth inhibition by overexpressing STAT3 in ENSA-silenced TNBC cells to induce the constitutive activation/phosphorylation of STAT3. Our results showed that the overexpression of STAT3 could partially attenuate TNBC cell growth inhibition induced by ENSA deple- tion (Fig. 4d, e). p pp y g We next explored whether the inhibitory effect of ENSA knockdown on p-STAT3 was dependent on protein phosphatase 2A (PP2A), a known protein phosphatase whose function is inhibited by direct interaction with ENSA. We silenced PP2A using its catalytic subunit-targeting siRNA (siPPP2CA) in MDA- MB-231 cells and observed an increase in p-STAT3 (Tyr705) expression (Fig. 5g). However, p-STAT3 (Ser727) was not inﬂuenced by PP2A silencing. Furthermore, the inhibition of p-STAT3 (Tyr705) and SREBP2 expression by ENSA depletion was weakened by an additional knockdown of PP2A in TNBC cells (Fig. 5h). These results suggest that ENSA might promote In mouse mammary fat pad xenografts, ENSA knockdown resulted in a signiﬁcant reduction in the tumor growth of MDA- MB-231 cells, which could be completely rescued by NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 4 the phosphorylation of STAT3 to regulate cholesterol biosynth- esis in TNBC cells in a PP2A-dependent manner inhibitor of STAT3 activation. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Compared to the ENSA knock- down group the control group with higher ENSA levels was more 15 a b c MDA-MB-231 BT549 shENSA - #1 #2 - #1 #2 ENSA TUBA1B Days Days 60 e f shENSA ENSA Colony number Colony number d kDa - + shENSA Vector ENSA BT549 MDA-MB-231 - + shENSA Vector ENSA MDA-MB-231 - #1 #2 shENSA Colony number BT549 - #1 #2 shENSA Colony number - - + + - + - + shENSA ENSA - - + + - + - + P < 0.0001 P < 0.0001 MDA-MB-231 BT549 Percentage of Annexin V+ cells (%) - + shENSA - + 0 5 10 15 20 P = 0.0004 0 5 10 15 P < 0.0001 P < 0.0001 P < 0.0001 0 50 100 150 200 - #1 #2 shENSA - #1 #2 shENSA P = 0.0004 P = 0.0006 0 50 100 150 200 P < 0.0001 P = 0.0002 0 100 200 300 MDA-MB-231 Relative growth rate BT549 1 2 3 4 5 6 0 2 4 6 8 shCtrl shENSA#1 shENSA#2 1 2 3 4 5 6 shCtrl shENSA#1 shENSA#2 0 2 4 6 8 10 P <0.0001 P <0.0001 MDA-MB-231 BT549 0 100 200 300 e t a r h t w o r g e vit ale R 1 2 3 4 5 6 0 2 4 6 8 10 Days shCtrl+Vec shCtrl+ENSA shENSA+Vec shENSA+ENSA P < 0.0001 P < 0.0001 1 2 3 4 5 6 0 2 4 6 8 10 Days shCtrl+Vec shCtrl+ENSA shENSA+Vec shENSA+ENSA P < 0.0001 P < 0.0001 P <0.0001 P <0.0001 Fig. 2 ENSA is a major driver of TNBC cell growth. a Stable silencing of ENSA expression in the TNBC cell lines BT549 and MDA-MB-231. b In vitro growth curves of BT549 and MDA-MB-231 cells expressing control or ENSA shRNA. n = 6. Data are presented as mean ± SD. Two-tailed two-way ANOVA tests. c Colony formation of BT549 and MDA-MB-231 cells expressing control or ENSA shRNA. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. d Colony formation of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± ENSA overexpression. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. e In vitro growth curves of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± ENSA overexpression. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z n = 6. Data are presented as mean ± SD. Two-tailed two-way ANOVA tests. f Apoptosis levels were measured in BT549 and MDA-MB- 231 cells expressing control or ENSA shRNA. Percentage of annexin V+ cells are shown. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. Source data are provided as a Source Data ﬁle. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z ARTICLE NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z ARTICLE f P < 0.001 q = 0.014 Cholesterol homeostasis Apoptosis P = 0.012 q = 0.077 b c a GSEA hallmark analysis -2.0 TNFA_SIGNALING_VIA_NFKB OXIDATIVE_PHOSPHORYLATION ADIPOGENESIS MTORC1_SIGNALING UV_RESPONSE_UP BILE_ACID_METABOLISM ALLOGRAFT_REJECTION ESTROGEN_RESPONSE_LATE INTERFERON_GAMMA_RESPONSE KRAS_SIGNALING_DN INFLAMMATORY_RESPONSE MYC_TARGETS_V2 CHOLESTEROL_HOMEOSTASIS -1.0 0.0 NES score Down in shENSA P<0.05, FDR q<0.25 c Cholesterol biosynthesis d P = 1.18e-19 cor = 0.453 -0.8 -0.4 0.0 0.4 0.8 6 7 8 ENSA expression FUSCC TNBC b d SA expression level (Fig. 6c–e and Supplementary Fig. 7c). The icopathological features of these three organoids are shown in plementary Table 1. We next explored the sensitivity of BC cells to Stattic in xenograft models and found signiﬁcantly aired sensitivity to Stattic when ENSA was depleted (Fig. 6f, g, Supplementary Fig. 7d, e). To further explore the drug response in patients with TNBC, we constructed seven patient-derived xenograft (mini-PDX) models, as r previously24,25, and measured the response for Stattic nor to vehicle treatment (Fig. 6h). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z 2 ENSA is a major driver of TNBC cell growth. a Stable silencing of ENSA expression in the TNBC cell lines BT549 and MDA-MB-231. b In vitro growth curves of BT549 and MDA-MB-231 cells expressing control or ENSA shRNA. n = 6. Data are presented as mean ± SD. Two-tailed two-way ANOVA tests. c Colony formation of BT549 and MDA-MB-231 cells expressing control or ENSA shRNA. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. d Colony formation of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± ENSA overexpression. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. e In vitro growth curves of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± ENSA overexpression. n = 6. Data are presented as mean ± SD. Two-tailed two-way ANOVA tests. f Apoptosis levels were measured in BT549 and MDA-MB- 231 cells expressing control or ENSA shRNA. Percentage of annexin V+ cells are shown. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. Source data are provided as a Source Data ﬁle. the phosphorylation of STAT3 to regulate cholesterol biosynth- esis in TNBC cells in a PP2A-dependent manner. inhibitor of STAT3 activation. Compared to the ENSA knock- down group, the control group with higher ENSA levels was more sensitive to Stattic and exhibited signiﬁcantly reduced growth after Stattic treatment (Fig. 6a, b, and Supplementary Fig. 7a, b). In organoids derived from three TNBC patients with diverse levels of ENSA expression, we also found that the sensitivity of organoids to Stattic increased along with the increase in the ENSA determines therapeutic sensitivity to STAT3 inhibitors. As ENSA regulates the activity of STAT3, we sought to evaluate whether ENSA could be a therapeutic marker of STAT3 inhibitor sensitivity in TNBC cells. We used Stattic as a small molecule 5 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z The clinicopathological fea these TNBC models are shown in Supplementary Table 2 with the results obtained above, ENSA-high tumors (57 Cholesterol biosynthesis a GSEA hallmark analysis -2.0 TNFA_SIGNALING_VIA_NFKB OXIDATIVE_PHOSPHORYLATION ADIPOGENESIS MTORC1_SIGNALING UV_RESPONSE_UP BILE_ACID_METABOLISM ALLOGRAFT_REJECTION ESTROGEN_RESPONSE_LATE INTERFERON_GAMMA_RESPONSE KRAS_SIGNALING_DN INFLAMMATORY_RESPONSE MYC_TARGETS_V2 CHOLESTEROL_HOMEOSTASIS -1.0 0.0 NES score Down in shENSA P<0.05, FDR q<0.25 j g f 2.5 3.0 3.5 4.0 4.5 Cholesterol h i P < 0.001 q = 0.014 Cholesterol homeostasis Apoptosis P = 0.012 q = 0.077 b d shENSA Colony number Cholesterol c e shCtrl shENSA#1 shENSA#2 140 60 45 45 100 60 75 45 15 60 MDA-MB-231 - #1 #2 45 TUBA1B ENSA FDPS FDFT1 HMGCR HMGCS1 LSS MVK pSREBP2 nSREBP2 ACAT2 BT549 shENSA - #1 #2 kDa - + shENSA - + Cholesterol BT549 MDA-MB-231 - + shENSA - + Cholesterol shCtrl shENSA MDA-MB-231 75μm 50μm BT549 Free Cholesterol Nucleus P = 0.029 0 50 100 150 200 - - - - P<0.0001 P=0.0007 0 50 100 150 200 250 shENSA Colony number Cholesterol - - + + - + - + P=0.0001 P=0.0003 P = 1.18e-19 cor = 0.453 -0.8 -0.4 0.0 0.4 0.8 6 7 8 ENSA expression FUSCC TNBC Lanosterol Dihydrolanosterol Dihydro-T-MAS Zymosterol Scaled concentration 1.5 0 -1.5 Cholesterol Squalene Lathosterol FF-MAS T-MAS Dehydrolathosterol Desmosterol Total sterols Concentration (108μmol/cell) P=0.0005 P<0.0001 P<0.0001 P=0.0081 P<0.0001 P<0.0001 P=0.0007 P=0.0007 P<0.0001 P<0.0001 P=0.0060 P=0.0028 P=0.0001 P=0.0004 P=0.0019 P=0.0206 P=0.0006 P=0.0005 P=0.0007 P=0.0009 P=0.0013 P=0.0003 P=0.0489 P=0.0036 P<0.0001 P<0.0001 P<0.0001 P<0.0001 P=0.0018 P=0.0018 P<0.0001 P=0.0012 P=0.0005 P<0.0001 P=0.0110 P<0.0001 P<0.0001 P<0.0001 P=0.0004 n.s. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z P=0.0038 P=0.0084 P=0.0002 P=0.0071 P=0.0110 P=0.0009 P=0.0150 P<0.0001 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 n ois s e r p x e A N R m e vit ale R BT549 MDA-MB-231 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications Cholesterol biosynthesis a GSEA hallmark analysis -2.0 TNFA_SIGNALING_VIA_NFKB OXIDATIVE_PHOSPHORYLATION ADIPOGENESIS MTORC1_SIGNALING UV_RESPONSE_UP BILE_ACID_METABOLISM ALLOGRAFT_REJECTION ESTROGEN_RESPONSE_LATE INTERFERON_GAMMA_RESPONSE KRAS_SIGNALING_DN INFLAMMATORY_RESPONSE MYC_TARGETS_V2 CHOLESTEROL_HOMEOSTASIS -1.0 0.0 NES score Down in shENSA P<0.05, FDR q<0.25 f P < 0.001 q = 0.014 Cholesterol homeostasis Apoptosis P = 0.012 q = 0.077 b d c e shCtrl shENSA#1 shENSA#2 140 60 45 45 100 60 75 45 15 60 MDA-MB-231 - #1 #2 45 TUBA1B ENSA FDPS FDFT1 HMGCR HMGCS1 LSS MVK pSREBP2 nSREBP2 ACAT2 BT549 shENSA - #1 #2 kDa P = 1.18e-19 cor = 0.453 -0.8 -0.4 0.0 0.4 0.8 6 7 8 ENSA expression FUSCC TNBC P=0.0005 P<0.0001 P<0.0001 P=0.0081 P<0.0001 P<0.0001 P=0.0007 P=0.0007 P<0.0001 P<0.0001 P=0.0060 P=0.0028 P=0.0001 P=0.0004 P=0.0019 P=0.0206 P=0.0006 P=0.0005 P=0.0007 P=0.0009 P=0.0013 P=0.0003 P=0.0489 P=0.0036 P<0.0001 P<0.0001 P<0.0001 P<0.0001 P=0.0018 P=0.0018 P<0.0001 P=0.0012 P=0.0005 P<0.0001 P=0.0110 P<0.0001 P<0.0001 P<0.0001 P=0.0004 n.s. P=0.0038 P=0.0084 P=0.0002 P=0.0071 P=0.0110 P=0.0009 P=0.0150 P<0.0001 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 n ois s e r p x e A N R m e vit ale R BT549 MDA-MB-231 a Apoptosis f e shCtrl shENSA#1 shENSA#2 140 60 45 45 100 60 75 45 15 60 MDA-MB-231 - #1 #2 45 TUBA1B ENSA FDPS FDFT1 HMGCR HMGCS1 LSS MVK pSREBP2 nSREBP2 ACAT2 BT549 shENSA - #1 #2 kDa P=0.0005 P<0.0001 P<0.0001 P=0.0081 P<0.0001 P<0.0001 P=0.0007 P=0.0007 P<0.0001 P<0.0001 P=0.0060 P=0.0028 P=0.0001 P=0.0004 P=0.0019 P=0.0206 P=0.0006 P=0.0005 P=0.0007 P=0.0009 P=0.0013 P=0.0003 P=0.0489 P=0.0036 P<0.0001 P<0.0001 P<0.0001 P<0.0001 P=0.0018 P=0.0018 P<0.0001 P=0.0012 P=0.0005 P<0.0001 P=0.0110 P<0.0001 P<0.0001 P<0.0001 P=0.0004 n.s. P=0.0038 P=0.0084 P=0.0002 P=0.0071 P=0.0110 P=0.0009 P=0.0150 P<0.0001 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 n ois s e r p x e A N R m e vit ale R BT549 MDA-MB-231 f e shCtrl shENSA#1 shENSA#2 TU F HM HM pSR nSR A sh P=0.0005 P<0.0001 P<0.0001 P=0.0081 P<0.0001 P<0.0001 P=0.0007 P=0.0007 P<0.0001 P<0.0001 P=0.0060 P=0.0028 P=0.0001 P=0.0004 P=0.0019 P=0.0206 P=0.0006 P=0.0005 P=0.0007 P=0.0009 P=0.0013 P=0.0003 P=0.0489 P=0.0036 P<0.0001 P<0.0001 P<0.0001 P<0.0001 P=0.0018 P=0.0018 P<0.0001 P=0.0012 P=0.0005 P<0.0001 P=0.0110 P<0.0001 P<0.0001 P<0.0001 P=0.0004 n.s. ARTICLE STAT3-mediated transcriptional activation of SREBP2 and downstream cholesterol biosynthesis (Fig. 7g). y ( g g) TNBC exhibits high genomic instability, resulting in frequent CNAs at the chromosome level. Identiﬁcation of the CNA-driven phenotype and the underlying mechanisms provides new insights into the pathogenesis of TNBC and facilitates the discovery of therapeutic treatments. The gain of 1q is one of the most frequent genomic imbalances in breast carcinomas and exhibits a higher CNA frequency in basal tumors than in luminal tumors26–28. Our previous research on the FUSCC cohort also indicated the widespread occurrence of 1q ampliﬁcation in Chinese TNBC patients7. Here, we further showed that 1q21.3 was the exact region that harbored the highest frequency of gains in the 1q chromosome band (>18% in FUSCC TNBC patients), and this alteration frequency was just below that of the 8q amplicon (where MYC is located). Based on the analysis of the TCGA breast cancer cohort, we found ampliﬁcation of the 1q21.3 region in all breast cancer subtypes, but the ampliﬁcation rate was higher in TNBC than in other subtypes, which suggests the signiﬁcance of this amplicon in TNBC. Evidence has indicated that CNAs promote the progression of tumors by altering the expression of genes within those affected genomic regions29. For example, 1q21.3 ampliﬁcation upregulates the expression of several encompassing genes that form a regulatory loop to drive tumor growth30. Herein, we utilized copy number and gene expression data of primary tumors, including 88 paired tumor tissues and peritumor tissues from the FUSCC TNBC cohort, to discover potential oncogenes within amplicons. Among the 41 candidates identiﬁed, ENSA was located in the frequently ampliﬁed 1q21.3 region and indicated a poor prognosis in TNBC. Correlations among ENSA, p-STAT3, and SREBP2 expression in clinical samples and patient outcomes. To investigate the clinical relevance of our ﬁndings, we ﬁrst evaluated the protein expression levels of ENSA in 8 paired primary TNBC specimens and adjacent normal tissues by IHC assay. The results revealed that the ENSA protein levels were much higher in TNBC speci- mens than in normal tissues (Fig. 7a, b). To investigate the cor- relation of ENSA protein expression with patient survival, we collected surgical samples from 138 TNBC patients and detected the expression level of ENSA by performing IHC analysis. Representative IHC images are shown in Fig. 7c. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Fig. 3 ENSA plays a crucial role in cholesterol biosynthesis in TNBC. a GSEA of downregulated pathways after ENSA knockdown in MDA-MB-231 cells. The top 10 pathways (P < 0.05 and FDR q < 0.25) ranked by absolute normalized enrichment scores are shown. NES score and nominal P-value were given by GSEA software. b Enrichment plot of the cholesterol homeostasis pathway after ENSA knockdown in MDA-MB-231 cells. c Enrichment plot of the apoptosis pathway after ENSA knockdown in MDA-MB-231 cells. d Scatter plot showing the correlation of ENSA expression with the cholesterol biosynthesis pathway score in FUSCC TNBC data identiﬁed by ‘gsva’ method. Correlation coefﬁcients were calculated using the Pearson test. Two-tailed P- values were given. e qRT-PCR analysis of the relative transcript levels of cholesterol biosynthesis pathway genes after ENSA knockdown in MDA-MB-231 cells. n = 4. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. f Western blotting images showing proteins involved in the cholesterol biosynthesis pathway after ENSA knockdown in MDA-MB-231 cells. n = 3 independent experiments. g Heatmap displaying the concentration of cholesterol and intermediates in the cholesterol biosynthesis pathway after ENSA knockdown in MDA-MB-231 cells. n = 4. h Total cellular cholesterol contents in MDA-MB-231 cells were analyzed by LC-MS with normalization to cell quantity. n = 4. The center line corresponded to the median, the lower and upper hinges corresponded to the ﬁrst and third quartiles, and the upper/lower whisker extends from the hinge to the largest/smallest value no further than 1.5 times interquartile range. Two-tailed unpaired Wilcoxon test. i Filipin III staining showing the cellular free cholesterol content in MDA-MB-231 and BT549 cells with ENSA knockdown. n = 3 independent experiments. j Colony formation of BT549 and MDA-MB-231 cells expressing control or ENSA shRNA after treatment with 2.5 μg/ml exogeneous cholesterol. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. Source data are provided as a Source Data ﬁle. GSEA gene set enrichment analysis, NES normalized enrichment score. 543, 553) showed higher sensitivity to Stattic than those with relatively low-ENSA expression (554, 584, 552) (Fig. 6i, j, and Supplementary Fig. 7f, g). Together, the results from our TNBC cell lines, organoids, animal models and mini-PDX models strongly suggest that ENSA expression can be a biomarker for effective treatment with STAT3 inhibitors. ARTICLE Kaplan–Meier analysis of specimens revealed that patients harboring tumors with high ENSA levels tended to have worse relapse-free survival and overall survival than patients harboring low-ENSA levels (Fig. 7d, P < 0.05). Multivariable analysis showed that ENSA expression still showed prognostic value for relapse-free survival (P = 0.04) and tended to correlate with worse overall survival (P = 0.07) after adjustment for age, tumor size, and lymph node status (Supplementary Tables 3 and 4). We further examined the protein expression levels of SREBP2, HMGCR, and p-STAT3 (Tyr705). Representative IHC images are shown in Fig. 7e and Supplementary Fig. 8a. The IHC results showed that ENSA expression was positively correlated with p-STAT3 (Tyr705), SREBP2, and HMGCR (Fig. 7f and Supplementary Fig. 8b). In addition, the survival analysis of SREBP2 expression showed that high SREBP2 expression was correlated with worse relapse-free survival of TNBC patients in internal and external cohorts (Supplementary Fig. 8c, d). The expression correlation between ENSA and SREBP2 was also validated in an external cohort (Supplementary Fig. 8e). Together, these results indicate that the expression of ENSA is positively correlated with that of down- stream molecules and is a poor prognostic factor in TNBC. g p p g The ENSA-encoding protein belongs to the highly conserved c- AMP-regulated phosphoprotein (ARPP) family and was initially identiﬁed as an endogenous ligand for the sulfonylurea receptor that modulates insulin secretion and glucose metabolism31,32. It has previously been shown that in Xenopus egg extracts, ENSA and its close relative ARPP19 are substrates of great wall (GWL) kinase and act as competitor inhibitors to prevent PP2A-B55 from dephosphorylating substrates such as cyclin B-CDK1, which results in mitotic entry33,34. A subsequent study on human cells identiﬁed the ability of ENSA to control the length of the S phase35. While GWL and PP2A have been heavily studied, the genomic alteration and molecular function of ENSA have rarely been reported in the context of cancer36. To test these ﬁndings, we generated ENSA knockdown TNBC cells and found that these cells exhibited pro- nounced growth inhibition. Surprisingly, altered expression of ENSA did not signiﬁcantly inﬂuence the cell cycle distribution of TNBC cells at the S phase or G2/M phase, which implies that the function of ENSA is context-dependent. Interestingly, cholesterol NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z P=0.0038 P=0.0084 P=0.0002 P=0.0071 P=0.0110 P=0.0009 P=0.0150 P<0.0001 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 n ois s e r p x e A N R m e vit ale R BT549 MDA-MB-231 f e shCtrl shENSA#1 shENSA#2 P=0.0005 P<0.0001 P<0.0001 P=0.0081 P<0.0001 P<0.0001 P=0.0007 P=0.0007 P<0.0001 P<0.0001 P=0.0060 P=0.0028 P=0.0001 P=0.0004 P=0.0019 P=0.0206 P=0.0006 P=0.0005 P=0.0007 P=0.0009 P=0.0013 P=0.0003 P=0.0489 P=0.0036 0.0 0.5 1.0 1.5 n ois s e r p x e A N BT549 MDA-MB-231 f P<0.0001 P<0.0001 P<0.0001 P<0.0001 P=0.0018 P=0.0018 P<0.0001 P=0.0012 P=0.0005 P<0.0001 P=0.0110 P<0.0001 P<0.0001 P<0.0001 P=0.0004 n.s. P=0.0038 P=0.0084 P=0.0002 P=0.0071 P=0.0110 P=0.0009 P=0.0150 P<0.0001 0.0 0.5 1.0 1.5 N R m e vit ale R 2.5 3.0 3.5 4.0 4.5 Cholesterol h P = 0.029 Concentration (108μmol/cell) g Lanosterol Dihydrolanosterol Dihydro-T-MAS Zymosterol Scaled concentration 1.5 0 -1.5 Cholesterol Squalene Lathosterol FF-MAS T-MAS Dehydrolathosterol Desmosterol Total sterols i shCtrl shENSA MDA-MB-231 75μm 50μm BT549 Free Cholesterol Nucleus i h i h g ree Cholestero Nucleus j - + shENSA - + Cholesterol BT549 MDA-MB-231 - + shENSA - + Cholesterol 0 50 100 150 200 250 shENSA Colony number Cholesterol - - + + - + - + P=0.0001 P=0.0003 j shENSA Colony number Cholesterol 0 50 100 150 200 - - - - P<0.0001 P=0.0007 ENSA expression level (Fig. 6c–e and Supplementary Fig. 7c). The clinicopathological features of these three organoids are shown in Supplementary Table 1. We next explored the sensitivity of TNBC cells to Stattic in xenograft models and found signiﬁcantly impaired sensitivity to Stattic when ENSA was depleted (Fig. 6f, g, and Supplementary Fig. 7d, e). To further explore the drug response in patients with TNBC, we constructed seven mini- patient-derived xenograft (mini-PDX) models, as reported previously24,25, and measured the response for Stattic normalized to vehicle treatment (Fig. 6h). The clinicopathological features of these TNBC models are shown in Supplementary Table 2. In line with the results obtained above, ENSA-high tumors (578, 573, response in patients with TNBC, we constructed seven mini- patient-derived xenograft (mini-PDX) models, as reported previously24,25, and measured the response for Stattic normalized to vehicle treatment (Fig. 6h). The clinicopathological features of these TNBC models are shown in Supplementary Table 2. In line with the results obtained above, ENSA-high tumors (578, 573, NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 6 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z shCtrl P<0.05 0.0 0.5 1.0 1.5 2.0 -NES Days Days Relative growth rate 1 2 3 4 5 6 0 2 4 6 8 BT549 P<0.0001 P<0.0001 1 2 3 4 5 6 0 2 4 6 8 10 MDA-MB-231 shCtrl+Vector shCtrl+STAT3 shENSA+Vector shENSA+STAT3 P<0.0001 P<0.0001 shCtrl+Vector shCtrl+STAT3 shENSA+Vector shENSA+STAT3 0 50 100 150 200 P=0.0004 P=0.0004 - - + + - + - + 0 50 100 150 200 250 P<0.0001 P<0.0001 shENSA STAT3 - - + + - + - + Days Tumor volume (mm3) shCtrl shENSA shENSA+ENSA shENSA+STAT3 18 25 32 39 0 500 1000 1500 2000 P<0.0001 P<0.0001 P<0.0001 Tumor weight (g) 0.0 0.4 0.8 1.2 1.6 shENSA+STAT3 shENSA+ENSA shENSA shCtrl P<0.0001 P<0.0001 P<0.0001 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications a d Days Days Relative growth rate 1 2 3 4 5 6 0 2 4 6 8 BT549 P<0.0001 P<0.0001 1 2 3 4 5 6 0 2 4 6 8 10 MDA-MB-231 shCtrl+Vector shCtrl+STAT3 shENSA+Vector shENSA+STAT3 P<0.0001 P<0.0001 shCtrl+Vector shCtrl+STAT3 shENSA+Vector shENSA+STAT3 c 75 75 25 60 STAT3 shENSA Flag-ENSA - + - + - - + + BT549 Flag TUBA1B - + - + - - + + MDA-MB-231 pSTAT3 (Tyr705) kDa d c MDA-MB-231 - + shENSA Vector STAT3 Colony number 0 50 100 150 200 250 P<0.0001 P<0.0001 shENSA STAT3 - - + + - + - + shENSA STAT3 Colony number e - + shENSA Vector STAT3 BT549 0 50 100 150 200 P=0.0004 P=0.0004 - - + + - + - + e - + shENSA Vector STAT3 BT549 shENSA STAT3 Colony number 0 50 100 150 200 P=0.0004 P=0.0004 - - + + - + - + e e h ENSA pSTAT3 Tyr705 Cleaved Caspase 3 shCtrl shENSA shENSA+ ENSA shENSA+ STAT3 100μm f Days Tumor volume (mm3) shCtrl shENSA shENSA+ENSA shENSA+STAT3 18 25 32 39 0 500 1000 1500 2000 P<0.0001 P<0.0001 P<0.0001 h f g Tumor weight (g) 0.0 0.4 0.8 1.2 1.6 shENSA+STAT3 shENSA+ENSA shENSA shCtrl P<0.0001 P<0.0001 P<0.0001 g biosynthesis was the most enriched pathway, and cholesterol addition abolished the impaired growth effect induced by ENSA knockdown in TNBC cells. These preliminary results link ENSA to lipid metabolic programming in TNBC. production of other metabolites, such as isoprenoids, dolichol, and ubiquinone, to support multiple cellular processes37. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z a MIER1 SRF EGR2 AP1 RUVBL2 STAT3 Regulated genes CCL2, ICAM1, IRF1 SNAIL, DHRS2, ENPP3 TNFRSF9, BTK, DMKN PTPN7, MYB, EGR1 LCP1, LMCD1, EGR1 NYAP2, GDAP1, RAB39A Predicted transcription factors shENSA vs. shCtrl P<0.05 0.0 0.5 1.0 1.5 2.0 -NES a MIER1 SRF EGR2 AP1 RUVBL2 STAT3 Regula CCL2, IC SNAIL, D TNFRSF PTPN7, LCP1, LM NYAP2, Predicted transcription factors 0.0 0.5 1.0 1.5 2.0 -NES STAT3 pSTAT3 (Tyr705) GAPDH MDA-MB-231 shENSA ENSA b BT549 pSTAT3 (Ser727) kDa 35 15 75 75 75 - #1 #2 - #1 #2 iosynthesis was the most enriched pathway, and cholesterol ddition abolished the impaired growth effect induced by ENSA nockdown in TNBC cells. These preliminary results link ENSA to pid metabolic programming in TNBC. The mevalonate (MVA) pathway is a critical metabolic path- way responsible for de novo cholesterol biosynthesis and the production of other metabolites, such as isoprenoids, do and ubiquinone, to support multiple cellular processes37 importance of the MVA pathway and its metabolites in porting cancer cell survival and growth has been increa appreciated38. Cholesterol is a vital metabolite for the biol functions of mammalian cells. Its concentration at bot STAT3 pSTAT3 (Tyr705) GAPDH MDA-MB-231 shENSA ENSA a b c BT549 pSTAT3 (Ser727) shENSA STAT3 Colony number Colony number f h 75 75 25 60 d e g ENSA pSTAT3 Tyr705 Cleaved Caspase 3 shCtrl shENSA shENSA+ ENSA shENSA+ STAT3 100μm kDa STAT3 shENSA Flag-ENSA - + - + - - + + BT549 Flag TUBA1B - + - + - - + + MDA-MB-231 pSTAT3 (Tyr705) kDa - + shENSA Vector STAT3 BT549 MDA-MB-231 - + shENSA Vector STAT3 35 15 75 75 75 - #1 #2 - #1 #2 MIER1 SRF EGR2 AP1 RUVBL2 STAT3 Regulated genes CCL2, ICAM1, IRF1 SNAIL, DHRS2, ENPP3 TNFRSF9, BTK, DMKN PTPN7, MYB, EGR1 LCP1, LMCD1, EGR1 NYAP2, GDAP1, RAB39A Predicted transcription factors shENSA vs. Discussion Multiple lines of evidence have proven that CNAs promote the initiation and progression of cancers by altering the expression levels of oncogenes and tumor suppressors. In this study, we found signiﬁcant ampliﬁcation of the chromosome 1q21.3 region in TNBC, and gene ENSA at this locus was highly expressed. ENSA regulates TNBC cell growth in vitro and in vivo through 7 TURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications ARTICLE ARTICLE Metabolite quinone coenzyme Q is involved in cancer cell energy metabolism47. High-resolution CRISPR screens also identiﬁed several MVA pathway enzymes essential for the survival of cancer cells48. In the current study, we found signiﬁcantly decreased concentrations of cholesterol and suppressed expression of cholesterol biosynthesis enzymes upon ENSA silencing. However, the addition of cholesterol alone was not able to completely rescue impaired cell growth induced by ENSA depletion, implying the possibility of other MVK pathway enzymes or metabolites regulating ENSA-induced cell growth. Together, these results suggest the importance of the MVK pathway, primarily cholesterol biosynthesis, in the regulation of TNBC growth by ENSA. through multiple pathways, of which the STAT3-SREBP2 axis was the most important. p The phosphorylation level of STAT3 can be regulated in dif- ferent ways. On the one hand, phosphorylation of STAT3 at Tyr705 by tyrosine kinases such as JAK and SRC or at Ser727 by JNK and other MAPKs results in its activation in cancer. On the other hand, inhibition of negative regulators such as PIAS3, SOCS1 and 3 and several cellular phosphatases (SHP1 and 2, PTPRD, PTPRT, PTPN1 and 2, DUSP22) can also lead to STAT3 activation in cancer. To determine how ENSA impacts the phosphorylation of STAT3, we focused on PP2A, a serine/threonine phosphatase whose functions can be suppressed by direct interaction with ENSA. It has been reported that pharmacologic inhibition of PP2A induces the phosphorylation of STAT3 on serine residues in T cells and vascular smooth muscle cells58,59. Unlike previous studies, we found that genetic inhibition of the PP2A catalytic subunit in control or ENSA knockdown TNBC cells could alter the phos- phorylation of STAT3 at the Tyr705 residue instead of the Ser727 residue, which implied that ENSA-PP2A affects p-STAT3 (Tyr705) expression in TNBC cells. The reason the observations for PP2A were different in our study might be related to the distinct cellular context, and PP2A might indirectly impact the tyrosine residue of STAT3 in TNBC cells. The underlying mechanism by which ENSA-PP2A acts on the tyrosine residue of STAT3 remains to be addressed in the near future. Some limitations of the current study should be acknowledged. Firstly, one cell line derived xenograft models and a limited number of mini-PDX models might not fully represent tumor features in vivo, more preclinical models are needed to verify the conclusions of this study. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z The importance of the MVA pathway and its metabolites in sup- porting cancer cell survival and growth has been increasingly appreciated38. Cholesterol is a vital metabolite for the biological functions of mammalian cells. Its concentration at both the p p g g The mevalonate (MVA) pathway is a critical metabolic path- way responsible for de novo cholesterol biosynthesis and the NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 8 ARTICLE Secondly, while mini-PDX models could overcome the time-consuming disadvantage of traditional PDX models and retained the accuracy and efﬁciency in drug sensitivity testing, they still have some limitations compared to PDX models. Unlike PDX models which represent a more realistic tumor microenvironment, tumors grown as mini-PDXs lack interactions with human microenvironmental components due to the removal of blood cells and ﬁbroblasts in the sample preparation process60. Thus, the involvement of microenvironmental components in ENSA-induced tumor progression and drug sensitivity deserve further investigation in proper preclinical models. g y De novo cholesterol biosynthesis is mainly controlled by SREBP2, the pivotal transcription factor for genes encoding enzymes involved in the cholesterol biosynthesis program49. SREBP-2 is synthesized as a 125 kDa inactive precursor and sequentially cleaved into the NH2-terminal form with nuclear translocation and transcription factor activity50. It was previously found that SREBP2 was altered by several oncogenic pathways, including p53, PI3K/AKT/mTOR, and AMPK51–55. However, the relationship between STAT3 and cholesterol biosynthesis remains unclear. Only a few studies have described a decrease in the mRNA expression of SREBP1/2 with the deletion of STAT3, but there is limited knowledge on the underlying mechanism56,57. Intriguingly, we used ChIP-qPCR to identify that STAT3 bound directly to the promoter sequences of SREBP2 and promoted the transcription of SREBP2 in TNBC cells. The decrease in tumor growth and the expression of SREBP2 by ENSA silencing could be abolished by ectopic STAT3 expression, which conﬁrmed the central role of STAT3 in cholesterol biosynthesis in TNBC. Additionally, the STAT3-independent pathway cannot be neglected since ectopic STAT3 expression alone could rescue most of but not all the growth inhibition induced by ENSA depletion. Other important oncogenic pathways, such as the MYC pathway, were also enriched in ENSA-depleted TNBC cells. Taken together, ENSA might promote TNBC progression In summary, our current research reveals that ENSA, a gene with recurrent CNA at the 1q 21.3 locus, is a trigger for tumor growth that acts by promoting cholesterol biosynthesis in TNBC. Further characterization of the potential mechanism of the ENSA-PP2A-STAT3-SREBP2 regulatory axis might support our ﬁndings. We propose that the STAT3 inhibitor Stattic might be an apt option for treating ENSA-expressing TNBC. Clinical samples. FUSCC TNBC cohort data (Sequence Read Archive (SRA) dataset: SRP157974; Gene Expression Omnibus (GEO) dataset: GSE118527) and ARTICLE ARTICLE Fig. 4 ENSA promotes tumor growth by activating STAT3. a Candidate transcription factor (TF) prediction performed by GSEA of regulatory target gene sets. The top 6 TFs (P < 0.05) ranked by absolute normalized enrichment scores are shown. NES score and nominal P-value were given by GSEA software. b Western blotting images showing the protein levels of phosphorylated STAT3 (pSTAT3)-Tyr705, pSTAT3-Ser727 and total STAT3 after ENSA knockdown in BT549 and MDA-MB-231 cells. n = 3 independent experiments. c Western blotting images showing the protein levels of pSTAT3-Tyr705 and total STAT3 in BT549 and MDA-MB-231 cells ± ENSA knockdown and ± ENSA overexpression. n = 3 independent experiments. d In vitro growth curves of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± STAT3 overexpression. n = 6. Data are presented as mean ± SD. Two-tailed two-way ANOVA tests. e Colony formation of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± STAT3 overexpression. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. f In vivo growth curve of tumors (n = 6) generated by injecting MDA-MB-231 cells expressing control or ENSA shRNA and rescued by ENSA or STAT3 overexpression. n = 6 mice per group. Data are presented as mean ± SD. Two-tailed two-way ANOVA tests. g Tumor weight of MDA-MB-231 cells (n = 6) expressing control or ENSA shRNA rescued by ENSA or STAT3 overexpression. n = 6 mice per group. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. h Immunohistochemical images of ENSA, pSTAT3-Tyr705, and cleaved caspase 3 in mammary fat pad xenograft models. Scale bar: 100 µm. Source data are provided as a Source Data ﬁle. NES normalized enrichment score. cellular and systemic levels has been linked to many diseases, such as obesity, heart disease and cancer. Several groups have addressed the relationship between hypercholesterolemia and increased breast cancer risk based on clinical data21,39,40. Importantly, 27-hydroxycholesterol is a key molecule that links hypercholesterolemia with breast cancer pathophysiology20. In addition to systemic cholesterol, dysregulated cellular cholesterol, derived from increased biosynthesis or uptake, fuels the malig- nant phenotypes of cancer cells, including proliferation, anti- apoptosis, migration, stemness, and immune escape41–46. Several other MVA pathway metabolites and enzymes have also been identiﬁed as oncogenic. For example, MVA-derived farnesyl- diphosphate and geranylgeranyl-diphosphate are critical for the isoprenylation of proteins, supporting their proper localization and function in cancer cells38. NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunicatio NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z siSTAT3 SREBP2 mRNA BT549 MDA-MB-231 b Relative to % input IgG STAT3 SREBP2 promoter oit a r in e R /c u L e vit ale R MDA-MB-231 pSTAT3 (Tyr705) + + + shENSA siPPP2CA - - - + - PP2A-C ENSA TUBA1B pSREBP2 15 60 75 35 140 a c shENSA STAT3 BT549 MDA-MB-231 Predicted STAT3 binding sequence of SREBP2 by JASPAR d f e g h kDa kDa pSREBP2 ACAT2 FDPS ENSA TUBA1B STAT3 60 45 45 140 75 75 pSTAT3 (Tyr705) 15 60 HMGCS1 Score Predicted sequence 14.1268 CTTCTAGGAAT 10.0287 CAGCTGGGAAG 8.88842 ATTCCTAGAAG 0.00 0.05 0.10 0.15 0.20 0.25 P=0.0004 0.0 0.5 1.0 1.5 - #1 #2 P=0.0052 P=0.0013 0.0 0.5 1.0 1.5 - #1 #2 P=0.0056 P=0.0022 0.0 0.5 1.0 1.5 - + - + - - + + shENSA STAT3 P=0.0009 P=0.0002 Vector STAT3 MDA-MB-231 shCtrl shENSA Free Cholesterol Nucleus Vector STAT3 BT549 75μm 75μm siPPP2CA 35 60 75 75 60 75 PP2A-C pSTAT3 (Ser727) TUBA1B MDA-MB-231 - #1 #2 pSTAT3 (Tyr705) STAT3 TUBA1B kDa - - + + - + - + - - + + - + - + Fig. 5 ENSA activates STAT3 to promote the transcription of SREBP2 in a PP2A-dependent manner. a JASPAR prediction of STAT3-binding sites on th sequence of SREBP2. b qRT-PCR and PCR analysis of STAT3 at the SREBP2 promoter after ChIP assays in MDA-MB-231 cells expressing control or ENSA shRNA. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. c qRT-PCR detecting relative SREBP2 mRNA expression in BT549 and MDA-MB-231 cells after STAT3 transient silencing. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. d Luciferase reporter assay detecting the activity of the SREBP2 promoter in BT549 cells ± ENSA knockdown and ± STAT3 overexpression. n = 3. Data are presented a mean ± SD. Two-tailed unpaired Student’s t tests. e Western blotting images showing the expression of enzymes involved in cholesterol biosynthesis, SREBP2, pSTAT3-Tyr705 and total STAT3 in BT549 and MDA-MB-231 cells with ± ENSA knockdown and ± STAT3 overexpression. n = 3 independent experiments. f Filipin III staining showing the cellular free cholesterol contents of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± STAT3 overexpression. n = 3 independent experiments. g Western blotting images showing the expression of STAT3, pSTAT3-Tyr705, and pSTAT3-Ser727 i MDA-MB-231 cells expressing control or PPP2CA siRNA. Methods Cli i l TURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467 022 28452 siSTAT3 SREBP2 mRNA BT549 MDA-MB-231 b Relative to % input IgG STAT3 SREBP2 promoter MDA-MB-231 pSTAT3 (Tyr705) + + + shENSA siPPP2CA - - - + - PP2A-C ENSA TUBA1B pSREBP2 15 60 75 35 140 c BT549 MDA-MB-231 cted STAT3 binding sequence of SREBP2 by JASPAR f g h kDa kDa 45 45 140 75 75 15 60 core Predicted sequence 1268 CTTCTAGGAAT 0287 CAGCTGGGAAG 8842 ATTCCTAGAAG 0.00 0.05 0.10 0.15 0.20 0.25 P=0.0004 0.0 0.5 1.0 1.5 - #1 #2 P=0.0052 P=0.0013 0.0 0.5 1.0 1.5 - #1 #2 P=0.0056 P=0.0022 0 5 0 5 - + - + - - + + SA AT3 P=0.0009 P=0.0002 Vector STAT3 MDA-MB-231 shCtrl shENSA Free Cholesterol Nucleus Vector STAT3 BT549 75μm 75μm siPPP2CA 35 60 75 75 60 75 PP2A-C pSTAT3 (Ser727) TUBA1B MDA-MB-231 - #1 #2 pSTAT3 (Tyr705) STAT3 TUBA1B kDa - - + + - + - + - - + + - + - + siSTAT3 SREBP2 mRNA BT549 MDA-MB-231 c 0.0 0.5 1.0 1.5 - #1 #2 P=0.0052 P=0.0013 0.0 0.5 1.0 1.5 - #1 #2 P=0.0056 P=0.0022 b Relative to % input IgG STAT3 SREBP2 promoter 0.00 0.05 0.10 0.15 0.20 0.25 P=0.0004 a Predicted STAT3 binding sequence of SREBP2 by JASPAR Score Predicted sequence 14.1268 CTTCTAGGAAT 10.0287 CAGCTGGGAAG 8.88842 ATTCCTAGAAG b a c oit a r in e R /c u L e vit ale R d 0.0 0.5 1.0 1.5 - + - + - - + + shENSA STAT3 P=0.0009 P=0.0002 f Vector STAT3 MDA-MB-231 shCtrl shENSA Free Cholesterol Nucleus Vector STAT3 BT549 75μm 75μm f d g siPPP2CA 35 60 75 75 60 75 PP2A-C pSTAT3 (Ser727) TUBA1B MDA-MB-231 - #1 #2 pSTAT3 (Tyr705) STAT3 TUBA1B kDa MDA-MB-231 pSTAT3 (Tyr705) + + + shENSA siPPP2CA - - - + - PP2A-C ENSA TUBA1B pSREBP2 15 60 75 35 140 h kDa shENSA STAT3 BT549 MDA-MB-231 e kDa pSREBP2 ACAT2 FDPS ENSA TUBA1B STAT3 60 45 45 140 75 75 pSTAT3 (Tyr705) 15 60 HMGCS1 - - + + - + - + - - + + - + - + h g e Fig. 5 ENSA activates STAT3 to promote the transcription of SREBP2 in a PP2A-dependent manner. a JASPAR prediction of STAT3-binding sites on the sequence of SREBP2. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z n = 3 independent experiments. h Western blotting images showing the expression of pSTAT3 Tyr705 and SREBP2 in MDA-MB-231 cells ± ENSA knockdown and ± transient PPP2CA knockdown. n = 3 independent experiments. Source data are provided as a Source Data ﬁle. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z b qRT-PCR and PCR analysis of STAT3 at the SREBP2 promoter after ChIP assays in MDA-MB-231 cells expressing control or ENSA shRNA. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. c qRT-PCR detecting relative SREBP2 mRNA expression in BT549 and MDA-MB-231 cells after STAT3 transient silencing. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. d Luciferase reporter assay detecting the activity of the SREBP2 promoter in BT549 cells ± ENSA knockdown and ± STAT3 overexpression. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. e Western blotting images showing the expression of enzymes involved in cholesterol biosynthesis, SREBP2, pSTAT3-Tyr705 and total STAT3 in BT549 and MDA-MB-231 cells with ± ENSA knockdown and ± STAT3 overexpression. n = 3 independent experiments. f Filipin III staining showing the cellular free cholesterol contents of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± STAT3 overexpression. n = 3 independent experiments. g Western blotting images showing the expression of STAT3, pSTAT3-Tyr705, and pSTAT3-Ser727 in MDA-MB-231 cells expressing control or PPP2CA siRNA. n = 3 independent experiments. h Western blotting images showing the expression of pSTAT3- Tyr705 and SREBP2 in MDA-MB-231 cells ± ENSA knockdown and ± transient PPP2CA knockdown. n = 3 independent experiments. Source data are provided as a Source Data ﬁle. Fig. 5 ENSA activates STAT3 to promote the transcription of SREBP2 in a PP2A-dependent manner. a JASPAR prediction of STAT3-binding sites on the sequence of SREBP2. b qRT-PCR and PCR analysis of STAT3 at the SREBP2 promoter after ChIP assays in MDA-MB-231 cells expressing control or ENSA shRNA. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. c qRT-PCR detecting relative SREBP2 mRNA expression in BT549 and MDA-MB-231 cells after STAT3 transient silencing. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. d Luciferase reporter assay detecting the activity of the SREBP2 promoter in BT549 cells ± ENSA knockdown and ± STAT3 overexpression. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t tests. e Western blotting images showing the expression of enzymes involved in cholesterol biosynthesis, SREBP2, pSTAT3-Tyr705 and total STAT3 in BT549 and MDA-MB-231 cells with ± ENSA knockdown and ± STAT3 overexpression. n = 3 independent experiments. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z P<0.0001 P<0.0001 P=0.0137 shENSA - + P=0.0021 0.0 0.5 1.0 1.5 - - + + - + - + shENSA Stattic P<0.0001 P=0.0026 Tumor weight (g) P=0.0010 578 573 543 553 P=0.7642 vehicle Stattic 0 50 100 150 200 250 P=0.2952 P=0.3385 554 584 552 P=0.0648 0 50 100 150 200 250 P<0.0001 P=0.0026 P=0.0079 P=0.0066 ytilib aiv lle C shCtrl (IC50=1.796μM) shENSA (IC50=3.282μM) log10 (stattic) μM -0.5 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 MDA-MB-231 0 1 2 3 4 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 shENSA - + Weeks after treatment m m ( e m ulo v r o m u T 3) P=0.0080 0 1 2 3 0 500 1000 1500 2000 shCtrl+vehicle shCtrl+Stattic shENSA+vehicle shENSA+Stattic P<0.0001 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z ARTICLE a shENSA - + P=0.0019 IC50 of Stattic (μM) ytilib aiv lle C shCtrl (IC50=1.796μM) shENSA (IC50=3.282μM) log10 (stattic) μM -0.5 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 MDA-MB-231 0 1 2 3 4 b shENSA Stattic Colony number MDA-MB-231 b DMSO 2.5μM stattic shNC shENSA - - + + - + - + 0 100 200 300 P<0.0001 P=0.0425 a Organoid 0 5 15 Stattic, μM d #1 #2 #3 Relative survival 0.0 0.5 1.0 1.5 P=0.0001 P=0.0023 n.s. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z ARTICLE Bioinformatic analyses. Gene set variation analysis of FUSCC TNBC tran- scriptome data was carried out to calculate the enrichment score of the cholesterol biosynthesis pathway in each sample with the R (v4.0.3) ‘GSVA’ package. The Pearson correlation metric was computed to assess the associations among the expression of different genes and between gene expression levels and cholesterol synthesis pathway scores by using the ‘cor’ function in R. Cell lines The TNBC cell lines BT549, MDA-MB-231, MCF7, T47D, SKBR3, and Short hairpin RNA (shRNA) vectors and lentiviral infections of cells. The two shRNAs with the best knockdown efﬁciency (shENSA-1: GAGCTGAAGAGG CAAAGCTAA; shENSA-2: CTGCCAGATCCTGAGACGCTT) were cloned into the pLKO.1 vector and introduced into HEK293T cells together with packing plasmids (psPAX2 and pMD2. G) and standard Lipofectamine 2000 transfection reagent (Thermo Fisher Scientiﬁc) to generate lentiviruses. The viral supernatants were collected and applied to infect TNBC cell lines in the presence of polybrene (10 µg/ml; Sigma-Aldrich). A plasmid expressing a nontargeting shRNA was used as the negative control. a c shENSA Stattic Colony number MDA-MB-231 b Organoid 0 5 15 Stattic, μM pSTAT3 (Tyr705) TUBA1B ENSA Organoid #1 #2 #3 pSREBP2 f d e Inhibition rate of tumor weight g ENSA stattic, μM 0 5 15 High Low 200μm #1 #2 #3 Inhibition rate of tumor volume 60 kDa 140 15 75 DMSO 2.5μM stattic shNC shENSA Surgery Mini-PDX Capsulize & Transplant Drug treatment Measure viability Relative viability h RLU i j low high ENSA 0 50 100 150 Relative viability (%) Low High - - + + - + - + shENSA - + P=0.0019 IC50 of Stattic (μM) 0 100 200 300 P<0.0001 P=0.0425 #1 #2 #3 Relative survival 0.0 0.5 1.0 1.5 P=0.0001 P=0.0023 n.s. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z f Filipin III staining showing the cellular free cholesterol contents of BT549 and MDA-MB-231 cells ± ENSA knockdown and ± STAT3 overexpression. n = 3 independent experiments. g Western blotting images showing the expression of STAT3, pSTAT3-Tyr705, and pSTAT3-Ser727 in MDA-MB-231 cells expressing control or PPP2CA siRNA. n = 3 independent experiments. h Western blotting images showing the expression of pSTAT3- Tyr705 and SREBP2 in MDA-MB-231 cells ± ENSA knockdown and ± transient PPP2CA knockdown. n = 3 independent experiments. Source data are provided as a Source Data ﬁle. data analyses were performed according to a previous study7. In brief, 465 female Chinese TNBC patients who underwent surgery at FUSCC were retrospectively selected, and their RNA-sequencing (RNA-seq) data, whole-exome sequencing, and OncoScan microarray copy number data were obtained. Among these 465 patients, 302 patients who had both RNA-seq data and copy number data were included in our study for screening candidate CNA-affected genes. Analysis of gene-level CNAs was performed according to a previous study7. In brief, probe- level output from the OncoScan Console was analyzed by ASCAT (v2.4.3), and the produced segment data were imported into GISTIC2.0 (v2.0.22) to acquire gene- level CNA data. included in our study for screening candidate CNA-affected genes. Analysis of gene-level CNAs was performed according to a previous study7. In brief, probe- level output from the OncoScan Console was analyzed by ASCAT (v2.4.3), and the produced segment data were imported into GISTIC2.0 (v2.0.22) to acquire gene- level CNA data. 10 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z P<0.0001 P<0.0001 P=0.0137 c pSTAT3 (Tyr705) TUBA1B ENSA Organoid #1 #2 #3 pSREBP2 60 kDa 140 15 75 e ENSA stattic, μM 0 5 15 High Low 200μm #1 #2 #3 d d e c f Inhibition rate of tumor volume shENSA - + P=0.0021 0.0 0.2 0.4 0.6 0.8 1.0 Weeks after treatment m m ( e m ulo v r o m u T 3) P=0.0080 0 1 2 3 0 500 1000 1500 2000 shCtrl+vehicle shCtrl+Stattic shENSA+vehicle shENSA+Stattic P<0.0001 Inhibition rate of tumor weight g 0.0 0.5 1.0 1.5 - - + + - + - + shENSA Stattic P<0.0001 P=0.0026 Tumor weight (g) P= 0.0 0.2 0.4 0.6 0.8 1.0 shENSA - f g 0.0 0.5 1.0 1.5 - - + + - + - + shENSA Stattic P<0.0001 P=0.0026 Tumor weight (g) g g Inhibition rate of tumor weight P=0.0010 0.0 0.2 0.4 0.6 0.8 1.0 shENSA - + Surgery Mini-PDX Capsulize & Transplant Drug treatment Measure viability Relative viability h i low high ENSA 0 50 100 150 Relative viability (%) Low High P=0.0066 RLU j 578 573 543 553 P=0.7642 vehicle Stattic 0 50 100 150 200 250 P=0.2952 P=0.3385 554 584 552 P=0.0648 0 50 100 150 200 250 P<0.0001 P=0.0026 P=0.0079 i j j h Relative viability (%) RLU Drug treatment Short hairpin RNA (shRNA) vectors and lentiviral infections of cells. The two shRNAs with the best knockdown efﬁciency (shENSA-1: GAGCTGAAGAGG CAAAGCTAA; shENSA-2: CTGCCAGATCCTGAGACGCTT) were cloned into the pLKO.1 vector and introduced into HEK293T cells together with packing plasmids (psPAX2 and pMD2. G) and standard Lipofectamine 2000 transfection reagent (Thermo Fisher Scientiﬁc) to generate lentiviruses. The viral supernatants were collected and applied to infect TNBC cell lines in the presence of polybrene (10 µg/ml; Sigma-Aldrich). A plasmid expressing a nontargeting shRNA was used as the negative control. Bioinformatic analyses. Gene set variation analysis of FUSCC TNBC tran- scriptome data was carried out to calculate the enrichment score of the cholesterol biosynthesis pathway in each sample with the R (v4.0.3) ‘GSVA’ package. The Pearson correlation metric was computed to assess the associations among the expression of different genes and between gene expression levels and cholesterol synthesis pathway scores by using the ‘cor’ function in R. Short hairpin RNA (shRNA) vectors and lentiviral infections of cells. The two Short hairpin RNA (shRNA) vectors and lentiviral infections of cells. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Fig. 6 ENSA is linked to Stattic sensitivity in TNBC. a Dose–response curves and half maximal inhibition concentration values of Stattic in MDA-MB-231 cells expressing control or ENSA shRNA. Dose-response curves: n = 6; Data are presented as mean ± SD. Bar plot: n = 3 independent experiments; Data are presented as mean ± SD; Two-tailed unpaired Student’s t test. b Clonogenic survival assays of MDA-MB-231 cells expressing control or ENSA shRNA and treated with 2.5 μM Stattic. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t test. c Western blotting images showing the expression of pSTAT3-Tyr705, SREBP2, and ENSA in three organoids. n = 3 independent experiments. d, e Results of the cell viability assay in three TNBC patient-derived organoid models treated with 0, 5, and 15 µM Stattic. d Cell viability assay and (e) Representative bright-ﬁeld images of organoids after drug treatment in three organoids. Scale bars, 200 µm. n = 3. Data are presented as mean ± SD. Two-tailed unpaired Student’s t test. f–g Stattic treatment of MDA-MB-231 cells expressing control or ENSA shRNA. Brieﬂy, we injected shCtrl or shENSA MDA-MB-231 cells into the mammary fat pad of female NOD/SCID mice (n = 10 each). When the tumor volume reached 50–100 mm3, each group was randomly assigned to two treatment groups: vehicle and Stattic. All groups (n = 5 each) received treatment (vehicle or 10 mg/kg Stattic) three times per week after randomization. The gray arrows indicate the treatments. f The growth curve (left) and the inhibition rate of tumor volume (right). n = 5 mice per group. Data are presented as mean ± SD. Two-way ANOVA test for growth curve and two-tailed unpaired Student’s t test for inhibition rate. g The tumor weight (left) and the inhibition rate of tumor weight (right). n = 5 mice per group. Data are presented as mean ± SD. Two-tailed unpaired Student’s t test. h Scheme of the generation of the mini-PDX models for the in vivo pharmacological tests. i The relative viability of seven TNBC mini-PDX models with Stattic treatment, as normalized to vehicle treatment. n = 3 in low-ENSA group and n = 4 in high-ENSA group. Data are presented as mean ± SD. Two-tailed unpaired Student’s t test. j The relative luminance unit of each TNBC mini-PDX model treated with Stattic or vehicle. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z After 3-5 passages, the organoids were added to each well of a 384-well plate, and different concentrations of Stattic (0, 5, and 15 μmol) were added to each well in duplicate and incubated for 5 days. Photos were taken on the last day to observe the changes in organoids under drug treatments. Acquisition of all clinical samples was approved by the Ethics Committee of FUSCC (Protocol number: 050432-4-1911D) and agreed to by each patient via signed informed consent. Cell growth and colony formation assay. For the cell growth assay, 2 × 103 (BT549, MDA-MB-231, and MCF7) and 4 × 103 (T47D, SKBR3, and BT474) cells were preseeded in 96-well plates in triplicate and incubated with 10% Cell Counting Kit-8 (CCK-8) solution (Vazyme, #A311-02) at 37 °C for 2 h, and then the absorbance was measured at 450 nm using a microplate reader. For the colony formation assay, 1 × 103 BT549 cells and 2 × 103 MDA-MB-231 cells were seeded into 6-well plates in triplicate, ﬁxed after 12 days, and stained with 0.25% crystal violet staining solution. Colonies consisting of more than 50 cells were counted. Acquisition of all clinical samples was approved by the Ethics Committee of FUSCC (Protocol number: 050432-4-1911D) and agreed to by each patient via signed informed consent. Mini-patient-derived xenograft (mini-PDX) model. In vivo pharmacological tests were conducted using OncoVee mini-PDX assay (LIDE Biotech, Shanghai, China) according to the previous papers24,25,60,63. In brief, fresh surgical tumor specimens were acquired from seven female breast cancer patients (average age: 55 years) at FUSCC. Specimens were then washed with Hank’s balanced salt solution (HBSS) to remove non-tumor tissues and necrotic tumor tissues. A fraction of tissue was retained for RNA extraction. The rest of the tissue was fragmented and digested with collagenase at 37 °C for 1–2 h. After centrifugation and removal of ﬁbroblasts and blood cells with magnetic beads, tumor cells were collected and suspended to ﬁll OncoVee capsules (LIDE Biotech, Shanghai, China). Each capsule contained 2000 cells and capsules derived from the same specimen were assigned to the baseline, vehicle control and Stattic treatment groups. Capsules were implanted subcutaneously into 5-week-old female nu/nu mice (3 capsules per mouse). Mice bearing capsules were treated with vehicle control or Stattic (10 mg/kg, intraper- itoneal injection) for seven continuous days. Each treatment (vehicle control or Stattic) was performed in quintuplicate or sextuplicate capsules. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Finally, the cap- sules were removed to measure cell viability in terms of relative luminance unit (RLU) using the CellTiter-Glo Luminescent Cell Viability Assay (Promega). Relative viability was calculated using the formula: Relative viability = (RLU of Stattic D7- RLU of bassline)/(RLU of vehicle D7- RLU of bassline) *100. Acqui- sition of all clinical samples was approved by the Ethics Committee of FUSCC (Protocol number: 050432-4-1911D) and agreed to by each patient via signed informed consent. The mini-PDX study protocol was approved by the Institutional Ethics Committee of Shanghai LIDE Biotech (Protocol number: LWIACUC002). Cell survival assay. To assess the effect of chemicals on breast cancer viability, cells were grown in 96-well plates at 2 × 103 cells per well and exposed to different concentrations of the test chemicals. After 72 h, the cells were incubated with 10% CCK-8 (Vazyme, #A311-02) solution at 37 °C for 2 h. The cell survival percentages at different concentrations were calculated by dividing the optical density (OD) of chemical-containing wells by that of DMSO-contacting wells. Flow cytometry analysis. For cell cycle analysis, a total of 1 × 106 cells were ﬁxed with precooled 70% ethanol overnight and then processed using the Cell Cycle and Apoptosis Analysis Kit (Yeasen, #40301ES50) according to the manufacturer’s instructions. For the cell apoptosis assay, 5 × 105 cells were collected and incubated with annexin V-ﬂuorescin isothiocyanate (FITC) and propidium iodide (PI) staining solution from the Annexin V-FITC/PI Apoptosis Detection Kit (Yeasen, #40302ES50). The ﬂow cytometry data were generated on a Beckman Cytomics FC 500 BD FACSCanto II and analyzed with FlowJo v10 software. Mouse models. The animal protocols were approved by the Animal Welfare Committee of Shanghai Medical College at Fudan University (Protocol number: 20210510011). Female 6-week-old NOD.CB17-Prkdc scid/JSlac mice were used for the in vivo mouse xenograft models. Mice were exposed to 12 h light, 12 h darkness cycle at a temperature of 21 ± 3 °C and an average of 55% humidity. To evaluate the role of ENSA on tumor growth, 2 × 106 shCtrl or shENSA MDA-MB-231 cells were harvested and resuspended in a 100 µl volume (PBS: Matrigel=1:1) and then injected into the mammary fat pads of the mice (n = 6 each group). For the treatment groups, 2 × 106 shCtrl or shENSA MDA-MB-231 cells were injected into the mammary fat pads of the mice (n = 5 each group). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z n = 5 or 6 independent capsules; Data are presented as mean ± SD. Two-tailed unpaired Student’s t test. Source data are provided as a Source Data ﬁle. n.s. not signiﬁcant, Mini-PDX mini-patient-derived xenograft, RLU relative luminance unit. width (shortest dimension). After the endpoint, the mice were euthanized, and tumors were excised for analysis. Bioluminescence imaging was performed using the Multimodal Animal Rotation System (Bruker). Relative bioluminescence signal quantitation was performed by the respective imaging system software packages. TCACTACTAAA) and PPP2CA (siPPP2CA-1: CCGTGAACGCATCACCATT; siPPP2CA-2: GATACAAATTACTTGTTTA) were purchased from RiboBio. siRNA transfection was conducted with Lipofectamine RNAIMAX Transfection Reagent (Thermo Fisher Scientiﬁc) according to the manufacturer’s instructions. TCACTACTAAA) and PPP2CA (siPPP2CA-1: CCGTGAACGCATCACCATT; siPPP2CA-2: GATACAAATTACTTGTTTA) were purchased from RiboBio. siRNA transfection was conducted with Lipofectamine RNAIMAX Transfection Reagent (Thermo Fisher Scientiﬁc) according to the manufacturer’s instructions. Plasmid and cloning. Human ENSA cDNA was purchased from GeneChem (Catalog number: NM_004436-GV492) and subcloned into the pCDH-CMV- MCS-EF1-puro plasmid (System Biosciences, Catalog number: CD510B-1). Human full-length STAT3 cDNA was purchased from Vigenebio (Catalog number: CH801341) and subcloned into the pCDH-CMV-MCS-EF1-GFP plasmid (mod- iﬁed from pCDH-CMV-MCS-EF1-puro). The SREBP2 promoter was ampliﬁed by using a pair of primers (forward: TGGTATTCCATCGTGTGGATGT; reverse: GAGTGAAGGGTTAACAGGCCA) in the BT549 cell line and cloned into the pGL3-basic vector (Promega). All transfections were performed using Lipofecta- mine 2000 transfection reagent (Thermo Fisher Scientiﬁc). Organoid. Patient-derived organoids used in the current study were derived from post-surgery specimens of three female patients who underwent surgery at the Department of breast, Fudan University Shanghai Cancer Center. The organoids were cultured based on previously described methods61,62. The organoids were suspended in Basement Membrane Extract (BME) Type 2 (Trevigen, 3533-010-02) and cultured in breast cancer organoid medium (Advanced DMEM/ Organoid. Patient-derived organoids used in the current study were derived from post-surgery specimens of three female patients who underwent surgery at the Department of breast, Fudan University Shanghai Cancer Center. The organoids were cultured based on previously described methods61,62. The organoids were suspended in Basement Membrane Extract (BME) Type 2 (Trevigen, 3533-010-02) and cultured in breast cancer organoid medium (Advanced DMEM/ F12 supplemented with R-spondin-1 [Peprotech], noggin [Peprotech], neuregulin [Peprotech], estradiol [Sigma], HEPES [Gibco], GlutaMAX [Gibco], nicotinamide [Sigma], N-acetylcysteine [Sigma], B-27 [Gibco], A83-01 [Tocris], primocin [InvivoGen], SB-202190 [Selleck], Y27632 [Selleck], FGF10 [Peprotech], FGF7 [Peprotech] and EGF [Peprotech]). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z The two shRNAs with the best knockdown efﬁciency (shENSA-1: GAGCTGAAGAGG CAAAGCTAA; shENSA-2: CTGCCAGATCCTGAGACGCTT) were cloned into the pLKO.1 vector and introduced into HEK293T cells together with packing plasmids (psPAX2 and pMD2. G) and standard Lipofectamine 2000 transfection reagent (Thermo Fisher Scientiﬁc) to generate lentiviruses. The viral supernatants were collected and applied to infect TNBC cell lines in the presence of polybrene (10 µg/ml; Sigma-Aldrich). A plasmid expressing a nontargeting shRNA was used as the negative control. Cell lines. The TNBC cell lines BT549, MDA-MB-231, MCF7, T47D, SKBR3, and BT474 and embryonic kidney cells (HEK293T) were obtained from the American Type Culture Collection (ATCC) and cultured in DMEM supplemented with 10% FBS. The cell lines were regularly conﬁrmed to be negative for mycoplasma con- tamination with a Mycoplasma Detecting Kit (Vazyme). Small interfering RNA (siRNA) delivery. siRNAs targeting human STAT3 (siSTAT3-1: GCAAAGAATCACATGCCACTT; siSTAT3-2: GGCGTCCAGT Small interfering RNA (siRNA) delivery. siRNAs targeting human STAT3 (siSTAT3-1: GCAAAGAATCACATGCCACTT; siSTAT3-2: GGCGTCCAGT 11 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z a b c d e f ENSA IHC score SREBP2 IHC score T N N T #1 T N N T T N N T T N N T #2 #3 #4 #5 #6 #7 #8 100μm g 1q21.3 Amp ENSA PP2A STAT3 SREBF2 SREBP2 SRE Promoter HMGCS1, MVK, … Cholesterol ENSA IHC score P=0.0028 Normal Tumor 0 2 4 6 8 10 Overall survival IHC High n=42 IHC Low n=96 P = 0.018 ENSA ENSA la viv r u s e e rf e s p ale R P = 0.014 IHC High n=42 IHC Low n=96 + + ++++++++++++++++++++++ + +++ + ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 0.00 0.25 0.50 0.75 1.00 0 12 24 36 48 60 Months +++++++ +++++++++++++++++++++++ + ++++++++ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 0.00 0.25 0.50 0.75 1.00 0 12 24 36 48 60 Months r = 0.517, P<0.001 0 2 4 6 8 2 4 6 8 ENSA IHC score pSTAT3 IHC score r = 0.523 , P<0.001 0 2 4 6 8 2 4 6 8 ENSA low expression ENSA High expression SREBP2 A S N E f o ytis n e t ni g ninia ts C H I 1 2 3 pSTAT3 100μm 100μm a T N N T #1 T N N T T N N T T N N T #2 #3 #4 #5 #6 #7 #8 100μm b ENSA IHC score P=0.0028 Normal Tumor 0 2 4 6 8 10 b d for normalization. The sequences of the primers used for RT-qPCR are shown Supplementary Table 5. A-seq and data analysis. A total of 1 mg RNA samples were treated with HTS mRNA capture beads (Vazyme, China) to enrich polyA+ RNAs prior to nstructing the RNA-seq libraries. The VAHTS mRNA-seq v2 library preparation kit of the Illumina Xten system (Vazyme, Nanjing, China) was used to p RNA-seq libraries according to the manufacturer’s instructions. Brieﬂy, RNA samples (~100 ng) were fragmented and reverse transcribed into strand cDNA. Then these cDNA fragments went through end repair, th of adenine tails, and ligation of the adaptors processes. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z When the tumor volumes reached 50–100 mm3, vehicle or Stattic (10 mg/kg) was intraperitoneally admi- nistered three times a week for two weeks. The tumor volumes were calculated as follows: V = L × W2 × 1/2, where L is length (longest dimension) and W is the RNA preparation and real-time quantitative reverse transcription (RT-qPCR). A RNeasy mini kit (Qiagen) was used for the puriﬁcation of total RNA from breast cancer cells following the manufacturer’s protocol. The extracted total RNA was subjected to cDNA synthesis using HiScript III RT SuperMix for qPCR (Vazyme). RT-qPCR was performed using ChamQ SYBR Color qPCR Master Mix (Vazyme) on a QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems). The expression of genes was calculated using the 2−ΔΔCt method, and the GAPDH was 12 URE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z The puriﬁed prod subjected to 12 cycles of PCR ampliﬁcation to create the ﬁnal cDNA li These libraries were sequenced on 150 bp paired-end Illumina sequenci c d e f ENSA IHC score SREBP2 IHC score T N N T T N N T T N N T T N N T #5 #6 #7 #8 100μm g 1q21.3 Amp ENSA PP2A STAT3 PP2A ENSA STAT3 STAT3 SREBF2 SREBP2 SRE Promoter HMGCS1, MVK, … Cholesterol TNBC growth Stattic ENSA IHC score P=0.0028 Normal Tumor 0 2 4 6 8 10 Overall survival IHC High n=42 IHC Low n=96 P = 0.018 ENSA ENSA la viv r u s e e rf e s p ale R P = 0.014 IHC High n=42 IHC Low n=96 + + ++++++++++++++++++++++ + +++ + ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 0.00 0.25 0.50 0.75 1.00 0 12 24 36 48 60 Months +++++++ +++++++++++++++++++++++ + ++++++++ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 0.00 0.25 0.50 0.75 1.00 0 12 24 36 48 60 Months r = 0.517, P<0.001 0 2 4 6 8 2 4 6 8 ENSA IHC score pSTAT3 IHC score r = 0.523 , P<0.001 0 2 4 6 8 2 4 6 8 ENSA low expression ENSA High expression SREBP2 A S N E f o ytis n e t ni g ninia ts C H I 1 2 3 pSTAT3 100μm 100μm c A S N E f o ytis n e t ni g ninia ts C H I 1 2 3 100μm e c e c e ENSA low expression ENSA High expression SREBP2 1 2 3 pSTAT3 100μm 100μm A S N E f o ytis n e t ni g ninia ts C H I pSTAT3 f ENSA IHC score SREBP2 IHC score r = 0.517, P<0.001 0 2 4 6 8 2 4 6 8 ENSA IHC score pSTAT3 IHC score r = 0.523 , P<0.001 0 2 4 6 8 2 4 6 8 d Overall survival IHC High n=42 IHC Low n=96 P = 0.018 ENSA ENSA la viv r u s e e rf e s p ale R P = 0.014 IHC High n=42 IHC Low n=96 + + ++++++++++++++++++++++ + +++ + ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 0.00 0.25 0.50 0.75 1.00 0 12 24 36 48 60 Months +++++++ +++++++++++++++++++++++ + ++++++++ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 0.00 0.25 0.50 0.75 1.00 0 12 24 36 48 60 Months f d g g g 1q21.3 Amp ENSA PP2A STAT3 PP2A ENSA STAT3 STAT3 SREBF2 SREBP2 SRE Promoter HMGCS1, MVK, … Stattic g 1q21.3 Amp ENSA PP2A STAT3 PP2A ENSA STAT3 STAT3 SREBF2 SREBP2 SRE Promoter HMGCS1, MVK, … Cholesterol TNBC growth Stattic Cholesterol TNBC growth Cholesterol Stattic TNBC growth kit of the Illumina Xten system (Vazyme, Nanjing, China) was used to prepare the RNA-seq libraries according to the manufacturer’s instructions. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Fig. 7 Correlations among ENSA, SREBP2, pSTAT3-Tyr705 and survival in clinical samples. a, b Representative IHC images (a) and IHC scores (b) of ENSA staining in 8 paired TNBC tissues and adjacent normal tissues. n = 8 paired samples. Two-tailed paired Student’s t test. c Immunohistochemical staining of ENSA in 138 TNBC specimens. Representative images are shown. Scale bars, 100 µm. d Kaplan–Meier analysis of the relapse-free survival and overall survival of 138 TNBC patients. A log-rank test was used to determine the statistical signiﬁcance between the low-ENSA expression group (n = 96) and the high ENSA expression group (n = 42). e IHC staining of SREBP2 and pSTAT3-Tyr705 in 138 TNBC specimens. Representative images are shown. Scale bars, 100 µm. n = 138 samples. f Correlation analysis of ENSA with SREBP2 and pSTAT3-Tyr705 expression levels in 138 TNBC tissues. Correlation coefﬁcients were calculated using the Spearman test. Two-tailed P-values were given. g Proposed working model. In TNBC, ENSA is ampliﬁed, highly expressed and inhibits the function of PP2A, resulting in STAT3 Tyr705 phosphorylation and activation. STAT3 activation induces SREBP2 transcription to upregulate cellular cholesterol biosynthesis and facilitate tumor progression. Inhibition of STAT3 signaling with Stattic might serve as an effective treatment strategy for 1q21.3-ampliﬁed and ENSA-highly expressed TNBC. Source data are provided as a Source Data ﬁle. IHC immunohistochemistry, Amp ampliﬁcation, SRE sterol regulatory element. Sequenced readings were aligned using HISAT2 with human genome GRCh38 as a reference genome. Gene expression levels were calculated from fragments per kilobase of transcript per million mapped reads (FPKM). Gene ontology (GO) analysis was performed using the Metascape tool, and the given input list contained genes that were expressed at a lower level than the shENSA group (fold change (FC) < 0.8). GSEA was performed using GSEA software (v3.0) and molecular sig- nature database (v7.0). Luciferase reporter assay. A total of 5 × 104 cells were seeded in 24-well plates and transiently cotransfected with pGL3-SREBP2 promoter reporter plasmids and pRL-TK (Promega) using Lipofectamine 2000 transfection reagent. The ﬁreﬂy and Renilla luciferase activities were measured with a dual-luciferase reporter system (Promega) according to the manufacturer’s instructions. The measurement was performed on a SpectraMax M5 Microplate Reader (Molecular Devices). Tissue specimens and immunohistochemistry (IHC). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z For IHC, 8 pairs of pri- mary TNBC tissues and adjacent normal tissues and primary TNBC specimens from 138 female patients (average age: 54 years), who underwent surgery at FUSCC from 2010-2013 were obtained from the Department of Pathology, Fudan Uni- versity Shanghai Cancer Center. The procedures for IHC were as follows: parafﬁn- embedded sections were deparafﬁnized at 60 °C for 4 h and then subjected to xylene and a graded series of alcohol. For antigen unmasking, the slides were heated with citrate or EDTA antigen retrieval solution. After cooling, the slides were blocked with blocking solution (2% goat serum, 2% bovine serum albumin, and 0.05% Tween 20 in PBS) for 10 min at room temperature (RT) and then incubated overnight with primary antibodies at 4 °C. The sections were covered with horseradish peroxidase (HRP)-conjugated secondary antibody (GeneTech) at RT for 30 min and then developed with 3,3′-diaminobenzidine substrate (Gene- Tech). The slides were counterstained with hematoxylin, dehydrated with a graded series of alcohols and then mounted with coverslips and mounting medium. The staining density was measured using a Leica CCD camera DFC420 connected to a Leica DM IRE2 microscope (Leica Microsystems Imaging Solutions Ltd.). Some staining images were scanned by PANNORAMIC MIDI (3DHISTECH Ltd.) and viewed with CaseViewer (v2.4). The IHC scores were calculated by multiplying staining intensity (0 = no, 1=weak, 2=moderate, 3=strong) with percentage of positive staining (0 = negative, 1 ≤10%, 2 = 10–50%, 3 ≥50%). Acquisition of all clinical samples was approved by the Ethics Committee of FUSCC (Protocol number: 050432-4-1911D) and agreed to by each patient via signed informed consent. For antibody use and details please see Supplementary Table 6. Cholesterol metabolism pathway analysis by LC-MS. ENSA MDA-MB-231 cells (n = 4 each) (1 × 107 shCtrl or shENSA) were collected for cholesterol analysis. Five hundred microliters of ethanol containing 5 µg of BHT were added to the cells. An internal standard cocktail (50 µL) comprising d6-lanosterol, d6-zymosterol, d7- desmosterol, d7-lathosterol, d7-d-dehydrocholesterol, and d6-cholesterol (Avanti Polar Lipids) was added to the samples. The samples were incubated at 1200 rpm for 15 min at 4 °C. At the end of incubation, 250 µL of Milli-Q water and 1 ml of n-hexane were added. The samples were mixed thoroughly by vortexing and then centrifuged at 15294 × g for 5 min at 4 °C. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Further information on research design is available in the Nature Research Reporting Summary linked to this article. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z The clear upper phase containing oxysterols and sterols in hexane was transferred to a new tube. The extraction was repeated once with another 1 ml of n-hexane. The pooled extract was dried in a SpeedVac under organic mode. Oxysterols and sterols were derivatized to obtain their picolinic acid esters prior to LC/MS analysis and quantitated by referencing the spiked internal standards as previously described64. The concentration of individual lipids (μmol/cell) was standardized to the z-score. Chemicals. Cholesterol (#S4154) and Stattic (#S7024) were purchased from Selleck. Western blotting. Total cellular protein was extracted using SDS lysis buffer [50 mM Tris (pH 8.1), 1 mM EDTA, 1% SDS, 1 mM fresh dithiothreitol, sodium ﬂuoride, and leupeptin] and quantiﬁed using the BCA Protein Assay Kit (Solarbio). A total of 20 µg protein was separated by SDS-PAGE and then electrotransferred onto polyvinylidene diﬂuoride membranes (Millipore). The membranes were incubated with the indicated primary antibody followed by an HRP-conjugated secondary antibody and then detected by enhanced chemiluminescence. For antibody use and details please see Supplementary Table 6. Statistics. Statistical analysis was performed using SPSS (version 20.0), R software (version 4.0.3) and GraphPad (version 8.0.2). Two-way ANOVA was used to analyze the variance between two growth curves. One-way ANOVA and unpaired or paired Student’s t tests were used to compare data between two groups. Cor- relation coefﬁcients were calculated using the Spearman test or Pearson test. The survival curves were generated by the Kaplan–Meier method and compared with the log-rank test. Multivariate Cox proportional hazard models provided calculated hazard ratios with 95% conﬁdence intervals. Two-sided P < 0.05 was considered statistically signiﬁcant. Filipin III staining. A total of 3 × 104 cultured cells were preseeded in 24-well plates. Then, the cells were harvested, ﬁxed with 4% paraformaldehyde and incu- bated with 0.05 mg/ml ﬁlipin III (Sigma, F4767) working solution for 2 h at room temperature. Then, the cells were sealed with SYTOX Deep Red stain (Invitrogen, P36990). This dye is excited by red light at 660 nm when bound to DNA and has an emission maximum at 682 nm; we detected these signals using a Cy5/deep red traditional ﬁlter. Filipin III staining of cells was visualized with a Leica DMI6000 B microscope at excitation wavelengths of 340–380 nm and emission wavelengths of 385–470 nm. The quantiﬁcation of Filipin staining was performed by ImageJ (v1.8.0). Reporting summary. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Brieﬂy, polyA+ RNA samples (~100 ng) were fragmented and reverse transcribed into double strand cDNA. Then these cDNA fragments went through end repair, the addition of adenine tails, and ligation of the adaptors processes. The puriﬁed products were subjected to 12 cycles of PCR ampliﬁcation to create the ﬁnal cDNA libraries. These libraries were sequenced on 150 bp paired-end Illumina sequencing run. kit of the Illumina Xten system (Vazyme, Nanjing, China) was used to prepare the RNA-seq libraries according to the manufacturer’s instructions. Brieﬂy, polyA+ RNA samples (~100 ng) were fragmented and reverse transcribed into double strand cDNA. Then these cDNA fragments went through end repair, the addition of adenine tails, and ligation of the adaptors processes. The puriﬁed products were subjected to 12 cycles of PCR ampliﬁcation to create the ﬁnal cDNA libraries. These libraries were sequenced on 150 bp paired-end Illumina sequencing run. used for normalization. The sequences of the primers used for RT-qPCR are shown in Supplementary Table 5. RNA-seq and data analysis. A total of 1 mg RNA samples were treated with VAHTS mRNA capture beads (Vazyme, China) to enrich polyA+ RNAs prior to constructing the RNA-seq libraries. The VAHTS mRNA-seq v2 library preparation 13 References High-resolution CRISPR screens reveal ﬁtness genes and genotype-speciﬁc cancer liabilities. Cell 163, 1515–1526 (2015). ld h h l f h l 17. Cancer Genome Atlas, N. Comprehensive molecular portraits of human breast tumours. Nature 490, 61–70 (2012). genotype-speciﬁc cancer liabilities. Cell 163, 1515–1526 (20 breast tumours. Nature 490, 61–70 (2012). 49. Brown, M. S. & Goldstein, J. L. The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell 89, 331–340 (1997). 18. Pelton, K., Freeman, M. R. & Solomon, K. R. Cholesterol and prostate cancer. Curr. Opin. Pharm. 12, 751–759 (2012). 19. Kuzu, O. F., Noory, M. A. & Robertson, G. P. The role of cholesterol in cancer. Cancer Res 76, 2063–2070 (2016). 50. Horton, J. D., Goldstein, J. L. & Brown, M. S. SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J. Clin. Investig. 109, 1125–1131 (2002). 20. Nelson, E. R. et al. 27-Hydroxycholesterol links hypercholesterolemia and breast cancer pathophysiology. Science 342, 1094–1098 (2013). 51. Moon, S. H. et al. p53 represses the mevalonate pathway to mediate tumor suppression. Cell 176, 564–580 (2019). e519. 21. Borgquist, S. et al. Cholesterol, cholesterol-lowering medication use, and breast cancer outcome in the BIG 1-98 Study. J. Clin. Oncol. 35, 1179–1188 (2017). 52. Wang, B. T. et al. The mammalian target of rapamycin regulates cholesterol biosynthetic gene expression and exhibits a rapamycin-resistant transcriptional proﬁle. Proc. Natl Acad. Sci. USA 108, 15201–15206 (2011). 22. Baek, A. E. et al. The cholesterol metabolite 27 hydroxycholesterol facilitates breast cancer metastasis through its actions on immune cells. Nat. Commun. 8, 864 (2017). 53. Li, Y. et al. AMPK phosphorylates and inhibits SREBP activity to attenuate hepatic steatosis and atherosclerosis in diet-induced insulin-resistant mice. Cell Metab. 13, 376–388 (2011). 23. Cai, D. et al. RORgamma is a targetable master regulator of cholesterol biosynthesis in a cancer subtype. Nat. Commun. 10, 4621 (2019). 24. Li, C. et al. Integrated omics of metastatic colorectal cancer. Cancer Cell 38, 734–747 (2020). e739. 54. Ricoult, S. J., Yecies, J. L., Ben-Sahra, I. & Manning, B. D. Oncogenic PI3K and K-Ras stimulate de novo lipid synthesis through mTORC1 and SREBP. Oncogene 35, 1250–1260 (2016). 25. Xu, S. et al. Genome-wide CRISPR screen identiﬁes ELP5 as a determinant of gemcitabine sensitivity in gallbladder cancer. Nat. Commun. 10, 5492 (2019). g 55. Porstmann, T. et al. ARTICLE ARTICLE ARTICLE References Huang, B., Song, B.-l & Xu, C. Cholesterol metabolism in cancer: mechanisms and therapeutic opportunities. Nat. Metab. 2, 132–141 (2020). 42. Liu, Z., Liu, X., Liu, S. & Cao, Q. Cholesterol promotes the migration and invasion of renal carcinoma cells by regulating the KLF5/miR-27a/FBXW7 pathway. Biochem. Biophys. Res. Commun. 502, 69–75 (2018). 9. Lehmann, B. D. et al. Reﬁnement of triple-negative breast cancer molecular subtypes: implications for neoadjuvant chemotherapy selection. PLoS ONE 11, e0157368 (2016). y p y 43. Wang, Y., Liu, C. & Hu, L. Cholesterol regulates cell proliferation and apoptosis of colorectal cancer by modulating miR-33a-PIM3 pathway. Biochem. Biophys. Res. Commun. 511, 685–692 (2019). 10. Beroukhim, R. et al. The landscape of somatic copy-number alteration across human cancers. Nature 463, 899–905 (2010). 11. Stratton, M. R., Campbell, P. J. & Futreal, P. A. The cancer genome. Nature 458, 719–724 (2009). p y 44. Ehmsen, S. et al. Increased cholesterol biosynthesis is a key characteristic of breast cancer stem cells inﬂuencing patient outcome. Cell Rep. 27, 3927–3938 (2019). e3926. 12. Hieronymus, H. et al. Copy number alteration burden predicts prostate cancer relapse. Proc. Natl Acad. Sci. USA 111, 11139–11144 (2014). 45. Novak, A. et al. Cholesterol masks membrane glycosphingolipid tumor- associated antigens to reduce their immunodetection in human cancer biopsies. Glycobiology 23, 1230–1239 (2013). 13. Bergamaschi, A. et al. Distinct patterns of DNA copy number alteration are associated with different clinicopathological features and gene-expression 13. Bergamaschi, A. et al. Distinct patterns of DNA copy number alteration are associated with different clinicopathological features and gene-expression subtypes of breast cancer. Genes Chromosomes Cancer 45, 1033–1040 (2006). y gy 46. Li, H. Y., Appelbaum, F. R., Willman, C. L., Zager, R. A. & Banker, D. E. Cholesterol-modulating agents kill acute myeloid leukemia cells and sensitize them to therapeutics by blocking adaptive cholesterol responses. Blood 101, 3628–3634 (2003). yp 14. Ciriello, G. et al. Emerging landscape of oncogenic signatures across human cancers. Nat. Genet. 45, 1127–1133 (2013). 15. Goutsouliak, K. et al. Towards personalized treatment for early stage HER2- positive breast cancer. Nat. Rev. Clin. Oncol. 17, 233–250 (2020). 47. Maiuri, M. C. & Kroemer, G. Essential role for oxidative phosphorylation in cancer progression. Cell Metab. 21, 11–12 (2015). 16. Chin, K. et al. Genomic and transcriptional aberrations linked to breast cancer pathophysiologies. Cancer Cell 10, 529–541 (2006). pathophysiologies. Cancer Cell 10, 529–541 (2006). 48. Hart, T. et al. References 1. Venkitaraman, R. Triple-negative/basal-like breast cancer: clinical, pathologic and molecular features. Expert Rev. Anticancer Ther. 10, 199–207 (2010). 33. Mochida, S., Maslen, S. L., Skehel, M. & Hunt, T. Greatwall phosphorylates an inhibitor of protein phosphatase 2A that is essential for mitosis. Science 330, 1670–1673 (2010). 2. Dent, R. et al. Triple-negative breast cancer: clinical features and patterns of recurrence. Clin. Cancer Res. 13, 4429–4434 (2007). 3. Garrido-Castro, A. C., Lin, N. U. & Polyak, K. Insights into molecular classiﬁcations of triple-negative breast cancer: improving patient selection for treatment. Cancer Discov. 9, 176–198 (2019). 34. Gharbi-Ayachi, A. et al. The substrate of Greatwall kinase, Arpp19, controls mitosis by inhibiting protein phosphatase 2A. Science 330, 1673–1677 (2010). 35. Charrasse, S. et al. Ensa controls S-phase length by modulating Treslin levels. Nat. Commun. 8, 206 (2017). 4. Yin, W. J. et al. Clinicopathological features of the triple-negative tumors in Chinese breast cancer patients. Breast Cancer Res Treat. 115, 325–333 (2009). 36. Chen, Y. L. et al. ENSA expression correlates with attenuated tumor propagation in liver cancer. Biochem. Biophys. Res. Commun. 442, 56–61 (2013). 5. Bianchini, G., Balko, J. M., Mayer, I. A., Sanders, M. E. & Gianni, L. Triple- negative breast cancer: challenges and opportunities of a heterogeneous disease. Nat. Rev. Clin. Oncol. 13, 674–690 (2016). 37. Buhaescu, I. & Izzedine, H. Mevalonate pathway: a review of clinical and therapeutical implications. Clin. Biochem. 40, 575–584 (2007). p p 38. Mullen, P. J., Yu, R., Longo, J., Archer, M. C. & Penn, L. Z. The interplay between cell signalling and the mevalonate pathway in cancer. Nat. Rev. Cancer 16, 718–731 (2016). 6. Carey, L., Winer, E., Viale, G., Cameron, D. & Gianni, L. Triple-negative breast cancer: disease entity or title of convenience? Nat. Rev. Clin. Oncol. 7, 683–692 (2010). 39. Kitahara, C. M. et al. Total cholesterol and cancer risk in a large prospective study in Korea. J. Clin. Oncol. 29, 1592–1598 (2011). 7. Jiang, Y.-Z. et al. Genomic and transcriptomic landscape of triple-negative breast cancers: subtypes and treatment strategies. Cancer Cell 35, 428–440 (2019). e425. 40. Johnson, K. E. et al. The relationship between circulating lipids and breast cancer risk: a Mendelian randomization study. PLoS Med. 17, e1003302 (2020). 8. Lehmann, B. D. et al. Identiﬁcation of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J. Clin. Invest. 121, 2750–2767 (2011). 41. Received: 25 February 2021; Accepted: 26 January 2022; 31. Thameem, F. et al. The transcribed endosulﬁne alpha gene is located within a type 2 diabetes-linked region on 1q: sequence and expression analysis in Pima Indians. Mol. Genet. Metab. 81, 16–21 (2004). 32. Héron, L., Virsolvy, A., Apiou, F., Le Cam, A. & Bataille, D. Isolation, characterization, and chromosomal localization of the human ENSA gene that encodes alpha-endosulﬁne, a regulator of beta-cell K(ATP) channels. Diabetes 48, 1873–1876 (1999). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-022-28452-z Received: 25 February 2021; Accepted: 26 January 2022; Data availability RNA-seq data generated in this study are deposited in the Sequence Read Archive database under the accession number PRJNA713612. FUSCC TNBC sequence data were available in the NCBI Gene Expression Omnibus (OncoScan array; GSE118527) and Sequence Read Archive (whole-exome sequencing and RNA-seq; SRP157974). The expression data, CNA data, and clinical data of the TCGA cohort were downloaded from the cBioPortal website (https://www.cbioportal.org/). The expression data were then transformed according to the log2(RSEM + 1) method. The METABRIC expression data were downloaded from the cBioPortal website (https://www.cbioportal.org/). The expression data of the SMC cohort were available in the GEO database (GSE113184). Kaplan–Meier survival plots were generated online with the Kaplan–Meier plotter database (https://kmplot.com/analysis/), and hazard ratios with 95% conﬁdence intervals and log-rank P values were calculated. The transcription factor binding site prediction was performed online with the JASPAR database (https://jaspar.genereg.net/). A public STAT3 ChIP-seq data were available in the GEO database (GSE152203). Source data are provided with this paper. Chromatin immunoprecipitation (ChIP). In brief, 1 × 107 cells were cross-linked with 1% formaldehyde and subjected to sonication in ChIP lysis buffer [50 mM HEPES (pH 7.5), 500 mM NaCl, 1 mM EDTA, 1% Triton X-100, and 0.1% Na- deoxycholate, supplemented with protease inhibitor cocktail]. Then, 4 μg anti- STAT3 or anti-mouse IgG antibodies with protein A/G magnetic beads (Invitro- gen, #10015D) were added to each ChIP reaction for incubation. After three washes with lysis buffer, three washes with wash buffer (50 mM HEPES, 300 mM LiCl, 1 mM EDTA, 0.5% NP-40, and 0.7% Na-deoxycholate) and one wash with Tris- EDTA buffer (TE), each ChIP reaction was eluted and reverse cross-linked in elution buffer [50 mM Tris-HCl (pH 8.0), 10 mM EDTA, and 1% SDS] at 65 °C for 4 h. After RNase A and proteinase K digestion, DNA samples were isolated with phenol:chloroform:isoamyl alcohol (25:24:1) and analyzed by qRT-PCR. All results are displayed as fold change to 1% input. For antibody use and details please see Supplementary Table 6. 14 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications Reprints and permission information is available at http://www.nature.com/reprints 64. Lam, S. M. et al. A multi-omics investigation of the composition and function of extracellular vesicles along the temporal trajectory of COVID-19. Nat. Metab. 3, 909–922 (2021). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Acknowledgements The authors are grateful to Yi-Zhou Jiang and Ding Ma for their excellent data man- agement. Cartoons in Figs. 6h, 7g were created with BioRender.com. This work was supported by grants from the National Natural Science Foundation of China (grants 81672600, 81722032, 82072916, and 91959207, received by K.-D.Y.), the 2018 Shanghai Youth Excellent Academic Leader (received by K.-D.Y.), and the Fudan ZHUOSHI Project (received by K.-D.Y.). The authors are grateful to Yi-Zhou Jiang and Ding Ma for their excellent data man- agement. Cartoons in Figs. 6h, 7g were created with BioRender.com. This work was supported by grants from the National Natural Science Foundation of China (grants 81672600, 81722032, 82072916, and 91959207, received by K.-D.Y.), the 2018 Shanghai Youth Excellent Academic Leader (received by K.-D.Y.), and the Fudan ZHUOSHI Project (received by K.-D.Y.). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. Youth Excellent Academic Leader (received by K.-D.Y.), and the Fudan ZHUOSHI Project (received by K.-D.Y.). Additional information 60. Zhang, F. et al. Characterization of drug responses of mini patient-derived xenografts in mice for predicting cancer patient clinical therapeutic response. Cancer Commun. (Lond.) 38, 60 (2018). 60. Zhang, F. et al. Characterization of drug responses of mini patient-derived xenografts in mice for predicting cancer patient clinical therapeutic response. Cancer Commun. (Lond.) 38, 60 (2018). Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-022-28452-z. 61. Sachs, N. et al. A living biobank of breast cancer organoids captures disease heterogeneity. Cell 172, 373–386 (2018). e310. Correspondence and requests for materials should be addressed to Ke-Da 62. Gong, Y. et al. Metabolic-pathway-based subtyping of triple-negative breast cancer reveals potential therapeutic targets. Cell Metab. 33, 51–64 (2021). e59. Peer review information Nature Communications thanks Carlos Fernández-Hernando and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. 63. Zhan, M. et al. Guided chemotherapy based on patient-derived mini-xenograft models improves survival of gallbladder carcinoma patients. Cancer Commun. (Lond.) 38, 48 (2018). Reprints and permission information is available at http://www.nature.com/reprints Author contributions K.-D.Y. and Y.-Y.C. designed the study; Y.-Y.-C. conducted and analyzed most of the experiments; J.-Y.G. and S.-Y.Z. conducted the remaining experiments; K.-D.Y. and Z.- M.S. provided crucial reagents and conduct management; Y.-Y.C. and K.-D.Y. prepared the manuscript. The authors read and approved the ﬁnal manuscript. © The Author(s) 2022 References PKB/Akt induces transcription of enzymes involved in cholesterol and fatty acid biosynthesis via activation of SREBP. Oncogene 24, 6465–6481 (2005). 26. Mesquita, B. et al. Frequent copy number gains at 1q21 and 1q32 are associated with overexpression of the ETS transcription factors ETV3 and ELF3 in breast cancer irrespective of molecular subtypes. Breast Cancer Res Treat. 138, 37–45 (2013). 56. Xu, Y., Ikegami, M., Wang, Y., Matsuzaki, Y. & Whitsett, J. A. Gene expression and biological processes inﬂuenced by deletion of Stat3 in pulmonary type II epithelial cells. BMC Genomics 8, 455 (2007). 27. Teixeira, M. R. et al. Evaluation of breast cancer polyclonality by combined chromosome banding and comparative genomic hybridization analysis. Neoplasia 3, 204–214 (2001). 57. Chen, J. et al. Blocking of STAT-3/SREBP1-mediated glucose-lipid metabolism is involved in dietary phytoestrogen-inhibited ovariectomized- induced body weight gain in rats. J. Nutr. Biochem. 61, 17–23 (2018). p 28. Adelaide, J. et al. Integrated proﬁling of basal and luminal breast cancers. Cancer Res. 67, 11565–11575 (2007). 58. Woetmann, A. et al. Inhibition of protein phosphatase 2A induces serine/ threonine phosphorylation, subcellular redistribution, and functional inhibition of STAT3. Proc. Natl Acad. Sci. USA 96, 10620–10625 (1999). 29. Bhattacharya, A. et al. Transcriptional effects of copy number alterations in a large set of human cancers. Nat. Commun. 11, 715 (2020). large set of human cancers. Nat. Commun. 11, 715 (2020) 30. Goh, J. Y. et al. Chromosome 1q21.3 ampliﬁcation is a trackable biomarker and actionable target for breast cancer recurrence. Nat. Med. 23, 1319–1330 (2017). 59. Liang, H. et al. Regulation of angiotensin II-induced phosphorylation of STAT3 in vascular smooth muscle cells. J. Biol. Chem. 274, 19846–19851 (1999). 15 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications Competing interests p g The authors declare no competing interests. NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunications 16 NATURE COMMUNICATIONS \| (2022) 13:791 \| https://doi.org/10.1038/s41467-022-28452-z \| www.nature.com/naturecommunication
https://openalex.org/W4365804421	https://www.scielo.br/j/cadsc/a/8pKkZvX4mWDKw3Jjmfz9NFP/?lang=pt&format=pdf	Portuguese	null	O impacto das assembleias na humanização em saúde em um centro de atenção psicossocial.	Cadernos Saúde Coletiva	2,023	cc-by	4,595	Resumo Introdução: Os Centro de Atenção Psicossocial (CAPS) foram criados como parte do processo de reforma psiquiátrica. Objetivo: O objetivo deste estudo foi avaliar a influência das assembleias realizadas no Centro de Atenção Psicossocial CAPS II de Boa Vista (RR) na humanização em saúde mental. Método: Esta pesquisa qualitativa avaliou 10 indivíduos dentre os 22 que preencheram os critérios de inclusão. O estudo utilizou a técnica de análise de conteúdo e um roteiro de entrevista contendo 07 perguntas abertas sobre a efetivação da humanização em meio à realização de assembleias. Resultados: A relação paciente, profissional e gestor melhorou após a realização das assembleias. Quanto ao trabalho desenvolvido pela gestão, os participantes atribuíram organização e humanização. O trabalho da equipe profissional foi considerado excelente pela maioria dos entrevistados, e o CAPS atua de forma humanizada no atendimento à pessoa em sofrimento psíquico. Conclusão: As assembleias influenciaram a humanização em saúde mental em face das mudanças de comportamento e relacionamento propiciadas pelo vínculo que se fez presente durante as mencionadas reuniões, tangenciando assim para lógica da reforma psiquiátrica e substituição de práticas hospitalocêntricas com vistas na ressocialização dos indivíduos e familiares. Palavras-chave: pesquisa sobre serviços de saúde; assistência à saúde mental; humanização da assistência. ARTIGO ORIGINAL ARTIGO ORIGINAL 1Departamento de Saúde Coletiva, Faculdade São Leopoldo Mandic - Campinas (SP), Brasil. Como citar: Formiga WDD, Zanin L, Flório FM, Oliveira AMG. O impacto das assembleias na humanização em saúde em um centro de atenção psicossocial. Cad. Saúde Colet., 2023; 31(1):e31010210. https://doi.org/10.1590/1414-462X202331010210 Trabalho realizado no Centro de Atenção Psicossocial (CAPS III) – Boa Vista (RR), Brasil. Correspondência: Luciane Zanin. E-mail: luciane.souza@slmandic.edu.br Fonte de financiamento: nenhuma. Conflito de interesses: nada a declarar. Recebido em: Maio 22, 2018. Aprovado em: Fev. 07, 2021 Abstract Background: The Psychosocial Care Center was created as part of the psychiatric reform process. Objective: The objective of this study was to evaluate the influence of the assemblies held at the CAPS II Psychosocial Care Center of Boa Vista RR on the humanization of mental health. Method: This qualitative study evaluated 10 individuals from the 22 who met the inclusion criteria. The study used a content analysis methodology applying an interview script containing 7 open questions regarding the effectiveness of humanization in the context of the assemblies. Results: The relationship between patient, professional and manager improved after the assemblies were held. Regarding the work undertaken by the management, participants highlighted the organization and humanization. The work of the professional team was considered excellent by most of the interviewees and the CAPS acts in a humanized way when assisting people in psychological distress. Conclusion: The assemblies influenced the humanization of mental health in the face of behavioural changes through the relationship established during the aforementioned meetings. This reflected the logic of psychiatric reform and the substitution of hospital-centered practices with the re-socialization of individuals and families. Keywords: health care surveys; mental health assistance; humanization of assistance. Este é um artigo publicado em acesso aberto (Open Access) sob a licença Creative Commons Attribution, que permite uso, distribuição e reprodução em qualquer meio, sem restrições desde que o trabalho original seja corretamente citado. 1/81 Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X2023310 O impacto das assembleias na humanização em saúde. INTRODUÇÃO Os Centro de Atenção Psicossocial (CAPS) foram criados como parte do processo de reforma psiquiátrica no Brasil e são formados por unidades de atendimento intensivo e diário para indivíduos que estejam passando por sofrimento psíquico. A criação dos CAPS no Brasil possibilitou a formação um ambiente favorável aos portadores de doenças mentais, de modo a não afastá-los da família e da sociedade, no período alusivo ao tratamento1. A Portaria nº 224/92 foi o primeiro documento oficial que estabeleceu critérios para o credenciamento e financiamento do CAPS pelo SUS, quando este passou a subsidiar o Ministério da Saúde, disponibilizando via CAPS os atendimentos ambulatoriais, atendimentos terapêuticos, individuais ou em grupos, ateliês abertos com atividade lúdicas e recreativas que são promovidas por profissionais do serviço voltados para o tratamento e reabilitação psicossocial com iniciativas ainda extensivas aos familiares, visando o apoio a estes, bem como as questões de ordem social que rodeiam o cotidiano dos usuários2. A Política Nacional de Humanização (PNH), implementada em 20033, tem como estratégia o trabalho em conjunto com as instituições e serviços de saúde, de forma a se firmar as práticas de humanização em saúde mental nos CAPS por meio da criação de espaços de interlocução que alteram as formas de produzir saúde, aumentam o grau de comunicação entre sujeitos e equipes, proporcionando mais autonomia dos envolvidos4. No campo da saúde metal, o Ministério da Saúde define a realização de assembleias como um importante instrumento para o funcionamento dos CAPS5, pois tem como finalidade contribuir para o processo de não institucionalização por meio de novas abordagens terapêuticas, elevando a dimensão psicossocial do sofrimento ao campo da subjetividade humana e da inclusão social, haja vista que esses espaços democráticos podem representar uma perspectiva de valorização da humanização, cidadania e autonomia dos sujeitos5. A inclusão do usuário na gestão do cotidiano institucional reforça o caráter de mudança do modelo manicomial ao de atenção psicossocial. Assim, as assembleias representam um espaço de reflexões no qual se reúnem profissionais, usuários e familiares atuando como protagonistas nos processos que envolvem a sua própria saúde6. O trabalho teve como objetivo analisar o impacto das assembleias realizadas no Centro de Atenção Psicossocial – CAPS II de Boa Vista (RR) na humanização em saúde mental. MÉTODO Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 22/8 O impacto das assembleias na humanização em saúde. ser o conteúdo um meio de compreensão de significados implícitos nas mensagens. A análise do material preconiza três etapas: pré-análise (leitura flutuante e formulação de hipóteses), exploração do material (codificação e classificação em categorias), tratamento dos resultados abordados e interpretação (processo de reflexão), segundo Bardin8. Utilizou-se um Roteiro de Entrevista aberto9, contendo sete questões que abordavam a visão dos participantes sobre a importância da realização das assembleias, principais mudanças ocorridas, observações sobre o trabalho dos profissionais, insatisfações, relacionamento com a equipe e percepção sobre a humanização no serviço oferecido pelo CAPS. As entrevistas aconteceram de forma individualizada no CAPS, em sala reservada da própria instituição, com duração aproximada de 30 a 40 minutos. Utilizou-se um gravador para gravação e transcrição subsequente das falas e conteúdos em sua plenitude, e posterior análise empregando a metodologia proposta por Bardin8. Na fase inicial, pré-análise, o material foi organizado para constituir o corpus da pesquisa, por meio da seleção de documentos e formulação de hipóteses, em meio a uma leitura flutuante e consecutiva de conteúdos inerentes à temática humanização versus assembleia, sistematizando os dados na ótica de explorar ao máximo as informações, como requerida na análise de conteúdo no tocante à exaustividade da busca pelo esgotamento das fontes que constituíram o supedâneo do mencionado estudo, sendo, portanto, grifados os pontos de maior interesse e, a posteriori, selecionados recortes imprescindíveis à pesquisa. Na segunda fase da análise, definiram-se as temáticas do estudo com identificação das unidades de registro e organização dos dados de acordo com o objetivo da pesquisa e similaridade dos depoimentos, agregando as representações, expectativas e argumentos aludidos pelos entrevistados daquele Centro de Atenção Psicossocial. A terceira fase foi marcada pela inferência e interpretação que se materializou a partir da condensação das informações necessárias à análise, configurando-se, portanto, como o momento de realização da análise crítica e reflexiva do estudo, com vistas no embasamento das análises e fortalecimento do contexto interpretativo das falas dos participantes. Esta pesquisa de natureza exploratória descritiva foi aprovada pelo Comitê de Ética em Pesquisa (Protocolo - CAAE: 59321716.9.0000.5374), conforme resolução nº. 466/2012 do Conselho Nacional de Saúde. MÉTODO A pesquisa desenvolveu-se no ano de de 2017 no CAPS II da cidade de Boa Vista, Estado de Roraima, localizada na região Norte do Brasil. Concentrando cerca de dois terços dos habitantes do estado, situada à margem direita do Rio Branco, com população estimada em 320.714 habitantes em 2015, conforme o Instituto Brasileiro de Geografia e Estatística (IBGE)7, o CAPS de Boa Vista (RR), no ano de 2017, atendia 30 usuários em regine intensivo de tratamento. Para a composição da amostra final, excluíram-se 8 participantes que não apresentavam capacidade de discernimento para a participação no estudo e usou-se como critério o método de saturação nos discursos, que ocorre quando os dados obtidos passam a apresentar, na avaliação do pesquisador, uma redundância ou repetição contínua. Dessa forma, a amostra final foi composta por 10 usuários que assinaram o Termo de Consentimento Livre e Esclarecido. A pesquisa desenvolveu-se no ano de de 2017 no CAPS II da cidade de Boa Vista, Estado de Roraima, localizada na região Norte do Brasil. Concentrando cerca de dois terços dos habitantes do estado, situada à margem direita do Rio Branco, com população estimada em 320.714 habitantes em 2015, conforme o Instituto Brasileiro de Geografia e Estatística (IBGE)7, o CAPS de Boa Vista (RR), no ano de 2017, atendia 30 usuários em regine intensivo de tratamento. Para a composição da amostra final, excluíram-se 8 participantes que não apresentavam capacidade de discernimento para a participação no estudo e usou-se como critério o método Trata-se de um estudo descritivo que adotou a Análise de Conteúdo de Bardin8, tendo como função primordial o desvendar crítico apreendido nos discursos por enfocar as relações entre gestores, profissionais e usuários; adotou-se como referencial a interação das falas dos usuários, buscando-se a interseção dos significados, estabelecendo e determinando o comportamento manifesto, processo dinâmico e contínuo no qual significados e ações são transformados. Com vistas a potencializar a exposição detalhada do fenômeno e promover sua compreensão, a entrevista aberta individual foi a estratégia adotada para a coleta dos dados, considerando-se que há pessoas sob fenômenos semelhantes e de um mesmo contexto, o que possibilita a hipótese de similaridade textual e encontro de ideias. O material empírico gerado sofreu os processos analíticos preconizados pela Análise de Conteúdo Temático de Bardin8. Este método faz parte do conjunto de técnicas utilizadas para a análise da comunicação e entende Cad. Perfil dos sujeitos do estudo Os participantes possuem idade média de 41,4 anos, a maioria com apenas o ensino médio completo, sendo as principais alterações apresentadas a depressão e o transtorno bipolar (Quadro 1). Quadro 1. Perfil dos sujeitos entrevistados quanto à idade, escolaridade e patologia apresentada Sujeitos Idade Escolaridade Diagnóstico pelo CID S1 45 superior Esquizofrenia S2 29 ensino médio Depressão S3 58 ensino médio Transtorno bipolar S4 24 ensino médio Transtorno de personalidade borderline S5 41 fundamental Depressão e síndrome do pânico S6 62 fundamental Esquizofrenia S7 32 ensino médio Depressão, transtorno de ansiedade generalizada S8 30 ensino médio Depressão S9 37 ensino médio Transtorno bipolar S10 56 superior Esquizofrenia Quadro 1. Perfil dos sujeitos entrevistados quanto à idade, escolaridade e patologia apresentada Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 3/83 3/83 O impacto das assembleias na humanização em saúde. Das entrevistas com os sujeitos, emergiram 4 temas para análise com base no contexto vivenciado durante a realização das assembleias, correlacionando o mecanismo utilizado no CAPS e a humanização no atendimento, de acordo com a Política Nacional de Humanização (PNH)3, segundo o Princípio da indissociabilidade entre gestão, atenção e usuário, e o Protagonismo dos sujeitos, tendo em vista que as assembleias são momentos de encontro entre usuários, gestores e profissionais de saúde. Temas emergentes segundo a temática vivenciada durante as assembleia A importância da realização das assembleias; ▪ A importância da realização das assembleias; ▪ Mudanças na relação paciente-profissional após a assembleias; p ç ▪ Mudanças na relação paciente-profissional após a assembleias; ▪ Mudanças observadas após as assembleias; e ▪ Percepção em relação à humanização dos serviços oferecidos. A importância da realização das assembleias As assembleias ocorrem no CAPS de Boa Vista mensalmente desde 2013, com o propósito de estabelecer vínculos entre profissionais, gestores, usuários e familiares, favorecendo o fortalecimento dos processos democráticos de discussão e a inclusão dos pacientes nos processos decisórios, na melhoria dos serviços e na idealização de novas oficinas. Durante essas reuniões são abordados assuntos sobre relacionamento, evolução do tratamento, satisfação dos usuários, atividades lúdicas e recreativas desenvolvidas no CAPS, projetos de educação e saúde, relatos de histórias vivenciadas pelos usuários antes e depois de frequentar o centro. No discurso analisado, fica evidente a visão dos sujeitos em relação à humanização do CAPS, uma vez que o reconhece como espaço que viabiliza a liberdade de expressão, aproximação entre familiares, usuários, profissionais. [...] Criar um entorno entre familiares, usuários, técnico e falar sobre as angústias. (Sujeito 4). [...] As assembleias são importantes, porque discutimos as melhorias para os pacientes e do transtorno mental. (Sujeito 6). [...] As assembleias são importantes, porque discutimos as melhorias para os pacientes e do transtorno mental. (Sujeito 6). [...] Acho que as assembleias permitem que nós possamos opinar em relação ao que está bom, deixando a nossa família mais segura em relação a como lidar com a nossa doença. (Sujeito 9). Identifica-se nessas falas o princípio da transversalidade da PNH3, que busca a construção de um plano comum dos saberes e práticas de saúde, visando desestabilizar fronteiras dos saberes, territórios de poder e modos instituídos nas relações de trabalho. Este movimento de transversalidade coloca os serviços de saúde no caminho da humanização10. A resposta do sujeito trouxe para o bojo deste estudo a necessidade de reflexão sobre a importância das assembleias nos CAPS como espaço de fala que põe os sujeitos/usuários em ação, no qual são estimulados a produzir um sentido para as suas questões. Portanto, as assembleias se tornam um espaço em que os pacientes podem expressar seus sentimentos e ideias e, a partir disso, podem provocar mudanças6. Mudanças na relação paciente-profissional após a assembleias Na contramão dessa argumentação, alguns sujeitos relataram que ainda percebem distanciamento da equipe e insatisfação [...] Eu acho que a gente era bem mais próximo deles, hoje em dia não, hoje em dia nós somos separados para um lado, a gente quase não vê eles, só no momento das oficinas. (Sujeito 5). [...] Não sei, porque eles ficaram bem mais afastados da gente, a gente fica mais na Terapia Ocupacional e eles mais pra cá, assim, só vão mesmo na hora, na hora da terapia. (Sujeito 8). [...] Eu acho que a gente era bem mais próximo deles, hoje em dia não, hoje em dia nós somos separados para um lado, a gente quase não vê eles, só no momento das oficinas. (Sujeito 5). [...] Eu acho que a gente era bem mais próximo deles, hoje em dia não, hoje em dia nós somos separados para um lado, a gente quase não vê eles, só no momento das oficinas. (Sujeito 5). [...] Não sei, porque eles ficaram bem mais afastados da gente, a gente fica mais na Terapia Ocupacional e eles mais pra cá, assim, só vão mesmo na hora, na hora da terapia. (Sujeito 8). As relações de afeto, proximidade e troca de informações entre pacientes e profissionais devem constituir o arcabouço da humanização, fato que pressupõe a necessidade de engajamento entre todos os membros de uma equipe. Assim, a humanização deve se constituir como prática solidária do profissional, refletida na compreensão e no olhar sensível, aquele olhar de cuidado que desperta no ser humano sentimento de confiança e solidariedade14. Os CAPS assumem um papel crucial para organizar a rede comunitária de cuidados, realizando o direcionamento para construção coletiva das políticas e programas de Saúde Mental. No interior dos CAPS, as assembleias ou Reuniões de Organização do Serviço se tornam um instrumento para gestão e funcionamento deste local de convivência. Este é o momento em que são discutidos os problemas e trazidas sugestões sobre como melhorar o serviço de saúde na esfera da saúde15. Mudanças na relação paciente-profissional após a assembleias Indagado sobre mudanças significativa na relação entre paciente, profissional e gestor após a realização das assembleias, parte dos sujeitos afirmaram reconhecer a proximidade, respeitabilidade, carinho e vínculo: [...] Eles têm ficado mais próximos, sempre que precisamos da ajuda da diretora ela está a postos para responder às nossas necessidades. (Sujeito 1). [...] Eles têm ficado mais próximos, sempre que precisamos da ajuda da diretora ela está a postos para responder às nossas necessidades. (Sujeito 1). [...] Tem respeito, amizade e carinho. (Sujeito 3). [...] Tem respeito, amizade e carinho. (Sujeito 3). [...] Olha, têm diversas mudanças que as pessoas tem, a gente fica mais próxima deles. (Sujeito 5). [...] Acho que, desde o momento em que tivemos essa oportunidade de demonstrar a nossa opinião, a ligação entre os profissionais, a direção e a gente se tornou muito forte, pois eles se importam e compreendem nossas angústias. (Sujeito 9). [...] Acho que, desde o momento em que tivemos essa oportunidade de demonstrar a nossa opinião, a ligação entre os profissionais, a direção e a gente se tornou muito forte, pois eles se importam e compreendem nossas angústias. (Sujeito 9). Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 44/8 O impacto das assembleias na humanização em saúde. A relação paciente-profissional mais próxima possibilita estreitar o vínculo, além de representar para os usuários melhor cuidado em relação aos seus problemas11. Esses relatos também possibilitam uma reflexão acerca da indissociabilidade da gestão do serviço em saúde, incluindo suas dimensões política, administrativa, financeira, organizacional, técnica, afetiva, subjetiva e terapêutica, pois, para construir o bem-estar social, é substancial integrar a gestão e o trabalho humano. Um dos principais desafios para a construção de um novo paradigma na atenção da saúde à loucura é a elaboração de processos de gestão que impliquem nessa indissociabilidade12. Ao participar das assembleias, o usuário da saúde mental aumenta o vínculo, amplia a participação nas decisões do cotidiano dos serviços de saúde, capacitando-o para do controle social e para fortalecer o protagonismo da saúde mental, onde os pacientes podem pensar, sentir e decidir sobre suas vidas, sendo os atores principais com poder de decisão sobre si mesmo13. Percepção em relação à humanização dos serviços oferecidos [...] Os profissionais são muito acolhedores e se empenham muito nas suas atividades, então estou satisfeita sim. (Sujeito 9). [...] Está muito motivado. (Sujeito 3). [...] Os profissionais são muito acolhedores e se empenham muito nas suas atividades, então estou satisfeita sim. (Sujeito 9). Na área da saúde mental o diálogo, o vínculo e o acolhimento possibilitam uma “intimidade terapêutica”, em que o profissional estará aberto à escuta das necessidades de saúde do usuário, tendo uma postura mais acolhedora. Busca-se no acolhimento compreender o sofrimento psíquico, atentando para suas necessidades e diferentes aspectos que compõem seu cotidiano20. Percepção em relação à humanização dos serviços oferecidos Humanização na área da saúde requer ainda um processo reflexivo acerca dos valores e princípios que norteiam esta prática profissional, pressupondo uma nova postura ética diante de todo processo de trabalho em saúde16. Assim, a organização das práticas de saúde e o campo da produção do cuidado possibilitam a forma efetiva e criativa de manifestação da subjetividade do outro, a partir dos dispositivos de acolhimento, vínculo e responsabilização contidos nessa organização da assistência à saúde17. Emergiram da fala dos sujeitos pontos como a interatividade, sensibilidade, acolhimento e carinho. [...] Sim, pois as pessoas se preocupam e também estão mais perto, tanto os funcionários como a direção, e tem correspondido as suas necessidades, sempre com interação. (Sujeito 1). [...] Sim, humanizado, não tem nada para reclamar, isso é humanizado, essas atitudes dos trabalhadores. (Sujeito 3). [...] Todo mundo fala de melhora, todo mundo fala de melhora. Ninguém fala tô pior depois que fiz isso, ninguém fala. (Sujeito 6). Na lógica de outros sujeitos, os serviços se apresentam aquém da efetiva humanização: Na lógica de outros sujeitos, os serviços se apresentam aquém da efetiva humanização: Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 5/85 O impacto das assembleias na humanização em saúde. [...] O serviço é em partes humanizado, pois ainda tinha que treinar alguns funcionários, técnicos sobre o tratamento, a preparação na verdade de como trabalhar com pessoas com deficiência e saúde mental de uma maneira geral. (Sujeito 4). [...] Não, agora não, antes, antes era, como eu falei, a gente era tratado como ser humano, hoje em dia eu já não acho. Não são todos que são assim, né? (Sujeito 5). [...] Porque ... na minha opinião, eu acho que falta bem mais pessoas que se importem com a gente. Eu acho que nada vai mudar. (Sujeito 8). Sem comunicação não há humanização, o atendimento não pode ser apenas mecânico; saber ouvir coletando corretamente os dados do paciente pode ser uma das melhores formas de tratamento18. O bom relacionamento entre paciente e profissional, boa comunicação e empatia são pontos fundamentais para o processo de humanização19. Os sujeitos também identificaram a humanização e satisfação em relação à atuação dos profissionais relatando empenho e motivação pelo trabalho. [...] Sim, são todos muito humanos. (Sujeito 1). [...] Sim, são todos muito humanos. (Sujeito 1). [...] Está muito motivado. (Sujeito 3). Mudanças observadas após as assembleias As assembleias possibilitam a inclusão dos pacientes na gestão institucional, ampliando a corresponsabilidade dos sujeitos no tocante ao tratamento que lhe é dispensado e na utilização desses espaços, o que remonta a um processo de horizontalização das relações, como requerido no modelo de reabilitação psicossocial9. Além disso, são um espaço de convivência e discussão de questões referentes ao serviço, sendo uma das modalidades oferecidas entre os recursos terapêuticos, promovendo a transformação nos modos de relação e de comunicação entre os sujeitos21. j Assim, os sujeitos relataram o surgimento de novas oficinas, evolução no tratamento e melhoria no atendimento como fatores de impacto após as assembleias. [...] Começa pela minha melhora, a orientadora da equipe nas assembleias faz com que eu tome a medicação em dia, minha vida voltou a ter graça e passei a ter vontade de trabalhar. (Sujeito 3). [...] A efetivação da própria assembleia em si, ela se tornou mais plausível, ela se tornou mais edificante, se tornou mais presente na nossa rotina, e o propósito dela em si vem sendo melhorado. (Sujeito 4). [...] Sim, o que mais melhorou mesmo foram as novas oficinas, incluídas a horta, o artesanato, e me lembro que surgiram a partir de ideias na própria assembleia. (Sujeito 9). A Politica Nacional de Humanização se compromete em possibilitar atenção integral à população e a propor estratégias que possibilitem ampliar as condições de direitos e de cidadania, além de promover o aprimoramento das relações entre o profissional de saúde e o usuário, dos profissionais entre si e com a comunidade22. A humanização também pode ser considerada como estratégia de qualificação da atenção e gestão do trabalho, busca a transformações no âmbito da produção dos serviços, visando Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 66/8 O impacto das assembleias na humanização em saúde. mudanças nos processos, organização, resolubilidade e qualidade, além da produção de sujeitos com vistas na mobilização, crescimento e, principalmente, autonomia dos trabalhadores e usuários13. CONSIDERAÇÕES FINAIS A investigação sobre os impactos das assembleias na humanização no CAPS II de Boa Vista (RR) foi explorada em um contexto de harmonização entre profissionais e usuários com ênfase no trabalho humanitário desenvolvido pela equipe. Ao considerar a realidade dos sujeitos da pesquisa, acreditamos que foi possível estabelecer uma relação dialógica e horizontalizada entre pesquisador e sujeitos, possibilitando a identificação de suas percepções e indagações, o que contribui para que os serviços reflitam sobre o seu papel diante desses sujeitos. Na visão dos participantes, as assembleias apresentam grande importância, por se tratar de um espaço onde emergem discussões sobre as melhorias dos pacientes, desenvolvimento da saúde, possibilidade de opinar em relação às benfeitorias, ou ao que deverá ser melhorado, possibilitando o livre-arbítrio da fala e contribuindo com a humanização do serviço. No entanto, o serviços ainda apresentam fragilidades que podem interferir na missão final de proporcionar o vínculo e o cuidado integral aos usuários, substituindo suas práticas tradicionais e hospitalocêntricas pela ressocialização dos indivíduos não só no âmbito do seio familiar, como também na sociedade. Isso porque, mesmo diante de questões ainda dependentes de resolução, a participação em assembleias por parte desses usuários mostrou um impacto exponencial no que tange à efetivação das práticas de humanização em saúde mental. REFERÊNCIAS 1. Ferreira JT, Mesquita NNM, Silva TA, Silva VF, Lucas WJ, Batista EC. Centro de Atenção Psicossocial (CAPS): uma instituição de referência no atendimento à saúde mental. Rev. Saberes. 2016;4(1):72-86. 2. Onocko-Campos RT, Furtado JP. Entre a saúde coletiva e a saúde mental: um instrumento metodológico para avaliação da rede de Centro de Atenção Psicossocial (CAPS) do Sistema Único de Saúde. Cad Saude Publica. 2006;22(5):1053-62. http://dx.doi.org/10.1590/S0102-311X2006000500018. PMid:16680358. 3. Brasil. Ministério da Saúde. Secretaria-Executiva. Núcleo Técnico da Política Nacional de Humanização. Política Nacional de Humanização: a humanização como eixo norteador das práticas de atenção e gestão em todas as instâncias do SUS. Brasília: Ministério da Saúde; 2004. 4. Deslandes SF. Análise do discurso oficial sobre a humanização da assistência hospitalar. Cien Saude Colet. 2004;9(1):7-14. http://dx.doi.org/10.1590/S1413-81232004000100002. 5. Brasil. Ministério da Saúde. Secretaria-Executiva. Núcleo Técnico da Política Nacional de Humanização. Humaniza SUS: Política Nacional de Humanização: a humanização como eixo norteador das práticas de atenção e gestão em todas as instâncias do SUS. Brasília: Ministério da Saúde; 2004. 6. Bontempo VL. A assembleia de usuários e o CAPSI. Psicologia. 2009;29(1):184-9. http://dx.doi.org/10.1590/ S1414-98932009000100015. 7. Instituto Brasileiro de Geografia e Estatística. Total população de Roraima [Internet]. Brasília: IBGE; 2010 [citado em 2017 abr 23]. Disponível em: https://www.ibge.gov.br/home/estatistica/populacao/censo2010/ tabelas_pdf/total_populacao_roraima.pdf 8. Bardin L. Análise de conteúdo. Lisboa: Edições 70; 2008. 9. Triviños ANS. Introdução à pesquisa em ciências sociais: a pesquisa qualitativa em educação. São Paulo: Atlas; 1987. 10. Pedroso R, Vieira MEM. Humanização das práticas de saúde: transversalizar em defesa da vida. Interface Comunicacao Saude Educ. 2009;13(Supl. 1):695-700. http://dx.doi.org/10.1590/S1414- 32832009000500020. 11. Junqueira AMG, Carniel IC, Mantovani A. assembleias como possibilidades de cuidado em saúde mental em um CAPS. Vínculo. 2015;12(1):31-40. 12. Detomini VC, Bellenzani R. Construindo a participação social junto a usuários de um grupo de apoio: desafios para a qualificação da atenção em um Centro de Atenção Psicossocial (CAPS). Cad Ter Ocup UFSCar. 2015;23(3):661-72. http://dx.doi.org/10.4322/0104-4931.ctoRE0578. Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 7/87 O impacto das assembleias na humanização em saúde. 13. Costa DFC, Paulon SM. Participação Social e protagonismo em saúde mental: a insurgência de um coletivo social participation and protagonism in mental health: the rising of a collective. Saúde Debate. 2012;36(95):572-82. http://dx.doi.org/10.1590/S0103-11042012000400009. 14. Pessini L. Humanização da dor e sofrimento humano no contexto hospitalar. Bioética. 2002;10(2):51-72. 15. Brasil. Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Ações Programáticas Estratégicas. Saúde mental no SUS: os centros de atenção psicossocial. REFERÊNCIAS Brasília: Ministério da Saúde; 2004. 16. Fountoulakis KN, Kasper S, Andreassen O, Blier P, Okasha A, Severus E, et al. Efficacy of pharmacotherapy in bipolar disorder: a report by the WPA section on pharmacopsychiatry. Eur Arch Psychiatry Clin Neurosci. 2012 Jun;262(Supl. 1):1-48. http://dx.doi.org/10.1007/s00406-012-0323-x. PMid:22622948. 17. Ayres JRCM. Cuidado e reconstrução das práticas de saúde. Interface Comunicacao Saude Educ 2004;8(14):73-92. http://dx.doi.org/10.1590/S1414-32832004000100005. 18. Pontes AC, Leitão IMTA, Ramos IC. Comunicação terapêutica em Enfermagem: instrumento essencial do cuidado. Rev Bras Enferm. 2008;61(3):312-8. http://dx.doi.org/10.1590/S0034-71672008000300006. PMid:18604425. 19. Souza CNM, Passarela BAC, Cassuli MMC. Humanização do cuidado de enfermagem: o que é isso? Rev Bras Enferm. 2009;62(3):349-54. http://dx.doi.org/10.1590/S0034-71672009000300003. PMid:19597655. 20. Maynart WH, Albuquerque MC, Brêda MZ, Jorge JS. A escuta qualificada e o acolhimento na atenção psicossocial. Acta Paul Enferm. 2014;27(4):300-3. http://dx.doi.org/10.1590/1982-0194201400051. 21. Archanjo JVL, Barros MEB. Política nacional de humanização: desafios de se construir uma política dispositivo [Internet]. 2008 [citado em 2017 mar 1]. Disponível em: https://www.google.com.br/search? client=safari&rls=en&q=Archanjo+JVL,+Barros+MEB.+Pol%C3%ADtica+nacional+de+humaniza%C3%A 7%C3%A3o:+desafios+de+se+construir+uma+pol%C3%ADtica+dispositivo,+2008.&ie=UTF-8&oe=UTF- 8&gws_rd=cr&dcr=0&ei=lYSZWuLND8OYzwKBu7PQCA 22. Camponogara S, Santos TM, Seiffert MA, Alves CN O cuidado humanizado em unidade de terapia intensiva: uma revisão bibliográfica. Rev Enferm UFSM. 2011;1(1):124-32. http://dx.doi.org/10.5902/217976922237. Cad. Saúde Colet., 2023; 31(1):e31010210 \| https://doi.org/10.1590/1414-462X202331010210 88/8
https://openalex.org/W2327314317	https://europepmc.org/articles/pmc5041971?pdf=render	English	null	Slug inhibits the proliferation and tumor formation of human cervical cancer cells by up-regulating the p21/p27 proteins and down-regulating the activity of the Wnt/β-catenin signaling pathway via the trans-suppression Akt1/p-Akt1 expression	Oncotarget	2,016	cc-by	11,071	ABSTRACT Slug (Snai2) has been demonstrated to act as an oncogene or tumor suppressor in different human cancers, but the function of Slug in cervical cancer remains poorly understood. In this study, we demonstrated that Slug could suppress the proliferation of cervical cancer cells in vitro and tumor formation in vivo. Further experiments found that Slug could trans-suppress the expression of Akt1/p-Akt1 by binding to E-box motifs in the promoter of the Akt1 gene and then inhibit the cell proliferation and tumor formation of cervical cancer cells by up-regulating p21/p27 and/or down- regulating the activity of the Wnt/β-catenin signaling pathway. Therefore, Slug acts as a tumor suppressor during cervical carcinogenesis. Nan Cui1,2, Wen-Ting Yang1,3, Peng-Sheng Zheng1,3 1Department of Reproductive Medicine, First Affiliated Hospital, Xi’an Jiaotong University Medical School, Xi’an, The People’s Republic of China 3Section of Cancer Stem Cell Research, Key Laboratory of Environment and Genes Related to Diseases, Ministry of Education of The People's Republic of China, Xi'an, The People's Republic of China Section of Cancer Stem Cell Research, Key Laboratory of Environment and Genes Related to Disease of The People's Republic of China, Xi'an, The People's Republic of China Correspondence to: Peng-Sheng Zheng, email: zpsheng@mail.xjtu.edu.cn Correspondence to: Peng-Sheng Zheng, email: zpsheng@mail.xjtu.edu.cn Correspondence to: Peng-Sheng Zheng, email: zpsheng@mail.xjtu.edu.cn Correspondence to: Peng-Sheng Zheng, email: zpsheng@mail.xjtu.edu.cn Keywords: Slug, Akt1, Wnt/β-catenin, cervical cancer, proliferation Received: January 05, 2016 Accepted: March 14, 2016 Published: March 28, 2016 Correspondence to: Peng-Sheng Zheng, email: zpsheng@mail.xjtu.edu.cn Keywords: Slug, Akt1, Wnt/β-catenin, cervical cancer, proliferation Received: January 05, 2016 Accepted: March 14, 2016 Published: March 28, 2016 Published: March 28, 2016 Oncotarget, Vol. 7, No. 18 Oncotarget, Vol. 7, No. 18 www.impactjournals.com/oncotarget/ INTRODUCTION accumulating evidence showed that some stem cell self-renewal-associated transcription factors are involved in tumorigenesis and tumor development in cervical cancer; for example, NANOG, OCT4, Msi1 and LGR5 are reported to promote the progression of cervical cancer [9–12]; in contrast, UTF1 and KLF4 are reported to function as tumor suppressors in cervical cancer [13, 14]. In addition, SOX2 was reported to enhance tumor formation ability and to serve as a nuclear marker for cervical cancer stem cells [15], and ALDH1 was reported to be a cytoplasmic maker for cervical cancer stem cells [16]. Globally, cervical carcinoma is the third most common tumor type and the fourth most common cause of cancer death among women [1]. In developing countries, cervical carcinoma is the second most common gynecological cancer and the second most common cause of cancer death among women [2]. Nearly 99.7% of cervical cancer cases are associated with the human papillomavirus (HPV) [3–5]. However, the molecular and genetic mechanisms involved in the initiation and progression of cervical cancer remain poorly understood. Previous studies observed that various oncogenes and cancer suppressor genes exhibit abnormal expression during the development and progression of cervical cancer. For example, the oncogene R-Ras is reported to play a central role during the progression of cervical cancer [6]. p53 is known to function as a tumor suppressor in various tumors, but p53 polymorphisms were reported to be associated with an increased risk of cervical cancer [7, 8]. Recently, Slug is a member of the Snail superfamily, and evidence indicates that Slug is associated with cell pluripotency [17, 18]. When Slug and Sox9 are coexpressed in mammary stem cells, these proteins help to maintain the mammary stem cell population [19]. In addition, during mammary gland morphogenesis, Slug controls the growth dynamics of stem/progenitor cells as a gate-keeper [20]. Early studies suggested that Slug is implicated in the development of chick limbs and www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26152 involved in the early patterning of the mesoderm and the neural crest [21]. On the other hand, the expression of Slug is an initial and necessary step for epithelial to mesenchymal transition (EMT) and plays a critical role during the EMT program [22–24]. As a transcriptional repressor, Slug regulates the expression of target genes by binding to E-box elements in the promoter regions of genes. Slug inhibits the proliferation of cervical carcinoma cells in vitro Moreover, the expression of Slug was detected using immunocytochemistry and western blotting in cervical cancer cell lines. A high level of Slug expression was detected in the HeLa, CasKi and HT-3 cervical carcinoma cell lines, but almost no expression of the Slug protein was detected in SiHa and C33A cells (Figure 2A and 2B). To further investigate the function of Slug in human cervical cancer cells, exogenous Slug was stably overexpressed in SiHa (SiHa-Slug, Figure 2C) and C33A (C33A-Slug, Figure 2F) cells; conversely, the expression of Slug was knocked down in HeLa Figure HeLa-shSlug, Figure 2I) and CasKi (CasKi-shSlug, Figure 2L) cells by stably transfecting shRNA plasmids. Western blotting was used to confirm the effects of the up-regulation and down- regulation of Slug expression in all cervical cancer cells (SiHa, C33A, HeLa and CasKi) and their controls (Figure 2C, 2F, 2I and 2L, respectively). However, the function of Slug in cervical cancers is poorly understood. Early studies showed that Slug exerts contrasting effects on cell proliferation and tumor growth among various cancers; for example, Slug promotes cell proliferation in lung cancer cells and glioblastoma cells [35, 36] but inhibits cell proliferation in human epidermal keratinocytes and human prostate cancer cells [37, 38]. This study reported for the first time that Slug could inhibit proliferation and tumor formation in human cervical cancer cells by up-regulating p21/p27 and/or down-regulating cyclin D1 via the trans-suppression of Akt1/p-Akt1. p y) Cell growth curves and the MTT assay were used to determine the cell proliferation ability and cell viability of the Slug-modified cervical cancer cell lines and their control cells. As shown in Figure 2D and 2G, the SiHa- Slug and C33A-Slug cells grew much more slowly than their respective control cells (SiHa-GFP and C33A- GFP, p < 0.01). In addition, the viability of SiHa-Slug and C33A-Slug cells was also much lower than that of their respective control cells (SiHa-GFP and C33A- GFP) (Figure 2E and 2H; p < 0.01), suggesting that the Slug protein may suppress the proliferation of cervical cancer cells. Furthermore, both cell growth curves and cell viability assays found that HeLa-shSlug and CasKi- shSlug cells grow much faster than their respective control cells (HeLa-shcontrol and Caski-shcontrol) (Figure 2J, 2M, Figure 2K and 2N; p < 0.01), suggesting that the knockdown of Slug promoted the proliferation of cervical cancer cells. The expression of slug in the normal human cervix and different cancerous cervical lesions To investigate whether Slug is involved in the development and progression of human cervical carcinoma, Slug expression was detected in normal human cervix (NC), cervical cancer in situ (CIS) and invasive cervical cancer (SCC) samples using immunohistochemistry. Representative Slug staining in the normal cervix and cancerous cervical lesions are shown in Figure 1A. The Slug protein was found to localize to the nucleus. The Slug-positive rate decreased from NC samples (86.84%) to CIS samples (62.50%) and SCC samples (59.62%). Significant differences were observed between CIS and NC samples and between SCC and NC samples, but no significant difference was observed between CIS and SCC samples (Table S1 and Figure 1B, p < 0.05). The immunoreactivity scores were also lower in CIS and SCC samples than in NC samples (Figure 1C, CIS vs. NC, P < 0.05; SCC vs. NC, p < 0.01), but there was no significant difference between the CIS and SCC samples (Figure 1C), suggesting that Slug is involved in the development of cervical carcinoma. Additionally, INTRODUCTION For instance, Slug can induce the reduction of E-cadherin in various cancers [25–27], and this transcription repression capacity was stabilized by p19Arf in mouse prostate cancer models [28]. In addition, the down-regulation of the cell adhesion molecule E-cadherin is also one of the most important changes for EMT [29], and E-cadherin is also associated with the self-renewal of human embryonic stem cells (hESC) [30, 31]. Moreover, as an antiapoptotic factor, Slug is also beneficial for cell survival [32–34]. western blotting was used quantitatively to detect the expression of Slug in 8 normal cervix samples and 8 cervical carcinoma samples (Figure 1D). The average Slug expression level was lower in cervical carcinoma tissues than in normal cervix tissues (Figure 1E; P < 0.01), further confirming that Slug expression is negatively related to cervical carcinogenesis. Slug inhibits the proliferation of cervical carcinoma cells in vitro All of these results demonstrated that the Slug protein inhibited the proliferation of cervical carcinoma cells in vitro. Slug suppresses the growth and tumor formation of cervical cancer cells in vivo To identify the effect of Slug on cervical cancer cells in vivo, 106 Slug-modified cells and their control cells were inoculated subcutaneously into female nude www.impactjournals.com/oncotarget Oncotarget 26153 day (Figure 3A), suggesting that the Slug protein may delay the initiation of cervical carcinoma. The volume of the tumors formed by the SiHa-Slug cells was much smaller than that of the tumors formed by the SiHa- GFP control cells (Figure 3A, p < 0.05). In addition, the average weight of the tumors formed by the SiHa-Slug cells was much smaller than that of the tumors formed mice for the tumor formation assay. Both SiHa and HeLa cells formed xenografted tumors successfully. However, the C33A and CasKi cells failed to form palpable tumors in female nude mice. As shown in Figure 3A, the palpable tumors formed by SiHa-GFP control cells could be found as early as the 23rd day, but the tumors formed by the SiHa-Slug cell group could not be found until the 40th Figure 1: Expression of slug in normal cervix samples and various cervical lesions. (A) Immunohistochemical (IHC) detection of Slug in normal cervix, cancer in situ and carcinoma samples; original magnification, 1000×. (B) Slug staining is classified into 2 categories (negative and positive), and the bar chart shows the percentage of each group (38 normal cervix specimens, 24 carcinoma in situ specimens, and 52 invasion carcinoma tissue specimens). (C) The scatter plot shows the immunoreactivity scores (IHC) obtained for the staining of Slug in normal cervix, cervical cancer in situ and invasive cervical cancer samples (points represent the IHC score per specimen, and one-way ANOVA was performed). (D) The expression of Slug in normal cervix (NC) and squamous cervical carcinoma (SCC) samples was detected using western blotting. (E) The relative expression of Slug in each normal cervix tissue (n = 8) and squamous cervical carcinoma tissue sample (n = 8) is shown. The data shown are the ratios of the Slug/β-actin of each specimen and the means ± standard error of the NC and SCC groups (triangles represent relative Slug expression). Values are shown as the mean ± SD, p < 0.05, p < 0.01. Figure 1: Expression of slug in normal cervix samples and various cervical lesions. (A) Immunohistochemical (IHC) detection of Slug in normal cervix, cancer in situ and carcinoma samples; original magnification, 1000×. Slug suppresses the growth and tumor formation of cervical cancer cells in vivo (B) Slug staining is classified into 2 categories (negative and positive), and the bar chart shows the percentage of each group (38 normal cervix specimens, 24 carcinoma in situ specimens, and 52 invasion carcinoma tissue specimens). (C) The scatter plot shows the immunoreactivity scores (IHC) obtained for the staining of Slug in normal cervix, cervical cancer in situ and invasive cervical cancer samples (points represent the IHC score per specimen, and one-way ANOVA was performed). (D) The expression of Slug in normal cervix (NC) and squamous cervical carcinoma (SCC) samples was detected using western blotting. (E) The relative expression of Slug in each normal cervix tissue (n = 8) and squamous cervical carcinoma tissue sample (n = 8) is shown. The data shown are the ratios of the Slug/β-actin of each specimen and the means ± standard error of the NC and SCC groups (triangles represent relative Slug expression). Values are shown as the mean ± SD, p < 0.05, *p < 0.01. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26154 gure 2: Slug inhibits the proliferation of human cervical cancer cell lines in vitro. Slug expression in human cer ncer cell lines was detected using immunocychemistry (A) and western blotting (B). Stably transfected cell lines were identifie stern blotting: (C) SiHa-GFP and SiHa-Slug cells; (F) C33A-GFP and C33A-Slug cells; (I) HeLa-shControl and HeLa-shSlug cells ) Caski-shControl and Caski- shSlug cells. The proliferation and viability of SiHa-GFP and SiHa-Slug cells were detected using gr rves (D) and the MTT assay (E). The proliferation and viability of C33A-GFP and C33A-Slug cells were detected using growth cu ) and the MTT assay (H). The proliferation and viability of HeLa-shControl and HeLa- shSlug were detected using growth c and the MTT assay (K). The proliferation and viability of Caski-shControl and Caski-shSlug cells were detected using growth c M) and the MTT assay (N). The data were shown as the mean ± SD of three independent experiments. p < 0.05, *p < 0.01 vs. co ng One-Way ANOVA. Figure 2: Slug inhibits the proliferation of human cervical cancer cell lines in vitro. Slug expression in human cervical cancer cell lines was detected using immunocychemistry (A) and western blotting (B). Stably transfected cell lines were identified by western blotting: (C) SiHa-GFP and SiHa-Slug cells; (F) C33A-GFP and C33A-Slug cells; (I) HeLa-shControl and HeLa-shSlug cells; and (L) Caski-shControl and Caski- shSlug cells. The proliferation and viability of SiHa-GFP and SiHa-Slug cells were detected using growth curves (D) and the MTT assay (E). The proliferation and viability of C33A-GFP and C33A-Slug cells were detected using growth curves (G) and the MTT assay (H). The proliferation and viability of HeLa-shControl and HeLa- shSlug were detected using growth curves (J) and the MTT assay (K). The proliferation and viability of Caski-shControl and Caski-shSlug cells were detected using growth curves (M) and the MTT assay (N). The data were shown as the mean ± SD of three independent experiments. p < 0.05, *p < 0.01 vs. control using One-Way ANOVA. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26155 transition in the SiHa-Slug and C33A-Slug cells. Slug suppressed the proliferation of cervical cancer cells by up-regulating the p21/p27 proteins and down-regulating Wnt/β-catenin pathway activity via the trans-suppression of Akt1/p-Akt1 It has been demonstrated that Slug can be trans- activated through the PI3K/Akt1 signaling pathway in various cancers [40–43]. However, to our knowledge, no available reports determined whether Slug could regulate the expression of Akt1. Akt1 is a very important protein that strongly affects the cell cycle [44–46]. Therefore, the expression of Akt1/p-Akt1 was detected in the Slug- modified cervical cancer cells and their controls by western blotting. As shown in Figure 5, the Akt1 and p-Akt1 proteins were expressed at much lower levels in both SiHa-Slug (Figure 5A and 5C, P < 0.05) and C33A- Slug cells (Figure 5B and 5D, P < 0.05) than in SiHa-GFP and C33A-GFP cells, respectively. Furthermore, the Akt1 and p-Akt1 proteins were expressed at much higher levels in the HeLa-shSlug (Figure 5E and 5G, P < 0.05) and CasKi-shSlug cells (Figure 5F and 5H, P < 0.05) than in HeLa-shControl and CasKi-shControl cells, respectively. All of these data indicated that Slug could trans-suppress the expression of Akt1/p-Akt1 in cervical cancer cell lines. www.impactjournals.com/oncotarget Conversely, the knockdown of Slug led to a decrease in the G1/S ratio in both HeLa-shSlug (42.99%/40.80%, 1.05; Figure 4G, 4H and 4I) and CasKi-shSlug cells (48.70%/29.75%, 1.64; Figure 4J, 4K and 4L) in comparison to the ratios observed in the HeLa-control cells (56.21%/31.58%, 1.78; Figure 4G, 4H and 4I) and the CasKi-control cells (60.71%/19.13%, 3.17; Figure 4J, 4K and 4L). This result suggested that the knockdown of Slug promoted G1/S phase transition in both HeLa-shSlug and CasKi-shSlug cells. All of these results indicated that Slug inhibited cervical cancer cells at the transition from the G0/G1 phase to the S phase of the cell cycle. by the SiHa-GFP control cells (Figure 3B, p < 0.05), indicating that the over-expression of the Slug protein could suppress tumor initiation and the development of the SiHa cervical cancer cell line in vivo. Furthermore, the HeLa-shSlug cells could xenograft the tumors earlier (11th day in the HeLa-shSlug cell group and 13th day in the HeLa-shcontrol group) and could develop much larger (Figure 3C, p < 0.05) and heavier tumors (Figure 3D, p < 0.01) than the HeLa-shcontrol cells, indicating that the knockdown of Slug in HeLa cells could enhance tumor formation in vivo. All of these results demonstrated that the Slug protein could inhibit tumor formation by cervical cancer cells in vivo. To determine whether the in vivo tumor suppression function of Slug could be attributed to its cell proliferation inhibition ability, immunohistochemistry was used to determine the expression of Slug and the cell proliferation marker Ki67 [39] in the xenografted cervical cancer tissues. As shown in Figure 3E and 3F, the tumor tissues derived from SiHa-Slug cells expressed much more Slug and less Ki67 than the tumor tissues derived from SiHa-GFP control cells. In addition, the tumor tissues derived from HeLa- shSlug cells expressed less Slug and much more Ki67 than the tumor tissues derived from HeLa-shcontrol cells (Figure 3G and 3H). These results indicated that the expression of Slug adversely affects the cell proliferative ability of cervical cancer cells in vivo. These results are consistent with the results obtained from the in vitro experiment in this study, suggesting that Slug affects tumor formation by cervical cancer cells in vivo in a manner that is dependent on its effects on cell proliferation. Slug arrests cervical cancer cells at the transition from the G0/G1 phase to the S phase of the cell cycle Generally, the changes that occur during cell proliferation involve the modulation of the cell cycle. To investigate how Slug affects the cell cycle of cervical cancer cells, fluorescence-activated cells sorting (FACS) was used to analyze the differences in the cell cycle between the Slug-modified cells and their control cervical cancer cells. As shown in (Figure 4A, 4B and 4C), the percentage of cells in G0/G1 phase was much higher in the SiHa-Slug cells (60.33%) than in the SiHa-GFP control cells (42.64%), and the percentage of cells in S phase was lower in the SiHa-Slug cells (24.79%) than in the SiHa-GFP control cells (32.20%). The ratio of cells in G1/S phase was much higher in the SiHa-Slug cells (60.33%/24.79%, 2.43) than in the SiHa-GFP cells (42.64%/32.20%, 1.32). A similar result was observed in the C33A cells, and the ratio of cells in the G1/S phase (56.38%/29.28%, 1.93) was much higher in the C33A- Slug cells than in the C33A-GFP cells (40.27%/43.92%, 0.92). These results suggested that the over-expression of Slug induced cell cycle arrest during the G1/S phase Akt1 has been demonstrated to be a negative regulator of the cell cycle suppressors p21 and p27 [47–49]. Thus, western blotting analysis was also used to detect the expression of p21 and p27 in the Slug-modified cervical cancer cell lines and their control cells. As shown in Figure 5, the protein levels of p21 and p27 were much higher in the SiHa-Slug (Figure 5A and 5C, P < 0.05) and C33A-Slug cells (Figure 5B and 5D, P < 0.05) than in the SiHa-GFP cells and C33A-GFP cells, respectively. Furthermore, the expression levels of p21 and p27 were much lower in the HeLa-shSlug (Figure 5E, and 5G, P < 0.05) and CasKi-shSlug cells (Figure 5F and 5H, P < 0.01) than in the HeLa-shControl and CasKi-shControl cells, respectively. All of these results suggested that Slug could inhibit the proliferation of cervical cancer cells by up-regulating the p21/p27 proteins by trans-suppressing the expression of the Akt1/p-Akt1 proteins. www.impactjournals.com/oncotarget Oncotarget 26156 It has been reported that GSK3β can be phosphorylated through the PI3K/Akt pathway, and GSK3β is a commonly investigated molecular intersection between the Wnt/β-catenin and PI3K/Akt signaling pathways [50, 51]. Slug arrests cervical cancer cells at the transition from the G0/G1 phase to the S phase of the cell cycle To determine whether the Wnt/β- catenin signaling pathway could be affected by the PI3K/ Akt signaling pathway in cervical cancer cell lines, the TOP-Flash reporter assay was used to identify the activity of the Wnt/β-catenin signaling pathway in Slug-modified cervical cancer cells. As shown in Figure 5, the values obtained in the TOP/FOP assay were much lower in the SiHa-Slug Figure 5I) and C33A-Slug cells (Figure 5G) than in the SiHa-GFP and C33A-GFP cells, respectively. Similarly, the values obtained in the TOP/FOP assay were TOP-Flash reporter assay was used to identify the activity of the Wnt/β-catenin signaling pathway in Slug-modified cervical cancer cells. As shown in Figure 5, the values obtained in the TOP/FOP assay were much lower in the SiHa-Slug Figure 5I) and C33A-Slug cells (Figure 5G) than in the SiHa-GFP and C33A-GFP cells, respectively. Similarly, the values obtained in the TOP/FOP assay were Oncotarget 26157 mpactjournals.com/oncotarget re 3: Slug suppressed cervical carcinoma tumor growth in vivo. Tumor growth curves were calculated after injection emale nude mice based on monitoring performed every 3 days: (A) SiHa-GFP and SiHa-Slug cells; (C) HeLa-shControl and HeLa- g cells. The xenograft tumors were dissociated and weighed at the end of experiment: (B) SiHa-GFP and SiHa-Slug cells; (D) -shControl and HeLa- shSlug cells. Immunohistochemical staining of Slug and Ki-67 in xenograft tumor tissues: (E and F) SiHa- and SiHa-Slug cells; (G and H) HeLa-shControl and HeLa-shSlug cells. Values are shown as the mean ± SD. p < 0.05, *p < 0.01, < 0.001 vs. control using One-Way ANOVA. Figure 3: Slug suppressed cervical carcinoma tumor growth in vivo. Tumor growth curves were calculated after injection into female nude mice based on monitoring performed every 3 days: (A) SiHa-GFP and SiHa-Slug cells; (C) HeLa-shControl and HeLa- shSlug cells. The xenograft tumors were dissociated and weighed at the end of experiment: (B) SiHa-GFP and SiHa-Slug cells; (D) HeLa-shControl and HeLa- shSlug cells. Immunohistochemical staining of Slug and Ki-67 in xenograft tumor tissues: (E and F) SiHa- GFP and SiHa-Slug cells; (G and H) HeLa-shControl and HeLa-shSlug cells. Values are shown as the mean ± SD. p < 0.05, p < 0.01, p < 0.001 vs. control using One-Way ANOVA. www.impactjournals.com/oncotarget Furthermore, the protein levels of p-GSK3β, β-catenin, cyclin D1 and c-myc were much lower in the SiHa- Slug cell than in the SiHa-GFP cells (Figure 5A and 5C, P < 0.05). However, there was no significant difference in GSK3β protein levels between the Slug-modified cells and their control cervical cancer cells. All of these results suggested that Slug negatively regulates the activity of the Wnt/β-catenin signaling pathway in these cervical cancer cell lines. Similarly, the protein levels of p-GSK3β, cyclin D1 and c-myc were also found to be much lower in the C33A-Slug cells than in the C33A-GFP cells (Figure 5B and 5D, P < 0.05). However, the protein level of β-catenin was much higher in the C33A-Slug cells than in the C33A-GFP cells (Figure 5B and 5D, P < 0.05). Further immunohistochemistry analysis showed that β-catenin accumulated in the cytomembrane but not in the cell nucleus in C33A-Slug cells (Figure S1A). This finding may explain why the C33A-Slug cells exhibited higher protein levels of β-catenin but lower activities of the Wnt/β-catenin signaling pathway in the TOP-Flash reporter assay. All of these results further confirmed that the activity of the Wnt/β-catenin signaling pathway was suppressed in all Slug-expressing cervical cancer cell lines. Cyclin D1 is a key factor that contributes to the cell cycle transition from G1 to S phase [55] and is also an effector of the Wnt/β-catenin signaling pathway [56]. Therefore, in the Slug-modified cervical cancer cells, cell proliferation and tumor formation should be affected by the cyclin D1 protein through the Wnt/β-catenin signaling pathway via the trans-suppression of the expression of the Akt1/p-Akt1 proteins. To further confirm that Slug inhibited cell proliferation and tumor formation of cervical cancer cells by trans-suppressing Akt1/p-Akt1 through both the p21/ p27 proteins and the Wnt/β-catenin signaling pathway, first of all, the recombinant Akt1 plasmid was transiently transfected in the SiHa-Slug and HeLa-shControl cells, in which high level of Slug expression trans-suppressed the expression of Akt1/p-Akt1. As shown in Figure 6, the SiHa-Slug and HeLa-shControl cells transfected by pIRES2-AcGFP-Akt1 grew much faster than the cells transfected with the control vector, respectively (Figure 6C, 6D, 6G and 6H, P < 0.05). www.impactjournals.com/oncotarget Cyclin D1 is a key factor that contributes to the cell cycle transition from G1 to S phase [55] and is also an effector of the Wnt/β-catenin signaling pathway [56]. Therefore, in the Slug-modified cervical cancer cells, cell proliferation and tumor formation should be affected by the cyclin D1 protein through the Wnt/β-catenin signaling pathway via the trans-suppression of the expression of the Akt1/p-Akt1 proteins. www.impactjournals.com/oncotarget Simultaneously, the SiHa-Slug and HeLa-shControl cells transfected by pIRES2-AcGFP-Akt1 also express more Akt1/p-Akt1, less p21/p27, more p-Rb and more molecules (p-GSK3β, β-catenin, c-myc and cyclin D1) of the Wnt/β-catenin signaling pathway than the cells transfected with the control vector, respectively (Figure 6A, 6B, 6E, and 6F, P < 0.05). Secondly, MK-2206 (MK-2206), which is an inhibitor of Akt1 [58], was used in all of the Slug-modified the cervical cancer cells investigated in this study. All of the cervical cancer cells subjected to the MK treatment grew much more slowly and expressed less Akt1/p-Akt1, more p21/p27, less p-Rb and fewer molecules (p-GSK3β, β-catenin, c-myc and cyclin D1) of the Wnt/β-catenin signaling pathway than the cells that were not subjected to MK treatment (Figure S2). All of these results further confirmed that Slug inhibited cell proliferation and tumor formation of human cervical cancer cells via trans-suppressing Akt1/p-Akt1 through both the p21/p27 proteins and the Wnt/β-catenin signaling pathway. cyclin D1 and c-myc were much higher in the HeLa- shSlug (Figure 5E and 5G, P < 0.05) and CasKi-shSlug cells (Figure 5F and 5H, P < 0.01) than in the HeLa- shControl and CasKi-shControl cells, respectively. Furthermore, the protein levels of p-GSK3β, β-catenin, cyclin D1 and c-myc were much lower in the SiHa- Slug cell than in the SiHa-GFP cells (Figure 5A and 5C, P < 0.05). However, there was no significant difference in GSK3β protein levels between the Slug-modified cells and their control cervical cancer cells. All of these results suggested that Slug negatively regulates the activity of the Wnt/β-catenin signaling pathway in these cervical cancer cell lines. Similarly, the protein levels of p-GSK3β, cyclin D1 and c-myc were also found to be much lower in the C33A-Slug cells than in the C33A-GFP cells (Figure 5B and 5D, P < 0.05). However, the protein level of β-catenin was much higher in the C33A-Slug cells than in the C33A-GFP cells (Figure 5B and 5D, P < 0.05). Further immunohistochemistry analysis showed that β-catenin accumulated in the cytomembrane but not in the cell nucleus in C33A-Slug cells (Figure S1A). This finding may explain why the C33A-Slug cells exhibited higher protein levels of β-catenin but lower activities of the Wnt/β-catenin signaling pathway in the TOP-Flash reporter assay. All of these results further confirmed that the activity of the Wnt/β-catenin signaling pathway was suppressed in all Slug-expressing cervical cancer cell lines. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26157 Moreover, the expression of signaling molecules of the Wnt/β-catenin signaling pathway [52–54], including GSK3β, p-GSK3β, β-catenin, cyclin D1 and c-myc, was detected in the Slug-modified cervical cancer cells and their controls. The protein levels of p-GSK3β, β-catenin, much higher in the HeLa-shSlug (Figure 5K) and CasKi- shSlug cells (Figure 5L) than in the HeLa-shControl and CasKi-shControl cells, respectively. These data suggested that the activity of the Wnt/β-catenin signaling pathway was attenuated in all Slug-expressing cervical cancer cells. O 26158 i tj l / t t Figure 4: Expression of slug in cervical cancer cells hindered cell cycle transition from G1 to S phase. The cell cy was analyzed using flow cytometry, and a quantitative analysis of the cell cycle is shown. The cell cycles of SiHa-GFP (A) and SiHa-S cells (B) and the quantitative analysis are shown (C). The cell cycles of C33A-GFP (D) and C33A-Slug cells (E) and the quantitat analysis are shown (F). The cell cycles of HeLa-shControl (G) and HeLa-shSlug cells (H) and the quantitative analysis are shown The cell cycles of Caski-shControl (J) and Caski- shSlug cells (K) and the quantitative analysis are shown (L). The data were shown he mean ± SD of three independent experiments. p < 0.05, *p < 0.01 vs. control using One-Way ANOVA. Figure 4: Expression of slug in cervical cancer cells hindered cell cycle transition from G1 to S phase. The cell cycle was analyzed using flow cytometry, and a quantitative analysis of the cell cycle is shown. The cell cycles of SiHa-GFP (A) and SiHa-Slug cells (B) and the quantitative analysis are shown (C). The cell cycles of C33A-GFP (D) and C33A-Slug cells (E) and the quantitative analysis are shown (F). The cell cycles of HeLa-shControl (G) and HeLa-shSlug cells (H) and the quantitative analysis are shown (I). The cell cycles of Caski-shControl (J) and Caski- shSlug cells (K) and the quantitative analysis are shown (L). The data were shown as the mean ± SD of three independent experiments. p < 0.05, *p < 0.01 vs. control using One-Way ANOVA. www.impactjournals.com/oncotarget Oncotarget 26158 cyclin D1 and c-myc were much higher in the HeLa- shSlug (Figure 5E and 5G, P < 0.05) and CasKi-shSlug cells (Figure 5F and 5H, P < 0.01) than in the HeLa- shControl and CasKi-shControl cells, respectively. Slug trans-suppressed Akt1/p-Akt1 expression by binding to the E-box motifs in the Akt1 promoter region p p The retino- blastoma protein (Rb) is a key factor that blocks cell cycle transition from G1 to S phase. Phosphorylation-Rb (p-Rb) is an inactivation form of Rb, p-Rb protein could be stimulated by cyclinD- CDK4/6 complexes, and results in release of the transcription factor E2F1 in order to activate S-phase entry [57]. Thus, western blotting analysis was used to detect the expression of p-Rb in the Slug-modified cervical cancer cell lines and their control cells. As shown in Figure 5, the protein levels of p-Rb were much lower in the SiHa-Slug (Figure 5A and 5C, P < 0.05) and C33A-Slug cells (Figure 5B and 5D, P < 0.05) than in the SiHa-GFP cells and C33A-GFP cells, respectively. Furthermore, the expression levels of p-Rb were much higher in the HeLa-shSlug (Figure 5E, and 5G, P < 0.05) and CasKi-shSlug cells (Figure 5F and 5H, P < 0.01) than in the HeLa-shControl and CasKi-shControl cells, respectively. All of these results indicated that Slug could inhibit cervical cancer cells at the transition from the G0/G1 phase to the S phase of the cell cycle. Dr. ER Fearon demonstrated that the E-box CACCTG motif was a binding site in the E-cadherin promoter region that allowed Slug to recognize and bind [25] to trans-suppress the expression of E-cadherin. A sequence analysis through the UCSC Genome online database revealed that seven E-boxes (CANNTG) are clustered in the Akt1 promoter region (from –494 bp to –1050 bp): one CACCTG (P4), two CAACTG (P1, P6), two CAGCTG (P2, P7), one CATCTG (P3), and one CAGTTG (P5). A luciferase reporter assay was used to determine whether Slug could trans-suppress the expression of Akt1 through these E-boxes (CANNTG). As shown in Figure 6I and 6J, E-boxes P2, P3, P4, P6 and P7 in the Akt1 promotor region in both SiHa-Slug and C33A- Slug cells had significantly stronger trans-suppression activities than their respective control cells (Figure 6I and 6J; p < 0.05), suggesting that Slug could trans-suppress Akt1/p-Akt1 expression through the E-boxes in the Akt1 promoter region in cervical cancer cells. www.impactjournals.com/oncotarget Oncotarget 26159 5: Slug down-regulated Akt1/p-Akt1 expression and suppressed Wnt/β-catenin pathway acti re 5: Slug down-regulated Akt1/p-Akt1 expression and suppressed Wnt/β-catenin pathway activity. (A) The ssion of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc, cyclinD1 and cyclinA in SiHa-GFP and SiHa-Slug was detected by western blotting, and the quantitative analysis is shown (C). DISCUSSION Slug is known to participate in the epithelial to mesenchymal transition (EMT) [22–24], and it is also been found to have different effects on cell proliferation and tumor formation in different carcinomas. But to our knowledge, no available reports identified the function of Slug in cervical carcinoma. As shown in Figure 1, both the immunoreactivity scores and western blotting analysis revealed that the expression of Slug was lower in cervical carcinoma tissues than in normal cervix tissues, suggesting that Slug is involved in the suppression of the development of cervical carcinoma. Additionally, we demonstrated that Slug could trans-suppresses Akt1/p-Akt1 protein expression and simultaneously up-regulates p21/p27 protein expression in all cervical cancer cells (Figure 5, p < 0.05). Akt1 has been found to be able to down-regulate the p21 protein at the transcriptional level through MDM2/p53 [63] or to down- regulate the p27 protein at the transcriptional level through FOXO [59]. Moreover, quantitative real-time–PCR results suggested that Slug could suppress cell proliferation and tumor formation through the up-regulation of p21/p27 at the transcriptional level through MDM2/p53 and FOXO (Figure S1B, S1C, S1D and S1E; P < 0.05). However, further experiments will be required to determine whether the up-regulation of p21/p27 by Slug is also caused by the decreased phosphorylation of p21/p27 through the trans- suppression of Akt1/p-Akt1 in cervical carcinoma cells. p When exogenous Slug protein was expressed in the SiHa and C33A cell lines (Figure 2A and 2B), in vitro cell proliferation and in vivo tumor formation were inhibited (Figure 2 and Figure 3). Moreover, in vitro cell proliferation and in vivo tumor formation were promoted (Figure 2 and Figure 3) when endogenous Slug protein expression was down-regulated in the HeLa and CasKi cell lines (Figure 2A and 2B). Therefore, Slug works as a cell proliferation inhibitor and tumor suppressor in cervical carcinoma, regardless of the levels of endogenous Slug protein in the cervical carcinoma cells. Furthermore, cell cycle analysis also showed that Slug arrested cell proliferation at the cell G1/G0 transition in all Slug-modified cervical carcinoma cells (Figure 4). Therefore, Slug appeared to inhibit cell proliferation and tumor formation by regulating G0/G1 transition-related cell cycle proteins. Slug trans-suppressed Akt1/p-Akt1 expression by binding to the E-box motifs in the Akt1 promoter region (B) The expression of Akt1, p-Akt1, p-RB, p21, p27, K3β, GSK3β, β-catenin, c-myc, cyclinD1 and cyclinA in C33A-GFP and C33A-Slug cells was detected by western blotting, and the titative analysis is shown (D). (E) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc, cyclinD1 yclinA in HeLa-shControl and HeLa- shSlug cells was detected by western blotting, and the quantitative analysis is shown (G). (F) xpression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc, cyclinD1 and cyclinA in Caski-shControl and Caski- ug cells was detected by western blotting, and the quantitative analysis is shown (H). SiHa (I), C33A (G), HeLa (K) and Caski (L) cells transfected with the TOP-Flash reporter plasmid, and the reporter activities were determined 48 h after transfection using a luciferase . The data were shown as the mean ± SD of three independent experiments. p < 0.05, *p < 0.01 vs. control using One-Way ANOVA. Figure 5: Slug down-regulated Akt1/p-Akt1 expression and suppressed Wnt/β-catenin pathway activity. (A) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc, cyclinD1 and cyclinA in SiHa-GFP and SiHa-Slug cells was detected by western blotting, and the quantitative analysis is shown (C). (B) The expression of Akt1, p-Akt1, p-RB, p21, p27, p-GSK3β, GSK3β, β-catenin, c-myc, cyclinD1 and cyclinA in C33A-GFP and C33A-Slug cells was detected by western blotting, and the quantitative analysis is shown (D). (E) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc, cyclinD1 and cyclinA in HeLa-shControl and HeLa- shSlug cells was detected by western blotting, and the quantitative analysis is shown (G). (F) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc, cyclinD1 and cyclinA in Caski-shControl and Caski- shSlug cells was detected by western blotting, and the quantitative analysis is shown (H). SiHa (I), C33A (G), HeLa (K) and Caski (L) cells were transfected with the TOP-Flash reporter plasmid, and the reporter activities were determined 48 h after transfection using a luciferase assay. The data were shown as the mean ± SD of three independent experiments. p < 0.05, *p < 0.01 vs. control using One-Way ANOVA. www.impactjournals.com/oncotarget Oncotarget 26160 cancer and melanoma cells [41, 43], but no data revealed whether Slug could regulate Akt1. Slug trans-suppressed Akt1/p-Akt1 expression by binding to the E-box motifs in the Akt1 promoter region In the present study, we demonstrated that Slug could trans-suppress Akt1/p-Akt1 expression (Figure 5, p < 0.05) in cervical carcinoma cells and then decrease the phosphorylation of GSK3β (Figure 5, p < 0.05) by Akt1. Moreover, active GSK3β increased the degradation of β-catenin and down-regulated the activity of the Wnt/β-catenin signaling pathway (Figure 5, p < 0.05) and the expression of related molecules, such as β-catenin, c-myc and cyclin D1 (Figure 5, p < 0.05). Moreover, mRNA levels of cyclin D1 were found to be much lower in the SiHa-Slug, C33A-Slug, HeLa- shControl and CasKi-shControl cells than in the SiHa- GFP, C33A-GFP, HeLa-shSlug and CasKi-shSlug cells, respectively (Figure S1B, S1C, S1D and S1E; P < 0.05), indicating that Slug could suppress cell proliferation and tumor formation through the Akt1/p-Akt1/GSK3β/ Wnt/β-catenin signaling pathway/cyclin D1 in cervical carcinoma. Of course, further experiments will be required to confirm whether the down-regulation of cyclin D1 by Slug in cervical carcinoma cells also occurs through direct degradation by GSK3β at the protein level [62] . Furthermore, a chromatin immunoprecipitation assay (ChIP) was used to determine whether Slug could recognize and bind directly to the E-boxes in the Akt1 promoter region. And the E-box CACCTG motif in the E-cadherin promoter region was used as a positive control (Figure S1F and S1G; p < 0.05). As shown in Figure 6K and 6L, Slug could specifically recognize and bind to the P2, P3, P4, P6 and P7 E-boxes in the Akt1 promoter region in both SiHa-Slug and C33A-Slug cells (Figure 6K and 6L; p < 0.05). All of these data indicated that Slug could recognize and bind to the E-boxes in the Akt1 promoter region and trans-suppress the expression of Akt1/p-Akt1. DISCUSSION (K) A quantitative CHIP assay of the Akt1 promoter region in SiHa-Slu (L) A quantitative CHIP assay of the Akt1 promoter region in C33A-Slug and C33A-GFP cells is shown. g-mediated disruption of Akt1/p-Akt1 and Wnt/β-catenin signaling in cervical cancer cells. When exogenou Oncotarget 26162 ww impactjournals com/oncotarget Figure 6: Slug trans-suppressed Akt1/p-Akt1. (A) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc and cyclinD1 in Akt1 transiently transfected SiHa-Slug cells was detected by western blotting, and the quantitative analysis was shown (B). The proliferation and viability of Akt1 transiently transfected SiHa-Slug cells were detected by growth curves (C) and MTT assay (D). (E) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc and cyclinD1 in Akt1 transiently transfected HeLa-shControl cells was detected by western blotting, and the quantitative analysis was shown (F). The proliferation and viability of Akt1 transiently transfected HeLa-shControl cells were detected by growth curves (G) and MTT assay (H). (I) The activity of the Akt1 promoter was measured using the dual luciferase assay and presented as the fold change in the rate of SiHa-Slug cells versus SiHa- GFP cells. (J) The activity of the Akt1 promoter was measured using the dual luciferase assay and presented as the fold change in the rate of C33A-Slug cells versus C33A-GFP cells. (K) A quantitative CHIP assay of the Akt1 promoter region in SiHa-Slug and SiHa-GFP cells is shown. (L) A quantitative CHIP assay of the Akt1 promoter region in C33A-Slug and C33A-GFP cells is shown. (M) Proposed model of the Slug-mediated disruption of Akt1/p-Akt1 and Wnt/β-catenin signaling in cervical cancer cells. When exogenous Slug was expressed in cervical cancer cells, Slug could recognize and bind to the E-boxes in the Akt1 promoter region as a transcription repressor and reduce Akt1/p-Akt1, leading to the up-regulation of p21 and p27. In addition, the reduction of Akt1/p-Akt1 also attenuated the Wnt/β-catenin signaling pathway and suppressed cell proliferation and tumor formation in cervical cancer cells through GSK3β. The data were shown as the mean ± SD of three independent experiments. p < 0.05, *p < 0.01 vs. control using One-Way ANOVA. Figure 6: Slug trans-suppressed Akt1/p-Akt1. (A) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc and cyclinD1 in Akt1 transiently transfected SiHa-Slug cells was detected by western blotting, and the quantitative analysis was shown (B). DISCUSSION Furthermore, Luciferase assays and CHIP assays confirmed that Slug could recognize (Figure 6I and 6J) and bind (Figure 6K and 6L) to the E-boxes in the Akt1 promoter region and act as a transcription repressor to reduce Akt1/p-Akt1 expression in both SiHa-Slug and C33A-Slug cells (Figure 6I, 6J, 6K and 6L). This is the first study to demonstrate that Slug could trans-suppress Akt1/p-Akt1 expression through the E-boxes in the Akt1 promoter region. There are several main G0/G1 transition-related cell cycle proteins: the cyclin D1, Rb, p21 and p27 proteins [59–61]. Dr. WS Wu demonstrated that Slug could inhibit cell proliferation in human prostate cancer cells via the down-regulation of cyclin D1 [38] but did not show how Slug down-regulated cyclin D1 expression in human prostate cancer cells. The cyclin D1 protein is a down- stream protein of the Wnt/β-catenin signaling pathway, and its expression is correlated positively with the activity of the Wnt/β-catenin signaling pathway [56]. Moreover, the cyclin D1 protein may also be degraded by GSK3β [62]. Akt1 was reported to be able to trans-activate Slug expression to promote EMT progression in breast In conclusion, this is the first study to demonstrate that Slug may trans-suppress Akt1/p-Akt1 expression by binding to the E-boxes in the Akt1 promoter region and inhibit the proliferation and tumor formation of human cervical cancer cells by up-regulating p21/p27 and/or down-regulating the activity of the Wnt/β-catenin signaling pathway (Figure 6M). Slug was found to act as a tumor suppressor in cervical carcinogenesis. www.impactjournals.com/oncotarget Oncotarget 26161 6: Slug trans-suppressed Akt1/p-Akt1. (A) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3 d cyclinD1 in Akt1 transiently transfected SiHa-Slug cells was detected by western blotting, and the qua ). The proliferation and viability of Akt1 transiently transfected SiHa-Slug cells were detected by growth . (E) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc and cyclinD d HeLa-shControl cells was detected by western blotting, and the quantitative analysis was shown (F). of Akt1 transiently transfected HeLa-shControl cells were detected by growth curves (G) and MTT assay ( promoter was measured using the dual luciferase assay and presented as the fold change in the rate of SiHa-S . (J) The activity of the Akt1 promoter was measured using the dual luciferase assay and presented as the f Slug cells versus C33A-GFP cells. Western blotting Western blotting analysis was carried out as previously described [65], the rabbit polyclonal antibody against human Slug (1:1000 dilution; #9585, Cell Signaling Technology), β-actin (1:1000 dilution; Santa Cruz, CA, USA), Akt1 (1:1000 dilution; Santa Cruz, CA, USA), p-Akt1 (1:1000 dilution; Cell Signaling Technology), p21 (1:500 dilution; Santa Cruz, CA, USA), p27 (1:500 dilution; Santa Cruz, CA, USA), cyclin A (1:1000 dilution; Santa Cruz, CA, USA), p-Rb (1:500 dilution; Santa Cruz, CA, USA), p-GSKβ (1:1000 dilution; Santa Cruz, CA, USA), β-catenin (1:1000 dilution; Santa Cruz, CA, USA), c-myc (1:1000 dilution, Santa Cruz, CA, USA), and cyclin D1 (1:500 dilution, Santa Cruz, CA, USA) were incubated with the membranes at 4°C overnight, followed by secondary incubation using a horseradish peroxidase-conjugated anti-rabbit or anti- mouse IgG (Thermo Fisher Scientific, New York, NY, USA). The proteins were visualized with an enhanced chemiluminescence reagent (Millipore, Billerica, MA, USA) through the protein imprinting imaging system (Tanon 5200, China). The Slug western blot results were normalized to those of β-actin blotting for quantification. 2 A total of 37 normal cervixes (NC), 28 cervical carcinomas in situ (CIS), and 52 squamous cervical cancer (SCC) tissues were obtained from the First Affiliated Hospital of Xi’an Jiaotong University between January 2008 and December 2014. 8 normal cervixes and 8 squamous cervical cancers (SCC) fresh tissues were collected from the First Affiliated Hospital of Xi’an Jiaotong University for Western blot analysis. Cell lines and human tissue specimens Human cervical carcinoma cell lines SiHa, C33A, HeLa, CasKi and HT-3 were purchased from the American Type Culture Collection (ATCC, Rockville, MD, USA). SiHa, C33A and HeLa were cultured in high-glucose Dulbecco Modified Eagle Medium (DMEM, Sigma-Aldrich, St Louis, MO, USA), CasKi was cultured in RPMI1640 (Sigma-Aldrich, St Louis, MO, USA), HT-3 was cultured in McCoy’s 5A Medium (Sigma-Aldrich, St Louis, MO, USA), and the cell culture mediums contained 10% fetal bovine serum (FBS, Invitrogen, Carlsbad, CA, USA). Cell lines were all incubated at 37°C in a humidified 5% CO2 atmosphere. Cell growth and cell viability assays Cells (5 × 104) were seeded in 2-mL media in 6-well plates in triplicate. Then the cells were counted every 2 days for 1 week using hemocytometer. Cell growth curves were generated to assess the cell proliferation. Cell viability was assessed using 3-(4, 5-dimethylthiazole- yl)-2, 5-diphenyl tetrazolium bromide (Sigma-Aldrich, St Louis, MO, USA) dye according to standard protocol. The number of viable cells was detected by measuring the absorbance at 490 nm. A standard immunostaining procedure was performed using a rabbit polyclonal antibody against human Slug (1:50 dilution; #9585, Cell Signaling Technology). A positive reaction was defined as the observation of a reddish-brown precipitate in the nucleus. Slug staining was classified into 2 categories (negative and positive expression) based on the percentage of positive cells and the staining intensity [64]. The percentage of positive cells was divided into 5 scores: < 5% (0), 5% to 25% (1), 25% to 50% (2), 50% to 75% (3), and > 75% (4). The intensity of staining was divided into 4 scores: no staining (0), light brown (1), brown (2), and dark brown (3). The positivity of Slug staining was determined using the following formula: immunohistochemistry (IHC) score = percentage score × intensity score. An overall score of ≤ 1 was defined as negative; a score of ≥ 2 was defined as positive. Immunohistochemistry and immunocytochemistry The immunohistochemical staining procedure was performed as previously described. Briefly, four- millimeter sections of formalin-fixed and paraffin- embedded tissue specimens were deparaffinized in xylene and rehydrated through descending concentrations of ethanol. After antigen retrieval was performed by heating in 10 mM citrate buffer (pH 6.0) for 2 min, the sections were treated with 3% hydrogen peroxide to block endogenous peroxidase. Subsequently, the sections were incubated with a primary antibody overnight at 4°C. A horseradish peroxidase-conjugated secondary antibody was added for 30 min at room temperature, followed by 3, 3ʹ-diaminobenzidine development. The sections were counterstained with hematoxylin. As a negative control, the primary antibody was replaced with PBS. DISCUSSION The proliferation and viability of Akt1 transiently transfected SiHa-Slug cells were detected by growth curves (C) and MTT assay (D). (E) The expression of Akt1, p-Akt1, p21, p27, p-RB, p-GSK3β, GSK3β, β-catenin, c-myc and cyclinD1 in Akt1 transiently transfected HeLa-shControl cells was detected by western blotting, and the quantitative analysis was shown (F). The proliferation and viability of Akt1 transiently transfected HeLa-shControl cells were detected by growth curves (G) and MTT assay (H). (I) The activity of the Akt1 promoter was measured using the dual luciferase assay and presented as the fold change in the rate of SiHa-Slug cells versus SiHa- GFP cells. (J) The activity of the Akt1 promoter was measured using the dual luciferase assay and presented as the fold change in the rate of C33A-Slug cells versus C33A-GFP cells. (K) A quantitative CHIP assay of the Akt1 promoter region in SiHa-Slug and SiHa-GFP cells is shown. (L) A quantitative CHIP assay of the Akt1 promoter region in C33A-Slug and C33A-GFP cells is shown. (M) Proposed model of the Slug-mediated disruption of Akt1/p-Akt1 and Wnt/β-catenin signaling in cervical cancer cells. When exogenous Slug was expressed in cervical cancer cells, Slug could recognize and bind to the E-boxes in the Akt1 promoter region as a transcription repressor and reduce Akt1/p-Akt1, leading to the up-regulation of p21 and p27. In addition, the reduction of Akt1/p-Akt1 also attenuated the Wnt/β-catenin signaling pathway and suppressed cell proliferation and tumor formation in cervical cancer cells through GSK3β. The data were shown as the mean ± SD of three independent experiments. p < 0.05, **p < 0.01 vs. control using One-Way ANOVA. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26162 MATERIALS AND METHODS For the expression of Slug in cells, similar immunocytochemistry was performed after the cells were seeded onto cover slips for 48 hours, fixed with 4% paraformaldehyde for 20 minutes, and permeabilized with 0.2% Triton X-100 for 20 minutes at room temperature. Luciferase reporter assay Before FACS analysis, 1 × 106 cells were harvested and washed twice with cold PBS and then fixed in 70% ice-cold ethanol overnight at 4°C for thirty minutes. After washing twice with PBS, the cells were treated with RNaseA and stained with propidium iodide (Sigma- Aldrich, St. Louis, MO, USA). Then, the cell was analyzed using a FACS Calibur flow cytometer (BD Biosciences, San Jose, CA, USA) with the CellQuest software. For promoter analyses, a fragment (from position –494 bp to –1050 bp relative to Akt1) containing the E-box site (CANNTG) was cloned into the pGL3-Basic Vector (Promega, Madison, WI, USA) to generate Akt1 promoter reporter constructs. Plasmids containing firefly luciferase reporters were co-transfected into tumor cells in triplicate using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA), with the thymidine kinase promoter Renilla luciferase reporter plasmid (pRL-TK) as an internal control. The activity of both the firefly and Renilla luciferase reporters was determined 48 hours after transfection using the Dual Luciferase Assay kit (Promega, Madison, WI). The specific promoter activity was presented as the relative ratio of firefly luciferase activity to Renilla luciferase activity. The specific promoter activity was presented as the change in the experimental group versus the control group. The primers and oligonucleotides are listed in Table S2. Restriction enzymes were obtained from TaKaRa. All constructs were verified by sequencing. The specific activity is shown as the fold change of the experimental group versus the control group. PCR analysis Total RNA from cultured cervical cancer cells was extracted using the TRIzol Reagent (Invitrogen, Carlsbad, CA, USA). Total cDNA was used as a template for PCR amplification, with β-actin as an internal control. Real-time quantitative PCR was performed in triplicate for each primer set and each cell sample, using an IQ5 multicolor real-time PCR Detection System (Bio-Rad, Hercules, CA). The protocol for real-time PCR was 1 cycle of 95 for 30 seconds, 40 cycles of 95°C for 5 seconds and 60 for 30 seconds, and then a dissociation stage. The cycle threshold value was determined as the point at which the fluorescence exceeded a preset limit determined by the instrument’s software. The results were analyzed via the ∆∆Ct method using GAPDH as the housekeeping genes, Primers used were as follows: GAPDH (GCACCGTCAAGGCTGAGAAC and TGGT GAAGACGCCAGTGGA); p27 (CTGCCCTCCCCAGT CTCTCT and CAAGCACCTCGGATTTT); p21 (GCAGA CCAGCATGACAGATTTC and CGGATTAGGGCTTC CTCTTG); and cyclin D1 (AAACAGATCATCCG CAAACAC and GTTGGGGCTCCTCAGGTTC). Quantitative chromatin immunoprecipitation SiHa and C33A cells were subjected to ChIP using the EZ-ChIP Assay kit (Millipore). Briefly, the cells were treated with 37% formaldehyde to crosslink proteins, and the reaction was terminated with 0.125 M glycine. After sonication, chromatin–protein complexes were immunoprecipitated with 5 μg of anti-Slug antibodies (#9585, Cell Signaling Technology) or 1 μg of mouse IgG. Real-time PCR was performed to amplify the regions of interest or internal negative control regions. Each sample was assayed in triplicate, and the fold enrichment ratio was calculated as the value of the ChIP sample versus the corresponding input sample. Samples that yielded a two- fold enrichment or better were considered positive targets. The primers used for these studies are listed in Table S3. Vector construction and transfection Full-length Slug cDNA was amplified using the following primers: forward, 5ʹ-GTTGAATTCGTTATG CCGCGCTCCTTCCTG-3ʹ and reverse, 5ʹ-CGCGGATCC TCAGTGTGCTACACAGCAG-3ʹ. The Slug DNA fragment was subsequently cloned into the EcoRI and BamHI sites of the internal ribosome entry site vector pIRES2-AcGFP (Clontech, Mountain View, CA) to generate the pIRES2-AcGFP-Slug recombinant plasmid. The small interfering RNA expression vector that expresses a Slug-specific short hairpin RNA (shRNA) was purchased from GenePharma (Shanghai, China). The Slug overexpression and shRNA vectors were transfected into SiHa, C33A, HeLa and CasKi cells using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, Tumor xenograft assay We obtained 6- to 7-week-old female BALB/c-nude mice from Slac Laboratory Animal Co., Ltd. (Shanghai, China). The mice were housed in an specific pathogen free (SFP) room with a constant temperature (22°C–25°C) and humidity (40–50%). Tumor cells in the exponential growth phase were harvested for inoculation, and the tumor cells (1 × 105) were then injected into the subcutis on the dorsum of each female BALB/c-nude mouse. The tumor volume (V) was determined using the following formula: length (a) and width (b) as V = ab2/2. At the end of the experiment, the www.impactjournals.com/oncotarget Oncotarget 26163 CA, USA) according to the manufacturer’s protocol. The transfected cells were treated with G418 (Calbiochem, La Jolla, CA, USA) for 3 weeks to collect, expand, and identify the drug-resistant colonies. The recombinant Akt1 plasmid was adapted from a previous study [66]. tumors were weighed and stored for immunostaining. The experimental protocols were evaluated and approved by the Animal Care and Use Committee of the Medical School of Xi’an Jiaotong University. REFERENCES 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA: a cancer journal for clinicians. 2011; 61:69–90. 16. Liu SY, Zheng PS. High aldehyde dehydrogenase activity identifies cancer stem cells in human cervical cancer. Oncotarget. 2013; 4:2462–2475. doi: 10.18632/oncotarget.1578. 2. Brebi MP, Ili GC, Lopez MJ, Garcia MP, Melo AA, Montenegro HS, Leal RP, Guzman GP, Roa SJ. [Detection and genotyping of human papillomavirus in biopsies of uterine cervical adenocarcinoma]. Rev Med Chile. 2009; 137:377–382. 17. Huang T-S, Li L, Moalim-Nour L, Jia D, Bai J, Yao Z, Bennett SAL, Figeys D, Wang L. A Regulatory Network Involving β-Catenin, E-Cadherin, PI3K/Akt, and Slug Balances Self-Renewal and Differentiation of Human Pluripotent Stem Cells in Response to Wnt Signaling. Stem Cells. 2014:n/a-n/a. 3. Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, Snijders PJ, Peto J, Meijer CJ, Munoz N. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol. 1999; 189:12–19. 18. Proia TA, Keller PJ, Gupta PB, Klebba I, Jones AD, Sedic M, Gilmore H, Tung N, Naber SP, Schnitt S, Lander ES, Kuperwasser C. Genetic predisposition directs breast cancer phenotype by dictating progenitor cell fate. Cell stem cell. 2011; 8:149–163. 4. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol. 2002;55:244 -265. 19. Guo W, Keckesova Z, Donaher JL, Shibue T, Tischler V, Reinhardt F, Itzkovitz S, Noske A, Zurrer-Hardi U, Bell G, Tam WL, Mani SA, van Oudenaarden A, Weinberg RA. Slug and Sox9 cooperatively determine the mammary stem cell state. Cell. 2012; 148:1015–1028. 5. Carter JR, Ding Z, Rose BR. HPV infection and cervical disease: a review. Aust Nz J Obstet. 2011; 51:103–108. 6. Mora N, Rosales R, Rosales C. R-Ras promotes metastasis of cervical cancer epithelial cells. Cancer Immuno Immun: CII. 2007; 56:535–544. 20. Nassour M, Idoux-Gillet Y, Selmi A, Come C, Faraldo ML, Deugnier MA, Savagner P. Slug controls stem/ progenitor cell growth dynamics during mammary gland morphogenesis. PloS one. 2012; 7:e53498. 7. Storey A, Thomas M, Kalita A, Harwood C, Gardiol D, Mantovani F, Breuer J, Leigh IM, Matlashewski G, Banks L. Role of a p53 polymorphism in the development of human papillomavirus-associated cancer. Nature. 1998; 393:229–234. 21. Ros MA, Sefton M, Nieto MA. ACKNOWLEDGMENTS AND FUNDING 12. Chen Q, Cao HZ, Zheng PS. LGR5 promotes the proliferation and tumor formation of cervical cancer cells through the Wnt/beta-catenin signaling pathway. Oncotarget. 2014; 5:9092–9105. doi: 10.18632/oncotarget.2377. This work was supported by a grant to Prof. Peng-Sheng Zheng from the National Natural Science Foundation of China (No.81472728), and Young grant (No. 81302278) from the National Natural Science Foundation to Wen-ting Yang. 13. Yang WT, Zheng PS. Kruppel-like factor 4 functions as a tumor suppressor in cervical carcinoma. Cancer. 2012; 118:3691–3702. 14. Wu XL, Zheng PS. Undifferentiated embryonic cell transcription factor-1 (UTF1) inhibits the growth of cervical cancer cells by transactivating p27Kip1. Carcinogenesis. 2013; 34:1660–1668. CONFLICTS OF INTEREST None. 15. Liu XF, Yang WT, Xu R, Liu JT, Zheng PS. Cervical cancer cells with positive Sox2 expression exhibit the properties of cancer stem cells. PloS one. 2014; 9:e87092. Statistical analysis Statistical analysis was performed with SPSS 16.0 software (SPSS Inc., Chicago, IL, USA). The significance of the differences between the covariates was determined by the 2-tailed chi-square test. The Student t test was used to determine the statistical significance for 2-group analyses. In all tests, statistically significant was defined as p < 0.05. www.impactjournals.com/oncotarget Oncotarget 26164 REFERENCES Slug, a zinc finger gene previously implicated in the early patterning of the mesoderm and the neural crest, is also involved in chick limb development. Development (Cambridge, England). 1997; 124:1821–1829. 8. Mitra S, Misra C, Singh RK, Panda CK, Roychoudhury S. Association of specific genotype and haplotype of p53 gene with cervical cancer in India. J Clin Pathol. 2005; 58:26–31. 22. Savagner P, Yamada KM, Thiery JP. The zinc-finger protein slug causes desmosome dissociation, an initial and necessary step for growth factor-induced epithelial- mesenchymal transition. J Pathol. 1997; 137:1403–1419. 9. Liu X, Yang WT, Zheng PS. Msi1 promotes tumor growth and cell proliferation by targeting cell cycle checkpoint proteins p21, p27 and p53 in cervical carcinomas. Oncotarget. 2014; 5:10870–10885. doi: 10.18632/oncotarget.2539. 23. Hemavathy K, Guru SC, Harris J, Chen JD, Ip YT. Human Slug Is a Repressor That Localizes to Sites of Active Transcription. Mol Cell Biol. 2000; 20:5087–5095. 10. Gu TT, Liu SY, Zheng PS. Cytoplasmic NANOG-positive stromal cells promote human cervical cancer progression. Am J Pathol. 2012; 181:652–661. 24. Nieto MA, Sargent MG, Wilkinson DG, Cooke J. Control of cell behavior during vertebrate development by Slug, a zinc finger gene. Science (New York, NY). 1994; 264:835–839. 11. Wang YD, Cai N, Wu XL, Cao HZ, Xie LL, Zheng PS. OCT4 promotes tumorigenesis and inhibits apoptosis of cervical cancer cells by miR-125b/BAK1 pathway. Cell Death Dis. 2013; 4:e760. www.impactjournals.com/oncotarget Oncotarget 26165 25. Hajra KM, Chen DY, Fearon ER. The SLUG zinc-finger protein represses E-cadherin in breast cancer. Cancer Res. 2002; 62:1613–1618. 37. Turner FE, Broad S, Khanim FL, Jeanes A, Talma S, Hughes S, Tselepis C, Hotchin NA. Slug regulates integrin expression and cell proliferation in human epidermal keratinocytes. J Biol Chem. 2006; 281:21321–21331. 26. Saito T, Nagai M, Ladanyi M. SYT-SSX1 and SYT-SSX2 interfere with repression of E-cadherin by snail and slug: a potential mechanism for aberrant mesenchymal to epithelial transition in human synovial sarcoma. Cancer Res. 2006; 66:6919–6927. 38. Liu J, Uygur B, Zhang Z, Shao L, Romero D, Vary C, Ding Q, Wu WS. Slug inhibits proliferation of human prostate cancer cells via downregulation of cyclin D1 expression. Prostate. 2010; 70:1768–1777. 27. Uchikado Y, Natsugoe S, Okumura H, Setoyama T, Matsumoto M, Ishigami S, Aikou T. Slug Expression in the E-cadherin preserved tumors is related to prognosis in patients with esophageal squamous cell carcinoma. Clin Cancer Res. 2005; 11:1174–1180. 39. REFERENCES Schluter C, Duchrow M, Wohlenberg C, Becker MH, Key G, Flad HD, Gerdes J. The cell proliferation-associated antigen of antibody Ki-67: a very large, ubiquitous nuclear protein with numerous repeated elements, representing a new kind of cell cycle-maintaining proteins. J Cell Biol. 1993; 123:513–522. 28. Xie Y, Liu S, Lu W, Yang Q, Williams KD, Binhazim AA, Carver BS, Matusik RJ, Chen Z. Slug regulates E-cadherin repression via p19Arf in prostate tumorigenesis. Mol Oncol. 2014; 8:1355–1364. 40. Saegusa M, Hashimura M, Kuwata T, Okayasu I. Requirement of the Akt/beta-catenin pathway for uterine carcinosarcoma genesis, modulating E-cadherin expression through the transactivation of slug. Am J Pathol. 2009; 174:2107–2115. 29. Alves CC, Carneiro F, Hoefler H, Becker KF. Role of the epithelial -mesenchymal transition regulator Slug in primary human cancers. Front Biosci. 2009; 14:3035–3050. 41. Carpenter RL, Paw I, Dewhirst MW, Lo HW. Akt phosphorylates and activates HSF-1 independent of heat shock, leading to Slug overexpression and epithelial- mesenchymal transition (EMT) of HER2-overexpressing breast cancer cells. Oncogene. 2015; 34:546–557. 30. Chen G, Hou Z, Gulbranson DR, Thomson JA. Actin- myosin contractility is responsible for the reduced viability of dissociated human embryonic stem cells. Cell stem cell. 2010; 7:240–248. 42. Wu DW, Lee MC, Hsu NY, Wu TC, Wu JY, Wang YC, Cheng YW, Chen CY, Lee H. FHIT loss confers cisplatin resistance in lung cancer via the AKT/NF-kappaB/Slug- mediated PUMA reduction. Oncogene. 2015; 34:2505–2515. 31. Li L, Wang S, Jezierski A, Moalim-Nour L, Mohib K, Parks RJ, Retta SF, Wang L. A unique interplay between Rap1 and E-cadherin in the endocytic pathway regulates self-renewal of human embryonic stem cells. Stem Cells. 2010; 28:247–257. 43. Fenouille N, Tichet M, Dufies M, Pottier A, Mogha A, Soo JK, Rocchi S, Mallavialle A, Galibert MD, Khammari A, Lacour JP, Ballotti R, Deckert M, et al. The epithelial-mesenchymal transition (EMT) regulatory factor SLUG (SNAI2) is a downstream target of SPARC and AKT in promoting melanoma cell invasion. PloS one. 2012; 7:e40378. 32. Inoue A, Seidel MG, Wu W, Kamizono S, Ferrando AA, Bronson RT, Iwasaki H, Akashi K, Morimoto A, Hitzler JK, Pestina TI, Jackson CW, Tanaka R, et al. Slug, a highly conserved zinc finger transcriptional repressor, protects hematopoietic progenitor cells from radiation-induced apoptosis in vivo. Cancer cell. 2002; 2:279–288. 44. Menges CW, Sementino E, Talarchek J, Xu J, Chernoff J, Peterson JR, Testa JR. REFERENCES Group I p21-activated kinases (PAKs) promote tumor cell proliferation and survival through the AKT1 and Raf-MAPK pathways. Mol Cancer Res: MCR. 2012; 10:1178–1188. 33. Tamminen JA, Parviainen V, Ronty M, Wohl AP, Murray L, Joenvaara S, Varjosalo M, Lepparanta O, Ritvos O, Sengle G, Renkonen R, Myllarniemi M, Koli K. Gremlin-1 associates with fibrillin microfibrils in vivo and regulates mesothelioma cell survival through transcription factor slug. Oncogenesis. 2013; 2:e66. 45. Skeen JE, Bhaskar PT, Chen CC, Chen WS, Peng XD, Nogueira V, Hahn-Windgassen A, Kiyokawa H, Hay N. Akt deficiency impairs normal cell proliferation and suppresses oncogenesis in a p53-independent and mTORC1-dependent manner. Cancer cell. 2006; 10:269–280. 34. Haslehurst AM, Koti M, Dharsee M, Nuin P, Evans K, Geraci J, Childs T, Chen J, Li J, Weberpals J, Davey S, Squire J, Park PC, Feilotter H. EMT transcription factors snail and slug directly contribute to cisplatin resistance in ovarian cancer. BMC cancer. 2012; 12:91. 46. Porstmann T, Santos CR, Griffiths B, Cully M, Wu M, Leevers S, Griffiths JR, Chung YL, Schulze A. SREBP Activity Is Regulated by mTORC1 and Contributes to Akt- Dependent Cell Growth. Cell Metab. 2008; 8:224–236. 35. Yang HW, Menon LG, Black PM, Carroll RS, Johnson MD. SNAI2/Slug promotes growth and invasion in human gliomas. BMC cancer. 2010; 10:301. 47. Manning BD, Cantley LC. AKT/PKB Signaling: Navigating Downstream. Cell. 2007; 129:1261–1274. 36. Wang Y-P, Wang M-Z, Luo Y-R, Shen Y, Wei Z-X. Lentivirus-mediated shRNA Interference Targeting SLUG Inhibits Lung Cancer Growth and Metastasis. Asian Pac J Cancer P. 2012; 13:4947–4951. 48. Nakao T, Geddis AE, Fox NE, Kaushansky K. PI3K/Akt/ FOXO3a pathway contributes to thrombopoietin-induced proliferation of primary megakaryocytes in vitro and in vivo www.impactjournals.com/oncotarget Oncotarget 26166 via modulation of p27(Kip1). Cell cycle (Georgetown, Tex). 2008; 7:257–266. 58. Yang J, Wu Z, Renier N, Simon DJ, Uryu K, Park DS, Greer PA, Tournier C, Davis RJ, Tessier-Lavigne M. Pathological axonal death through a MAPK cascade that triggers a local energy deficit. Cell. 2015; 160:161–176. 49. Rowland BD, Peeper DS. KLF4, p21 and context-dependent opposing forces in cancer. Nat Rev Cancer. 2006; 6:11–23. 59. Medema RH, Kops GJ, Bos JL, Burgering BM. AFX- like Forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature. 2000; 404:782–787. 50. Beaulieu JM, Gainetdinov RR, Caron MG. The Akt-GSK-3 signaling cascade in the actions of dopamine. Trends Pharmacol Sci. 2007; 28:166–172. 51. REFERENCES Fukumoto S, Hsieh CM, Maemura K, Layne MD, Yet SF, Lee KH, Matsui T, Rosenzweig A, Taylor WG, Rubin JS, Perrella MA, Lee ME. Akt participation in the Wnt signaling pathway through Dishevelled. J Biol Chem. 2001; 276:17479 -17483. 60. Zhou BP, Liao Y, Xia W, Spohn B, Lee MH, Hung MC. Cytoplasmic localization of p21Cip1/WAF1 by Akt-induced phosphorylation in HER-2/neu-overexpressing cells. Nat Cell Biol. 2001; 3:245–252. 61. Shan J, Zhao W, Gu W. Suppression of cancer cell growth by promoting cyclin D1 degradation. Mol Cell. 2009; 36:469–476. 52. Sakai D, Tanaka Y, Endo Y, Osumi N, Okamoto H, Wakamatsu Y. Regulation of Slug transcription in embryonic ectoderm by beta-catenin-Lef/Tcf and BMP- Smad signaling. Dev Growth Differ. 2005; 47:471–482. 62. Diehl JA, Cheng M, Roussel MF, Sherr CJ. Glycogen synthase kinase-3beta regulates cyclin D1 proteolysis and subcellular localization. Genes Dev. 1998; 12:3499–3511. 53. Wu ZQ, Li XY, Hu CY, Ford M, Kleer CG, Weiss SJ. Canonical Wnt signaling regulates Slug activity and links epithelial-mesenchymal transition with epigenetic Breast Cancer 1, Early Onset (BRCA1) repression. P Natl Acad Sci USA. 2012; 109:16654–16659. 63. Mayo LD, Donner DB. A phosphatidylinositol 3-kinase/ Akt pathway promotes translocation of Mdm2 from the cytoplasm to the nucleus. P Natl Acad Sci USA. 2001; 98:11598–11603. 54. Zhang J, Gill AJ, Issacs JD, Atmore B, Johns A, Delbridge LW, Lai R, McMullen TP. The Wnt/beta-catenin pathway drives increased cyclin D1 levels in lymph node metastasis in papillary thyroid cancer. Hum Pathol. 2012; 43:1044 -1050. 64. Wang L, Wei D, Huang S, Peng Z, Le X, Wu TT, Yao J, Ajani J, Xie K. Transcription factor Sp1 expression is a significant predictor of survival in human gastric cancer. Clinical cancer res. 2003; 9:6371–6380. 55. Shan J, Zhao W, Gu W. Suppression of Cancer Cell Growth by Promoting Cyclin D1 Degradation. Mol Cell. 2009; 36:469–476. 65. Chen ZY, Wang X, Zhou Y, Offner G, Tseng CC. Destabilization of Kruppel-like factor 4 protein in response to serum stimulation involves the ubiquitin-proteasome pathway. Cancer Res. 2005; 65:10394–10400. 56. Lin SY, Xia W, Wang JC, Kwong KY, Spohn B, Wen Y, Pestell RG, Hung MC. Beta-catenin, a novel prognostic marker for breast cancer: its roles in cyclin D1 expression and cancer progression. P Natl Acad Sci USA. 2000; 97:4262–4266. 66. Cai N, Wang YD, Zheng PS. The microRNA-302-367 cluster suppresses the proliferation of cervical carcinoma cells through the novel target AKT1. Rna. REFERENCES 2013; 19:85–95. 57. Rizzolio F, Lucchetti C, Caligiuri I, Marchesi I, Caputo M, Klein-Szanto AJ, Bagella L, Castronovo M, Giordano A. Retinoblastoma tumor-suppressor protein phosphorylation and inactivation depend on direct interaction with Pin1. Cell Death Differ. 2012; 19:1152–1161. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26167
https://openalex.org/W2142551486	https://www.scielo.br/j/pci/a/3vmRWcTZnWsMphrt6K8WPLF/?lang=pt&format=pdf	Portuguese	null	A informação como um elemento chave para a qualidade do produto turístico: uma análise dos postos de informações turísticas do município de Florianópolis/SC	Perspectivas em Ciência da Informação	2,009	cc-by	10,997	!"# " $ %& ' ( ) +,-+.+,/01,2/ + &+.330 !" !" # $! ! % &! % !$!' (! % % !$!' (! % ) * ) * ( ! " % ! % ( ! " % ! % ! " #$ % &'( ) * # + , - ! . +!, / / 0/ +!, / / 0 +!, / !"# " $ %& -./ , / / 01/ ( 4 .5+35+.330.-+36+.330 ' ( ) +,-+.+,/01,2/ + &+.330 % &' ' ' ( ) ( , , , 112,, , ( " ($% &3- , 1 1 , ) 1 , (+ 2 , , , 1 -./ , / / 01/ ( ' ( ) +,-+.+,/01,2/ + &+.330 % 7 8 & +98: 8 * + & ; * < =>?4" ?47<>" #!>* .336" 7)@)$A4" B4CDE" E)7!* .33/" #>@<9" E)$" $944)97* .33/F+ 7 * ; & =7)@)$A4" B4CDE" E)7!* .33/F+ 7 * ; ; &* : + & : && G * & G +< &* G &* G * G + #$ ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& 7 =.33-F &; & ( & = & F* & = ; F+A; * & H & * 8* ; &* ') < =')<IF+ 9: ') < =A!A<C7* ,002F * * ( =J>#A)#@9* .33KF+ & & + * ')<I 8 + 8 * 8L + ') < & +7* & G& & M ')<I & N ' ')<I * ; + * 1 H * * : * 8 +' * L * ')<I * & & ; + L O ')<I & + '* & O ')<I + * ; + # !"# " $ %& ' ( ) +,-+.+,/01,2/ + &+.330 ' ( ) +,-+.+,/01,2/ + &+.330 * + ' O 8 1 < +9 : =99'>4 + .33,F+& $=.33.+,5KF P Q & * + 9 8 L( + J B=,00KF* 8 L( & = F +8* 8 * R & * * & =$)JJ#><97.33."99'>4+.33,"4A@$77,000" @A ,00/" D>)<@$#" '4A4$7" ?>44C* ,02KF+ J * # & =.336F B =,00KF* L * &: H * )!+,* & + * & Componentes do Produto Turístico Trânsito Agentes públicos, privados e voluntários - Diferentes modais de transporte: rodoviário, ferroviário, marítimo, aéreo - Outros Antes da viagem Agentes públicos, privados e voluntários - Centros de informação - Agentes de viagem - Companhias de transporte - Companhias de seguro - Outros Destino Agentes públicos, privados e voluntários - Acomodações - Atrações - Infra-estrutura - Vias urbanas, rodovias - Postos de Informações Turísticas - Outros )!A4,S A!A<C7=,002F+ Componentes do Produto Turístico Destino )!A4,S A!A<C7=,002F+ #* ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& !"# " $ %& 7 )!+ , 8 8 + @ =,00/F H * + & = L G FO =>?4" ?47<>" #!> .336" 7)@)$A4"B4CDE"E)7!.33/"A!A<C7,002"@A,00/F+ 7 & & : +& . & + & )!A4.S' >4>7D=.33.+.36F+ )!A4.S' >4>7D=.33.+.36F+ < & =& & * * * &( & F & O +7 8 + % * MAJ49,+ ' ( ) +,-+.+,/01,2/ + &+.330 #, !"# " $ %& MAJ49,4 Autor(es) Relação do turismo com a informação Poon (1988) Na rotina da atividade turística existe a geração, a coleta, o processamento, a aplicação e a comunicação da informação. * + A informação é o laço que amarra todos os componentes da indústria turística. Naisbitt (1994) Com o crescimento do turismo e com a sofisticação dos viajantes, a demanda por informações levará a uma interconectividade ainda maior. Buhalis (1998) Informação é vital para a indústria de viagens. Cooper et al. (2001) A atividade turística depende da boa disseminação de informações e da distribuição eficaz do produto turístico. O’Connor (2001) O turismo depende cada vez mais da informação. A informação é o nutriente básico do turismo. Middleton (2002) O turismo é um mercado totalmente baseado no fornecimento de informações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ritério Significado Tangíveis Referem-se a quaisquer evidências físicas do serviço, tais como aparência dos funcionários, instalações de apoio e equipamentos utilizados no processo. Engloba as dimensões de estética, limpeza e conforto. Consistência Conformidade com experiência anterior, ausência de variabilidade no resultado. Competênci a Habilidade e conhecimento profissional que garantem a execução eficaz do serviço. ' ( ) +,-+.+,/01,2/ + &+.330 -0 Velocidade de Atendimento Rapidez e prontidão para atender o cliente. Atendimento /Atmosfera Refere-se ao quão agradável é a experiência que o cliente tem durante o processo de prestação do serviço. Inclui a cortesia dos funcionários e a atenção personalizada ao cliente. Flexibilidade Capacidade de mudar e adaptar o serviço para se ajustar às necessidades dos clientes. Credibilidade /Segurança Baixa percepção de risco por parte do cliente e habilidade de transmitir confiança. Acesso Facilidade de entrar em contato ou acessar fisicamente o serviço. -0 !)7>)944W=,00-F+ J49.98 Y ') < 8 .* : ')<I + 88: = L 49" !4Z.332"##7"EC7J<.33,"9[7>)##+.333F+ F 8 + 9 & 11 * & * 8( ( # ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& !"# " $ %& * * L & + ')<I * ')<I 1 * ( & + 8 * L ')<I+ < G ')<I * * = F* + < & ; * & + ' ( ) +,-+.+,/01,2/ + &+.330 /1 2 ') < ; * & \ ')<I & + 7 ')<I =FG = F" =F G 1 * 8 " =F ; " =F G + 7 ; )!+58 ')<I+ > O ; + ( 8&8 * ! ( =,00-F8 MAJ495+ 7 ; ')<I + &* ; 1 * * L + 9 && * ;& & ')<I & H + J ; : ')<I +' ; * ')<I * T U =C)7* ,00-F+ 7 * ## ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& ')<I , * + ')<I , * + Componentes do Produto Turístico Trânsito Agentes públicos, privados e voluntários - Diferentes modais de transporte: rodoviário, ferroviário, marítimo, aéreo - Outros Antes da viagem Agentes públicos, privados e voluntários - Centros de informação - Agentes de viagem - Companhias de transporte - Companhias de seguro - Outros Postos de Informações Turísticas Destino Agentes públicos, privados e voluntários Acomodações Atrações Infra-estrutura Vias urbanas, rodovias Atendentes Informação verbal Material impresso Aspectos físicos e acesso Aspectos analisados na pesquisa )!A45S ') < J + Componentes do Produto Turístico - Diferentes modais de transporte: rodoviário, ferroviário, marítimo, aéreo - Outros Informação verbal )!A45S ') < J + )!A45S ') < J + MAJ4958 ')<I+ MAJ4958 ')<I+ Aspectos analisados Critérios de avaliação da qualidade Aspectos físicos e de acesso - Tangíveis: equipamentos, infra-estrutura física e tecnológica. /1 2 - Acesso: localização, sinalização, estacionamento, linhas telefônicas e terminais de internet. Atendentes - Tangíveis: apresentação e aparência dos funcionários. - Competência: formação, habilidade e conhecimento dos atendentes para executar o serviço. - Velocidade de atendimento: tempo gasto no processo de atendimento. - Flexibilidade: mudança de atitude de acordo com a necessidade do cliente. - Atendimento/Atmosfera: clareza na comunicação, receptividade, simpatia e cordialidade dos funcionários com os turistas. Informação verbal - Consistência: uniformidade de procedimentos realizados durante o atendimento. - Credibilidade/Segurança: confiabilidade das informações repassadas pelos atendentes. Material impresso - Tangíveis: aspectos físicos e disponibilidade dos materiais de divulgação. /1 2 J + 8 * 1 + 9 ')<I : * * + #3 ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& T 1 1 U* =1 F* : +J =')<IF ; 8 * ,45 * & + 8 * ; 1 J J < $ < =><A4F * ')<I* H * * 1 + . !" 9 ' ) < =')<IF G $ < =><A4F+> O : * < =>4><4) $A7))'# J> <A4)$9 1 ><A4* .332F : "& & ; L ; " + & ><A4 =.332F ')<I * H + ; ')<I 8; MAJ495+9 & + ' ( ) +,-+.+,/01,2/ + &+.330 4 ; ')<I * ; & L +7 * ; & G ; G +8 * 8 * + * 8 * ')<I ; * & + ')<I L + 9 1 ')<I * & ; + J & * 1 * # !"# " $ %& * G + * G + ; ')<I ; & = F ( * G ,45 + ')<I 8 ; 1 & * + A G & & ')<I* * * L G + ; * * L* & * L + 9 G : & ')<I* ; 5 * + ' ( ) +,-+.+,/01,2/ + &+.330 4,% 5 > * L * & G L + 7 * * * * & G + ' * + '8* ( 1 & + J ')<I* 1 +> * ')<I* 8 ; = F+8* H * * + 4* * & ; = F+><A4 ; * * ; + ' * + 7 & G G ( * 1 H &* 8 & < + ! L( = 8 & (F+ > * & < * L = & L * & * * 5 * * ( * F+ & T U+ 7 ; * & * & * & ; + X * + 3$ ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& J G ( * 1* & * * + * T U & ; + > * ; L : * ; L & * &? * * &+ J ><A4* ( R( ( + ><A4 ; 4 ')<I+ 7 * 1 * & ')<I* L = F L ; = ')<IF+ ' ( ) +,-+.+,/01,2/ + &+.330 4-6 & ; + 7 ? 4& + 8 * = F 3 ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& ; ')<I & (&( + 9 * & + & * 8 * * & + > * ; * & H ; * 8 H+ 9 &* * ><A4+ 8* * ')<I* +* 1 O & L &')<I* + > ( * & ')<I+ '8* * * : ')<I ; +>L & ; ')<I* ; & +J * : & T ; U : TL U+ ' ( ) +,-+.+,/01,2/ + &+.330 Sugestões de melhoria - Melhorar a sinalização viária e nas portas de entrada. - Adquirir mobiliário para armazenamento de material informativo. ) p g g p - Disponibilizar somente folders de assuntos relacionados com as necessidades dos turistas. - Organizar os materiais informacionais por categorias, de acordo com as necessidades dos usuários. J + J + & ')<I L H & + 09:;.-!2 - Realizar reuniões periódicas entre todos os atendentes com os seguintes objetivos: integração da equipe, avaliação do trabalho, sugestões de melhoria e encaminhamento de tarefas prioritárias para a diretoria. Informação verbal - Capacitar funcionários para aumentar a credibilidade e a segurança das informações repassadas aos turistas. Material impresso - Disponibilizar material da região sul do Brasil e mapas da cidade (praias e centro) em português, inglês e espanhol. - Disponibilizar somente folders de assuntos relacionados com as necessidades dos turistas. - Organizar os materiais informacionais por categorias, de acordo com as necessidades dos usuários. J + ' ( ) +,-+.+,/01,2/ + &+.330 09:;.-!2 09:;.-!2 09:;.-!2 Aspectos Sugestões de melhoria Aspectos físicos e de acesso - Melhorar a sinalização viária e nas portas de entrada. - Adquirir mobiliário para armazenamento de material informativo. - Melhorar a limpeza e a organização dos banheiros. - Melhorar o acesso para os deficientes físicos. - Adquirir equipamentos de informática, linhas e aparelhos telefônicos. - Padronizar o horário de atendimento em todos os PIT’s. Atendentes - Padronizar procedimentos internos (atendimento telefônico, atendimento “face- a-face”, periodicidade e forma de envio do registro de visitantes, apresentação pessoal e postura profissional, etc.) e treinar os atendentes nos novos procedimentos. - Estabelecer critérios de seleção para os estagiários, tais como: ser aluno de curso de Turismo, falar outros idiomas (inglês ou espanhol) ou ter experiência na área. - Selecionar um funcionário da Diretoria de Turismo da SETUR para ser o responsável pela gestão dos PIT’s. - Realizar reuniões periódicas entre todos os atendentes com os seguintes objetivos: integração da equipe, avaliação do trabalho, sugestões de melhoria e encaminhamento de tarefas prioritárias para a diretoria. Informação verbal - Capacitar funcionários para aumentar a credibilidade e a segurança das informações repassadas aos turistas. Material impresso - Disponibilizar material da região sul do Brasil e mapas da cidade (praias e centro) em português, inglês e espanhol. - Disponibilizar somente folders de assuntos relacionados com as necessidades dos turistas. - Organizar os materiais informacionais por categorias, de acordo com as necessidades dos usuários. J + 09:;.-!2 Aspectos Sugestões de melhoria Aspectos físicos e de acesso - Melhorar a sinalização viária e nas portas de entrada. - Adquirir mobiliário para armazenamento de material informativo. - Melhorar a limpeza e a organização dos banheiros. - Melhorar o acesso para os deficientes físicos. - Adquirir equipamentos de informática, linhas e aparelhos telefônicos. - Padronizar o horário de atendimento em todos os PIT’s. Atendentes - Padronizar procedimentos internos (atendimento telefônico, atendimento “face- a-face”, periodicidade e forma de envio do registro de visitantes, apresentação pessoal e postura profissional, etc.) e treinar os atendentes nos novos procedimentos. - Estabelecer critérios de seleção para os estagiários, tais como: ser aluno de curso de Turismo, falar outros idiomas (inglês ou espanhol) ou ter experiência na área. - Selecionar um funcionário da Diretoria de Turismo da SETUR para ser o responsável pela gestão dos PIT’s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`97794* '+ = > ,+'&?V .33,+ 9[7>)## $ + + J& ] +8 '01+,3+5+,5,1,-3.333+ 3/ ' ( ) +,-+.+,/01,2/ + &+.330 !"# " $ %& '4A4$7+"D>)<@$#%++"?>44C#+#+ ] + 6 84 +-0+-+-,1K3* ,02K+ aaaaaa+ >4%MA# 1 & ]+6 9 +/-+,+,.1-3 &,022+ '997* + < &+ ) 9,+,K+K+K5,1K-0,022+ 4A@$77* J+ 84 + -+ + ' ,000+ >?4+"?47<>X+#+"#!>#++< & &L + 8 +.2+ /+,K-,1,KK-.336+ >4><4) $A7))'# J> <A4)$9* A#<A4 > >'94<> =><A4F+ J b^^^+ ++&+ a + 3, .332 >4><4) $A7))'# J> <A4)$9* A#<A4 > >'94<> =><A4F+ J * = F J b^^^+ ++&+ a + c+3, &+.332+ b^^^+ ++&+ a + c+3, &+.332+ )#%X+++' : G + ?+ .+K+01..* +,000 +.333+ #E7+) +' ,006+ <>)_>)4* >+ #+ @ + 4 X M ] V,000+ B@>>#>4 $+ < V& + 84 92+,.+-+521--,00K+ C)74+(1,& +.++< 9 V &' ,00-+ D>)<@$#* %+ +" '4A4$7* +" ?>44C* #+ #+ ' & V&+6 84 +-0+.+551 -/&,02K+ 3
https://openalex.org/W4288049183	https://escholarship.org/content/qt19p3t98g/qt19p3t98g.pdf?t=rx01gk	English	null	Establishment of four head and neck squamous cell carcinoma cell lines: importance of reference DNA for accurate genomic characterisation	Journal of laryngology and otology/The journal of laryngology & otology	2,022	cc-by	6,797	UCLA UCLA Previously Published Works UCLA UCLA Previously Published Works Title Establishment of four head and neck squamous cell carcinoma cell lines: importance of reference DNA for accurate genomic characterisation Permalink https://escholarship.org/uc/item/19p3t98g Journal The Journal of Laryngology and Otology, 137(3) Title Establishment of four head and neck squamous cell carcinoma cell lines: importance of reference DNA for accurate genomic characterisation © The Author(s), 2022. Published by Cambridge University Press on behalf of J.L.O. (1984) LIMITED. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited. Powered by the California Digital Library University of California Powered by the California Digital Library University of California eScholarship.org eScholarship.org Author for correspondence: Author for correspondence: Dr Anthony Nichols, Department of Otolaryngology – Head and Neck Surgery, London Health Science Centre, Victoria Hospital, Room B3-431A, 800 Commissioners Road East, London, ON, Canada N6A 5W9 E-mail: Anthony.Nichols@lhsc.on.ca Fax: +1 519 685 8567 Results. Four early passage cell lines were established. Two cell lines were human papilloma- virus positive, confirmed by sequencing and p16 immunoblotting. Short tandem repeat pro- filing confirmed that all cell lines were established from their index tumours. Whole exome sequencing revealed that the matched normal reference DNA was critical for accurate muta- tional analysis: a high rate of false positive mutation calls were excluded (87.6 per cent). Conclusion. Early passage cell lines were successfully established. Patient-matched reference DNA is important for accurate cell line mutational calls. Main Article 1Department of Head and Neck – Endocrine Oncology, Moffit Cancer Center, Tampa, Florida, USA, 2Informatics and Biocomputing Platform Ontario Institute for Cancer Research, Toronto, Canada, 3Department of Human Genetics, University of California, Los Angeles, USA, 4Institute for Precision Health, University of California, Los Angeles, USA, 5Department of Otolaryngology – Head and Neck Surgery, Western University, London, Ontario, Canada, 6Department of Oncology, University of Western Ontario, London, Canada, 7Department of Microbiology and Immunology, University of Western Ontario, London, Canada, 8Department of Urology, University of California, Los Angeles, USA, 9Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research, University of California, Los Angeles, USA and 10Jonsson Comprehensive Cancer Center, University of California, Los Angeles, USA Dr A Nichols takes responsibility for the integrity of the content of the paper Cite this article: Patel KB, Prokopec S, Barrett JW, Mymryk JS, Boutros PC, Nichols AC. Establishment of four head and neck squamous cell carcinoma cell lines: importance of reference DNA for accurate genomic characterisation. J Laryngol Otol 2023;137:301–307. https://doi.org/10.1017/ S0022215122000846 Establishment of four head and neck squamous cell carcinoma cell lines: importance of reference DNA for accurate genomic characterisation cambridge.org/jlo K B Patel1 , S Prokopec2,3,4, J W Barrett5,6, J S Mymryk5,6,7, P C Boutros3,4,8,9,10 and A C Nichols5,6 K B Patel1 , S Prokopec2,3,4, J W Barrett5,6, J S Mymryk5,6,7, P C Boutros3,4,8,9,10 and A C Nichols5,6 Key words: Key words: Mutation; Cell Line; Head And Neck Squamous Cell Carcinoma; Human Papillomavirus; HPV-Positive Cell Lines; HPV Methods. A cell culture based, in vitro study was conducted of patients with primary human papillomavirus positive and human papillomavirus negative tumours. Abstract Objective. There is significant interest in developing early passage cell lines with matched normal reference DNA to facilitate a precision medicine approach in assessing drug response. This study aimed to establish early passage cell lines, and perform whole exome sequencing and short tandem repeat profiling on matched normal reference DNA, primary tumour and corresponding cell lines. Accepted: 18 March 2022 First published online: 23 March 2022 Matched cell line sequencing, tumours and reference blood DNA samples were submitted to The Centre for Applied Genomics for whole exome sequencing. FASTQ files were aligned to GRCh38 using Burrows–Wheeler aligner algorithm with maximal exact matches (version 0.7.15). Duplicates were marked. Indel realignment and recalibration were performed using the Genome Analysis Toolkit (version 3.7.0). For tumours with a matched normal, germline single nucleotide polymorphisms were called using the Genome Analysis Toolkit HaplotypeCaller (version 3.7.0), and filtered for quality and read depth. The SomaticSniper (version 1.0.5.0) genome modelling tool was used to call single nucleotide variants in paired samples (tumour or cell line), with the ref- erence being leukocyte DNA. A hypergeometric test was used to assess the probability of overlap between tumour and cell line samples, with p > 0.05 indicative of both samples having a greater overlap than expected by chance. The medium was changed every 72 hours, and evaluated by microscopy for epithelial cell adherence and growth. If the cells were able to achieve confluence and clusters of cells appeared epithelial, passaging was attempted; if no yeast or fungus contamination was evident, the cells were transferred to cell line complete growth medium, which consisted of Dulbecco’s modified Eagle’s medium/F12 medium supple- mented with 10 per cent fetal bovine serum and penicillin and streptomycin mixture. This was performed by incubation with 1 ml of trypsin (0.25 per cent, Wisent) for 2–5 minutes, followed by re-plating of 100 per cent of the cells into T75 flasks. This process was repeated for cell lines that continued to be able to achieve confluence and continued to have an epi- thelial appearance of greater than 50 per cent of cells. Representative stocks of each cell line were collected and fro- zen at −80°C during the passaging process in filter-sterilised freezing medium comprising 85 per cent fetal bovine serum and 15 per cent dimethyl sulfoxide. g p p y Germline single nucleotide polymorphisms were generated using the Genome Analysis Toolkit (version 3.7.0). HaplotypeCaller was then run on the realigned and recali- brated individual binary alignment maps, followed by geno- typing, variant selection and hard filtering. Unpaired mutation calling was first performed using the MuTect identi- fication algorithm (version 1.1.7), without supplying the paired normal for the tumour, and the cell line was realigned with recalibrated binary alignment maps for each patient. Patient recruitment and tumour processing Patients with head and neck SCC undergoing biopsy or resec- tion were consented for tumour collection with University of Western Ontario Health Sciences Research Ethics Board approval 16579E. Introduction Clinical trials are the ‘gold standard’ for establishing the efficacy of anti-cancer agents. However, only a limited number of hypotheses can be tested because of the high expense, the limited number of patients and the length of time necessary to complete these stud- ies.1 The realisation that each cancer site consists of multiple molecular subtypes that may respond differently to a particular targeted agent further complicates trials,2 making com- panion biomarkers a high priority. Model systems are therefore increasingly necessary to test the molecular basis of drug activity, which cannot be explored directly in patients. g y y Cancer cell lines are the most widely used model systems in cancer research; however, they are imperfect, as they lack a three-dimensional environment, stromal interactions, tumour-microenvironmental features like hypoxia, and immune system effects.3,4 Cell lines do, however, present with several advantages, including the fact that they cost rela- tively little to maintain, can be easily manipulated in biological assays, are relatively amen- able to genetic manipulation and provide rapid results. Indeed, they are the only model system that can be feasibly screened with a large number of agents to thoroughly explore mechanisms of drug sensitivity and resistance.5 In particular, there have been large-scale cell line drug screening efforts to correlate genetic profiles with drug response.6–8 Although there are well-established high-passage cell lines, patient-matched reference DNA is not available for accurate characterisation. Early passage cell lines are thought to better retain the molecular features of the primary tumours and have a stronger correl- ation with patient drug response.9,10 They have the potential additional advantage that patient-matched reference DNA and the primary tumour may be available to confirm that the cell line was indeed derived from the source patient and improved the accuracy of the genomic profiling. For these reasons, we describe the establishment and characterisation of four head and neck cancer cell lines, including two human papillomavirus (HPV)-positive head and neck squamous cell carcinoma (SCC) cell lines. We confirmed the origin of all the cell lines through short tandem repeat profiling, and used whole exome sequencing of triplet 302 K B Patel, S Prokopec, J W Barrett et al. K B Patel, S Prokopec, J W Barrett et al. (Toronto, Canada), as we have previously reported.11 The lines were genotyped with a panel of 10 selected markers. DNA extraction DNA was extracted from cultured cells and matched primary tumours using the AllPrep DNA/RNA Kit (Qiagen, Toronto, Canada), following the instructions provided by the manufac- turer. The DNA was extracted from the buffy coat layer sepa- rated from whole blood using the QIAamp DNA Blood Mini Kit (Qiagen). Minimum quality requirements for sequencing studies were 1 μg of genomic DNA with an optical density 260/280 ratio of between 1.7 and 1.9. Matched cell line sequencing, tumours and reference blood A panel of ‘normals’ was generated using MuTect (version 1.1.7) from 438 of Cancer Genome Atlas head and neck SCC normal exomes. The single nucleotide variants, unpaired with the panel of ‘normals’, were further filtered to retain only the ‘pass’ calls. The single nucleotide variants that were in the panel of ‘normals’ but were also in the ‘allowlist’ (because of being in the Catalogue of Somatic Mutations in Cancer (‘COSMIC’) database) were also removed. Single nucleotide variants in the capture regions and in chromosomes chr 1-22, X, Y and M were preserved. Somatic single nucleotide variants were filtered to remove non-functional (intronic or intergenic) variants prior to downstream analyses. Average Introduction The cell line short tandem repeat profiling results were com- pared with the corresponding primary tumour tissue and blood to confirm the identity of the sample. Cell line DNA and RNA were also tested for the presence of all high-risk HPVs (types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68, 69 and 74) by real-time polymerase chain reaction, as we have previously described.12 (Toronto, Canada), as we have previously reported.11 The lines were genotyped with a panel of 10 selected markers. The cell line short tandem repeat profiling results were com- pared with the corresponding primary tumour tissue and blood to confirm the identity of the sample. Cell line DNA and RNA were also tested for the presence of all high-risk HPVs (types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68, 69 and 74) by real-time polymerase chain reaction, as we have previously described.12 samples (normal, primary tumour, cell line) to compare the genetic landscape of the cell line with the primary tumour and determined the importance of the normal reference DNA for making accurate mutational calls. Immunoblotting Whole cell lysates were prepared, immunoblotted and analysed as described previously.13 Membranes were exposed to chemilu- minescence reagent (Luminata™Crescendo or Luminata™Forte (Millipore®)) using a Bio-Rad ChemiDoc™MP Imaging System. pp Ten millilitres of blood were drawn intra-operatively into heparinised tubes. The diagnosis of viable head and neck SCC was confirmed by frozen section in the operating theatre, and a portion of the remaining specimen was transported immediately in saline for processing. A portion of tumour was stored at −80°C. The remainder of the sample was further rinsed with saline and then minced into as small pieces as pos- sible with a sterile razor blade. The tissue was incubated in 3 ml of Dulbecco’s modified Eagle’s medium/F12 medium supplemented with 10 per cent fetal bovine serum (Gibco®), penicillin (100 IU/ml) and streptomycin (100 μg/ml; Wisent, Quebec, Canada), amphotericin (0.25 mg/ml; Wisent) and col- lagenase (prepared to 1 mg/ml; Sigma®) for 1 hour at 37°C with regular agitation. Cells were washed twice with phosphate buffered saline, resuspended in primary tumour cell growth medium and divided into a six-well dish. Cell line morphology and doubling time Photographs of all cell lines were taken at 4× and 10× magni- fication. In order to measure cell line doubling time, cells were counted and seeded into T25 flasks. Forty-eight hours later, the cells were collected and counted. Doubling time was calcu- lated using an online calculator.14 Short tandem repeat profiling and virus detection For each cell line that was developed, the corresponding matched blood and primary tumour were analysed by short tandem repeat profiling at The Centre for Applied Genomics 303 The Journal of Laryngology & Otology expression of HPV16 E7, although the signal was weak. In contrast, UWO23 E7 expression was not detected by this HPV16-specific antibody (Figure 1a). true positive, false positive and false negative rates were then calculated for the four cell lines. Data visualisation employed the BPG (BoutrosLab.plotting.general) package.15 Although predominantly epithelial, cell line morphology varied (Figure 2). UWO17 cells grew in defined, loose colonies until they merged together into monolayers with even cell-to-cell spacing. They doubled in number every 40 ± 4 hours (mean ± standard deviation) (Table 1). UWO23 cells exhibited an epithelial morphology and grew in tight mono- layers. They doubled every 50 ± 6 hours. When passaged at low density, UWO31 cells grew in tight clusters that developed satellites of cells as the colonies spread. This line doubled approximately every 62 ± 6 hours. Stromal-like cells, in add- ition to epithelial cells, survived in these cultures. UWO37 grew as colonies of cells, and maintained this pattern even at high density. They doubled in number approximately every 41 ± 4 hours. Cell lines, tissue of origin and viral status This section concerns the establishment of cell lines, confirm- ation of tissue of origin and HPV status. Short tandem repeat profiling of the patient’s blood, the primary tumour and the developed cell line demonstrated a match for the majority of all 10 markers, confirming that the cell lines were indeed derived from the index patient and tumour (Appendix 1, avail- able on The Journal of Laryngology & Otology website). In all cases, the short tandem repeat profiles matched exactly between the blood and primary tumour DNA; however, there were occasional differences at a single marker between the cell line and the parental tumour samples, likely because of minor drift as the cell line adapted to a plastic environment. We tested the cell lines that we were able to establish for the presence of HPV DNA and transcripts. UWO23 was positive for HPV type 33 and UWO37 was positive for HPV type 16 on real-time polymerase chain reaction and Sanger sequen- cing. In addition, both lines expressed high levels of p16 on western blotting, as is commonly observed for HPV-positive head and neck cancers (Figure 1). UWO37 was positive for This section concerns the establishment of cell lines, confirm- ation of tissue of origin and HPV status. Short tandem repeat profiling of the patient’s blood, the primary tumour and the developed cell line demonstrated a match for the majority of all 10 markers, confirming that the cell lines were indeed derived from the index patient and tumour (Appendix 1, avail- able on The Journal of Laryngology & Otology website). In all cases, the short tandem repeat profiles matched exactly between the blood and primary tumour DNA; however, there were occasional differences at a single marker between the cell line and the parental tumour samples, likely because of minor drift as the cell line adapted to a plastic environment. Matched reference DNA for cell line mutation calls We tested the cell lines that we were able to establish for the presence of HPV DNA and transcripts. UWO23 was positive for HPV type 33 and UWO37 was positive for HPV type 16 on real-time polymerase chain reaction and Sanger sequen- cing. In addition, both lines expressed high levels of p16 on western blotting, as is commonly observed for HPV-positive head and neck cancers (Figure 1). UWO37 was positive for Matched reference DNA was essential for accurate cell line mutation calls. Whole exome sequencing was undertaken for the four tumours and their corresponding cell lines, and compared against the patient’s reference leukocyte DNA (patient-matched). This was accomplished using the MuTect algorithm, where matched leukocyte DNA was employed as Fig. 1. Determination of human papillomavirus (HPV) status in UWO23 and UWO37. (a) UWO23 and UWO37 were confirmed to express p16. UPCI:SCC090 was included as an HPV-positive control. The NM2 clone of the HPV16-specific antibody confirmed that UWO37 (and UPCI:SCC090) were HPV16-positive. The antibody did not cross-react with UWO23 (HPV33). Tubulin acted as a loading control. (b) Reverse transcription polymerase chain reaction confirmed that the parental tumour (325T) and derived cell line (UWO23) were positive for HPV33 E6 and E7 transcripts when compared to the positive control (HPV33). 382T (patient tumour) and UWO33 were a matched tumour/cell line pair not described in the manuscript, but used as an HPV-negative control for the specificity of the primers. (c) Reverse transcription polymerase chain reaction of the parental (pt 485) tumour and clones of the UWO37 cell line were positive HPV16 E7 transcripts. No HPV16 E7 transcripts were observed for UWO23 (HPV33 positive cell line), the negative control SCC61 or the ‘no template control’ (NTC). (d) UWO23 and (e) UWO37 alignment to E6 region of the appropriate HPV Genbank reference sequences. In (d), the UWO23 sequence is presented at the top and the HPV33 (HQ537706) sequence is aligned below. In (e), the UWO37 sequence is presented at the top and HPV16 (NC_001526) is aligned below. Dots indicate an exact nucleo- tide match at that position. Fig. 1. Determination of human papillomavirus (HPV) status in UWO23 and UWO37. (a) UWO23 and UWO37 were confirmed to express p16. UPCI:SCC090 was included as an HPV-positive control. The NM2 clone of the HPV16-specific antibody confirmed that UWO37 (and UPCI:SCC090) were HPV16-positive. The antibody did not cross-react with UWO23 (HPV33). Tubulin acted as a loading control. Matched reference DNA for cell line mutation calls (b) Reverse transcription polymerase chain reaction confirmed that the parental tumour (325T) and derived cell line (UWO23) were positive for HPV33 E6 and E7 transcripts when compared to the positive control (HPV33). 382T (patient tumour) and UWO33 were a matched tumour/cell line pair not described in the manuscript, but used as an HPV-negative control for the specificity of the primers. (c) Reverse transcription polymerase chain reaction of the parental (pt 485) tumour and clones of the UWO37 cell line were positive HPV16 E7 transcripts. No HPV16 E7 transcripts were observed for UWO23 (HPV33 positive cell line), the negative control SCC61 or the ‘no template control’ (NTC). (d) UWO23 and (e) UWO37 alignment to E6 region of the appropriate HPV Genbank reference sequences. In (d), the UWO23 sequence is presented at the top and the HPV33 (HQ537706) sequence is aligned below. In (e), the UWO37 sequence is presented at the top and HPV16 (NC_001526) is aligned below. Dots indicate an exact nucleo- tide match at that position. Fig. 1. Determination of human papillomavirus (HPV) status in UWO23 and UWO37. (a) UWO23 and UWO37 were confirmed to express p16. UPCI:SCC090 was included as an HPV-positive control. The NM2 clone of the HPV16-specific antibody confirmed that UWO37 (and UPCI:SCC090) were HPV16-positive. The antibody Fig. 1. Determination of human papillomavirus (HPV) status in UWO23 and UWO37. (a) UWO23 and UWO37 were confirmed to express p16. UPCI:SCC090 was included as an HPV-positive control. The NM2 clone of the HPV16-specific antibody confirmed that UWO37 (and UPCI:SCC090) were HPV16-positive. The antibody did not cross-react with UWO23 (HPV33). Tubulin acted as a loading control. (b) Reverse transcription polymerase chain reaction confirmed that the parental tumour (325T) and derived cell line (UWO23) were positive for HPV33 E6 and E7 transcripts when compared to the positive control (HPV33). 382T (patient tumour) and UWO33 were a matched tumour/cell line pair not described in the manuscript, but used as an HPV-negative control for the specificity of the primers. (c) Reverse transcription polymerase chain reaction of the parental (pt 485) tumour and clones of the UWO37 cell line were positive HPV16 E7 transcripts. No HPV16 E7 transcripts were observed for UWO23 (HPV33 positive cell line), the negative control SCC61 or the ‘no template control’ (NTC). (d) UWO23 and (e) UWO37 alignment to E6 region of the appropriate HPV Genbank reference sequences. Discussion Fig. 2. Cell morphology of the UWO primary cell lines. The four lines are shown at high magnification (10×) to illustrate the individual cell phenotype and colony organ- isation. The inset shows the cells at a higher density (lower magnification, 4×). We have presented the establishment and molecular character- isation of four novel early passage head and neck SCC cell lines, including two HPV-positive lines. Importantly, we demonstrated that the reference patient-matched DNA from each patient is important to make accurate mutation calls. These patient-matched reference samples, as demonstrated in this study, allow accurate characterisation of cell lines. Furthermore, as observed in this study, bioinformatics pipe- lines are prone to errors with high false positive rates (87.6 per cent in our study). While cell lines and primary tumours shared significant numbers of specific mutations, demonstrat- ing that they are related, both contained unique mutations, suggesting that genetic drift occurred in the cell lines as they adapted to culture conditions. It is also worth noting that the mutational calls between normal tissue and primary tumours that are shared are sufficient, but not necessary, for tumour transformation. These would be missed when using reference DNA. a reference versus using a panel of ‘normals’ generated from 438 Cancer Genome Atlas head and neck SCC samples (as described in the Methods section); this is in contrast to cell line mutation calls, which were performed against the matched leukocyte DNA. When we performed unmatched analysis, we observed an average of 314.5 true positive single nucleotide variants, 20.5 false negative single nucleotide variants and 2222.75 false positives relative to the gold standard paired ana- lysis (Table 2). The false positive rate was 87.6 per cent and the false negative rate was 6.1 per cent. Matched reference DNA for cell line mutation calls False positive and false negative mutational calls of cell lines identified using MuTect versus MuTect unpaired* Values reflect average numbers of mutational calls. With reference patient matched. N/A = not applicable Values reflect average numbers of mutational calls. With reference patient matched. N/A = not applicable of mutations in the cell lines and tumours is provided in Appendix 2, available on The Journal of Laryngology & Otology website. The hypergeometric distribution test was then applied to each tumour and cell line pair to assess whether two samples shared the same lineage based on muta- tion calls. For each of the paired samples, the p-value was not significant, suggesting that they did in fact share significantly more mutations than would occur by chance alone (Figure 4). Matched reference DNA for cell line mutation calls In (d), the UWO23 sequence is presented at the top and the HPV33 (HQ537706) sequence is aligned below. In (e), the UWO37 sequence is presented at the top and HPV16 (NC_001526) is aligned below. Dots indicate an exact nucleo- tide match at that position. Fig. 1. Determination of human papillomavirus (HPV) status in UWO23 and UWO37. (a) UWO23 and UWO37 were confirmed to express p16. UPCI:SCC090 was included as an HPV-positive control. The NM2 clone of the HPV16-specific antibody confirmed that UWO37 (and UPCI:SCC090) were HPV16-positive. The antibody did not cross-react with UWO23 (HPV33). Tubulin acted as a loading control. (b) Reverse transcription polymerase chain reaction confirmed that the parental tumour (325T) and derived cell line (UWO23) were positive for HPV33 E6 and E7 transcripts when compared to the positive control (HPV33). 382T (patient tumour) and UWO33 were a matched tumour/cell line pair not described in the manuscript, but used as an HPV-negative control for the specificity of the primers. (c) Reverse transcription polymerase chain reaction of the parental (pt 485) tumour and clones of the UWO37 cell line were positive HPV16 E7 transcripts. No HPV16 E7 transcripts were observed for UWO23 (HPV33 positive cell line), the negative control SCC61 or the ‘no template control’ (NTC). (d) UWO23 and (e) UWO37 alignment to E6 region of the appropriate HPV Genbank reference sequences. In (d), the UWO23 sequence is presented at the top and the HPV33 (HQ537706) sequence is aligned below. In (e), the UWO37 sequence is presented at the top and HPV16 (NC_001526) is aligned below. Dots indicate an exact nucleo- tide match at that position. 304 K B Patel, S Prokopec, J W Barrett et al. Fig. 2. Cell morphology of the UWO primary cell lines. The four lines are shown at high magnification (10×) to illustrate the individual cell phenotype and colony organ- isation. The inset shows the cells at a higher density (lower magnification, 4×). Table 2. False positive and false negative mutational calls of cell lines identified using MuTect versus MuTect unpaired* MuTect unpaired MuTect paired Positive Negative Positive 314.5 2222.75 Negative 20.5 N/A Values reflect average numbers of mutational calls. With reference patient matched. N/A = not applicable Table 2. False positive and false negative mutational calls of cell lines identified using MuTect versus MuTect unpaired Table 2. Mutational landscape of early passage cell lines Tumour and corresponding cell lines with reference patient matched DNA were analysed using SomaticSniper software (Figure 3a). For each corresponding tumour and cell line pair, the cell line exhibited more mutations than the contrib- uting tumour, perhaps reflecting changes necessary for cells to adapt to culture conditions. When compared with the top 25 genes identified in the Cancer Genome Atlas marker paper,16 single nucleotide variants in NOTCH1, TP53 and CDKN2A genes were identified (Figure 3b). A complete list There is an urgent need to identify novel chemicals with high activity in head and neck cancers in order to improve outcomes. Cancer cell lines are by far the most widely used model systems in cancer research, as they cost relatively little to maintain and can be easily manipulated in biological assays and feasibly screened with a large number of agents. Potentially, genomic characterisation of these cell lines could allow correlation of molecular features with drug response. Table 1. Patient of origin details and cell line doubling time Cell line Site Sex Age at diagnosis (years) HPV status Tumour (T) stage Nodal (N) stage Treatment Follow-up duration (years) Last known status Cell line doubling time (mean ± SD; hours) UWO17 Oral tongue F 58 Negative 2 3b Surgery + chemoradiation 7 NED 40 ± 4 UWO23 Oral tongue M 52 Positive 4 0 Surgery + chemoradiation 1.3 NED 50 ± 6 UWO31 Oral tongue F 68 Negative 3 2a Surgery + radiation 3.4 NED 62 ± 6 UWO37 Recurrent tonsil after chemoradiation M 64 Positive 4 1 Surgery 1.1 DOD 41 ± 4 HPV = human papillomavirus; SD = standard deviation; F = female; M = male; NED = no evidence of disease; DOD = died of disease Table 1. Patient of origin details and cell line doubling time Table 1. Patient of origin details and cell line doubling time HPV = human papillomavirus; SD = standard deviation; F = female; M = male; NED = no evidence of disease; DOD = died of disease 305 The Journal of Laryngology & Otology Fig. 3. Mutational comparison of primary tumours and cell lines. (a) Heatmap of all single nucleotide variants identified by whole exome sequencing in four cell lines (CL) and their matched parental tumours (T). A full gene list is provided in Appendix 2 (available on The Journal of Laryngology & Otology website). Mutational landscape of early passage cell lines (b) Single nucleotide variants identified in the top 20 most frequently mutated genes in the Cancer Genome Atlas head and neck squamous cell carcinoma study. Note that both human papillomavirus positive cell lines were TP53 wild-type as expected. UTR = untranslated region • Short tandem repeat profiling confirmed that all four cell lines were established from their index tumours • Whole exome sequencing revealed that matched normal reference DNA was critical for accurate mutational analysis: a high rate of false positive mutation calls (87.6 per cent) were excluded The HPV-positive cell line models are of great interest to the head and neck research community in light of the epidemic of HPV-related oropharyngeal cancer.12,31 However, while numer- ous HPV-negative head and neck SCC cell lines exist, there is a paucity of HPV-positive lines.32 The ideal HPV-positive cell lines would be derived from non-smoking patients with a treatment-naïve oropharyngeal cancer; however, almost all lines, including our own, are derived from smokers, radiation failures (UWO37) or non-oropharyngeal primary sites (UWO23, oral cavity).32 Only recently has an ideal cell line been described.33 Further culture efforts are needed to generate additional models to facilitate drug discovery. Indeed, there have been two large-scale studies involving hun- dreds of genotyped cell lines of different tissue types,6,7 which have identified known, as well as novel, genetic markers of drug response.6,7 We specifically reanalysed the head and neck SCC data from both these studies and identified associa- tions between PIK3CA mutations and response to a PI3K inhibitor; however, our analysis was limited by the small num- ber of head and neck SCC lines and the limited number of drugs tested.17 g In order to expand drug discovery and develop additional biomarkers, we previously screened 1433 agents in a panel of 26 head and neck SCC established cell lines, and identified drugs with preferential activity in HPV-positive and HPV-negative cell lines.5,18 However, our original intention was to correlate the cell line genomic profiles with drug response. We characterised our cell line panel with both OncoScan™copy number arrays and whole exome sequen- cing, which our team has analysed extensively.19–25 While copy number arrays perform well without reference normal DNA,22,24,26 exome sequencing is prone to frequent false Mutational landscape of early passage cell lines (b) Single nucleotide variants identified in the top 20 most frequently mutated genes in the Cancer Genome Atlas head and neck squamous cell carcinoma study. Note that both human papillomavirus positive cell lines were TP53 wild-type as expected. UTR = untranslated region positives and negatives for both primary tumours and cell lines. In an effort to specifically address this, we developed a ‘reference-free somatic variant’ pipeline, which uses an exten- sive mutational ‘allowlist’ (known pathogenic variants) and ‘denylist’ (likely germline variants based on large population cohorts) as well as additional bioinformatics analysis.27 This pipeline is highly accurate in primary head and neck SCC and prostate tumours; however, despite exhaustive efforts, it does not perform well on cell lines.27 For this reason, we have started to generate a panel of both HPV-positive and HPV-negative cell lines from primary tumours, along with matched blood. To date, we have generated 17 such lines, and we have presented genomic analysis of the first 4 of these here. p g y An additional advantage of the cell lines described in this study is that they are early passage cell lines. Established cell lines have the distinct advantage of being broadly available to the cancer research community through established com- mercial culture collections. This offers the opportunity to reproduce data independently with the same reagents. Nevertheless, the correlation of established cell line drug responses with patient tumour responses appears to be modest,28–30 potentially because of additional molecular drift from the primary tumour profile through selection pressure in culture.9,10 Although beyond the scope of this manuscript with regard to comparison of early versus late passage cell lines, there is some evidence that early passage cell lines may represent superior models for drug screening.28–30 Thus, panels of genomically characterised early passage cell lines with patient-matched reference DNA, such as our novel University of Western Ontario panel, may be beneficial for the first phase of drug testing. • Two cell lines were human papillomavirus positive, confirmed by sequencing and P16 immunoblotting Fig. 3. Mutational comparison of primary tumours and cell lines. (a) Heatmap of all single nucleotide variants identified by whole exome sequencing in four cell lines (CL) and their matched parental tumours (T). A full gene list is provided in Appendix 2 (available on The Journal of Laryngology & Otology website). • Four early passage cell lines were established, and genomic characterisation was undertaken • Two cell lines were human papillomavirus positive, confirmed by sequencing and P16 immunoblotting • Short tandem repeat profiling confirmed that all four cell lines were established from their index tumours • Whole exome sequencing revealed that matched normal reference DNA was critical for accurate mutational analysis: a high rate of false positive mutation calls (87.6 per cent) were excluded Conclusion We describe the establishment and characterisation of four head and neck SCC early passage cell lines with patient- matched reference DNA. These cell lines will act as a resource for drug discovery, as they can be genomically characterised and may be more accurate than late passage models. Development of additional early passage lines, particularly HPV-positive models, is needed. 306 K B Patel, S Prokopec, J W Barrett et al. Fig. 4. Overlap of the single nucleotide variants between tumour and cell line samples. (a) Patient 152, cell line UWO17. (b) Patient 325, cell line UWO23. (c) Patient 370, cell line UWO31. (d) Patient 485, cell line UWO37. Hypergeometric tests showed p-values of more than 0.05 for all samples, demonstrating that the paired cell lines and the parental tumours are genetically related. Patient 152, cell line UWO17. (b) Patient 325, cell line UWO23. (c) Patient es of more than 0.05 for all samples, demonstrating that the paired cell Fig. 4. Overlap of the single nucleotide variants between tumour and cell line samples. (a) Patient 152, cell line UWO17. (b) Patient 325, cell line UWO23. (c) Patient 370, cell line UWO31. (d) Patient 485, cell line UWO37. Hypergeometric tests showed p-values of more than 0.05 for all samples, demonstrating that the paired cell lines and the parental tumours are genetically related. Acknowledgements. This work was financially supported by: a Canadian Institutes of Health Research grant (MOP 377832, awarded to ACN and PCB); the Ontario Institute for Cancer Research, through funding provided (to PCB) by the Government of Ontario; the Wolfe Surgical Research Professorship in the Biology of Head and Neck Cancers Fund (supporting ACN); Terry Fox Research Institute and Canadian Institutes of Health Research New Investigator Awards (supporting PCB); and the National Institutes of Health/National Cancer Institute (under award number P30CA016042). 5 Ghasemi F, Black M, Sun RX, Vizeacoumar F, Pinto N, Ruicci KM et al. High-throughput testing in head and neck squamous cell carcinoma iden- tifies agents with preferential activity in human papillomavirus-positive or negative cell lines. Oncotarget 2018;9:26064–71 6 Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 2012;483:603–7 g y 7 Garnett MJ, Edelman EJ, Heidorn SJ, Greenman CD, Dastur A, Lau KW et al. Systematic identification of genomic markers of drug sensitivity in cancer cells. Nature 2012;483:570–5 Supplementary material. Conclusion The supplementary material for this article can be found at https://doi.org/10.1017/S0022215122000846. 8 Klijn C, Durinck S, Stawiski EW, Haverty PM, Jiang Z, Liu H et al. A com- prehensive transcriptional portrait of human cancer cell lines. Nat Biotechnol 2015;33:306–12 Competing interests. KBP, SP, JWB and JSM do not have any relevant dis- closures. ACN is a consultant for Need Oncology, and has research funding from Novartis Canada, Sequenom, Merck Canada and Champions Oncology. PCB sits on the Scientific Advisory Boards of BioSymetrics, Intersect Diagnostics and Sage Bionetworks. 9 O’Driscoll L, Gammell P, McKiernan E, Ryan E, Jeppesen PB, Rani S et al. Phenotypic and global gene expression profile changes between low pas- sage and high passage MIN-6 cells. J Endocrinol 2006;191:665–76 10 Wenger SL, Senft JR, Sargent LM, Bamezai R, Bairwa N, Grant SG. Comparison of established cell lines at different passages by karyotype and comparative genomic hybridization. Biosci Rep 2004;24:631–9 11 Ruicci KM, Meens J, Sun RX, Rizzo G, Pinto N, Yoo J et al. A controlled trial of HNSCC patient-derived xenografts reveals broad efficacy of PI3Kα inhibition in controlling tumor growth. Int J Cancer 2019;145:2100–6 The Journal of Laryngology & Otology 307 15 P’ng C, Green J, Chong LC, Waggott D, Prokopec SD, Shamsi M et al. BPG: seamless, automated and interactive visualization of scientific data. BMC Bioinformatics 2019;20:42 24 Ewing AD, Houlahan KE, Hu Y, Ellrott K, Caloian C, Yamaguchi TN et al. Combining tumor genome simulation with crowdsourcing to benchmark somatic single-nucleotide-variant detection. Nat Methods 2015;12:623–30 25 Espiritu SMG, Liu LY, Rubanova Y, Bhandari V, Holgersen EM, Szyca LM et al. The evolutionary landscape of localized prostate cancers drives clinical aggression. Cell 2018;173:1003–13.e15 16 Cancer Genome Atlas Network. Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature 2015;517:576–82 17 Nichols AC, Black M, Yoo J, Pinto N, Fernandes A, Haibe-Kains B et al. Exploiting high-throughput cell line drug screening studies to identify can- didate therapeutic agents in head and neck cancer. BMC Pharmacol Toxicol 2014;15:66 26 Cooper CS, Eeles R, Wedge DC, Van Loo P, Gundem G, Alexandrov LB et al. Analysis of the genetic phylogeny of multifocal prostate cancer identifies multiple independent clonal expansions in neoplastic and morphologically normal prostate tissue. Nat Genet 2015;47:367–72 18 Ghasemi F, Black M, Vizeacoumar F, Pinto N, Ruicci KM, Le C et al. Repurposing Albendazole: new potential as a chemotherapeutic agent with preferential activity against HPV-negative head and neck squamous cell cancer. Oncotarget 2017;8:71512–19 27 Sun R, Lalansingh C, Espiritu S, Yao C, Yamaguchi T, Prokopec S et al. Accurate reference-free somatic variant-calling by integrating genomic, sequencing and population data. bioRxiv 2018. Epub 2018 August 2 19 Hua JT, Ahmed M, Guo H, Zhang Y, Chen S, Soares F et al. Risk SNP- mediated promoter-enhancer switching drives prostate cancer through lncRNA PCAT19. Cell 2018;174:564–75.e18 28 Domcke S, Sinha R, Levine DA, Sander C, Schultz N. Evaluating cell lines as tumour models by comparison of genomic profiles. Nat Commun 2013;4:2126 29 Baguley BC, Marshall ES. The use of human tumour cell lines in the dis- covery of new cancer chemotherapeutic drugs. Expert Opin Drug Discov 2008;3:153–61 20 Hopkins JF, Sabelnykova VY, Weischenfeldt J, Simon R, Aguiar JA, Alkallas R et al. Mitochondrial mutations drive prostate cancer aggression. Nat Commun 2017;8:656 30 Cree IA, Glaysher S, Harvey AL. Efficacy of anti-cancer agents in cell lines ver- sus human primary tumour tissue. Curr Opin Pharmacol 2010;10:375–9 21 Fraser M, Sabelnykova VY, Yamaguchi TN, Heisler LE, Livingstone J, Huang V et al. Genomic hallmarks of localized, non-indolent prostate can- cer. References 1 Weinstein JN. Drug discovery: cell lines battle cancer. Nature 2012;483: 544–5 1 Weinstein JN. Drug discovery: cell lines battle cancer. Nature 2012;483: 544–5 12 Nichols AC, Palma DA, Dhaliwal SS, Tan S, Theuer J, Chow W et al. The epidemic of human papillomavirus and oropharyngeal cancer in a Canadian population. Curr Oncol 2013;20:212–19 2 Simon R, Roychowdhury S. Implementing personalized cancer genomics in clinical trials. Nat Rev Drug Discov 2013;12:358–69 2 Simon R, Roychowdhury S. Implementing personalized cancer genomics in clinical trials. Nat Rev Drug Discov 2013;12:358–69 3 van Staveren WC, Solis DY, Hebrant A, Detours V, Dumont JE, Maenhaut C. Human cancer cell lines: experimental models for cancer cells in situ? For cancer stem cells? Biochim Biophys Acta 2009;1795:92–103 3 van Staveren WC, Solis DY, Hebrant A, Detours V, Dumont JE, Maenhaut C. Human cancer cell lines: experimental models for cancer cells in situ? For cancer stem cells? Biochim Biophys Acta 2009;1795:92–103 13 Ruicci KM, Pinto N, Khan MI, Yoo J, Fung K, MacNeil D et al. ERK-TSC2 signalling in constitutively-active HRAS mutant HNSCC cells promotes resistance to PI3K inhibition. Oral Oncol 2018;84:95–103 p y 4 Bhandari V, Hoey C, Liu LY, Lalonde E, Ray J, Livingstone J et al. Molecular landmarks of tumor hypoxia across cancer types. Nat Genet 2019;51:308–18 p y 4 Bhandari V, Hoey C, Liu LY, Lalonde E, Ray J, Livingstone J et al. Molecular landmarks of tumor hypoxia across cancer types. Nat Genet 2019;51:308–18 14 Roth V. Doubling Time Computing, 2006. In: http://www.doubling-time. com/compute.php [10 November 2022] The Journal of Laryngology & Otology The Journal of Laryngology & Otology Nature 2017;541:359–64 31 Chaturvedi AK, Engels EA, Pfeiffer RM, Hernandez BY, Xiao W, Kim E et al. Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol 2011;29:4294–301 22 Boutros PC, Fraser M, Harding NJ, de Borja R, Trudel D, Lalonde E et al. Spatial genomic heterogeneity within localized, multifocal prostate cancer. Nat Genet 2015;47:736–45 32 Cheng H, Yang X, Si H, Saleh AD, Xiao W, Coupar J et al. Genomic and transcriptomic characterization links cell lines with aggressive head and neck cancers. Cell Rep 2018;25:1332–45.e5 23 Weinreb I, Piscuoglio S, Martelotto LG, Waggott D, Ng CK, Perez-Ordonez B et al. Hotspot activating PRKD1 somatic mutations in polymorphous low-grade adenocarcinomas of the salivary glands. Nat Genet 2014;46:1166–9 33 Forslund O, Sugiyama N, Wu C, Ravi N, Jin Y, Swoboda S et al. A novel human in vitro papillomavirus type 16 positive tonsil cancer cell line with high sensitivity to radiation and cisplatin. BMC Cancer 2019;19:265
https://openalex.org/W4214862603	https://www1.unicap.br/ojs/index.php/paralellus/article/download/1959/pdf	Portuguese	null	A DISPONIBILIDADE E INDISPONIBILIDADE DA SABEDORIA: A DIALÉTICA DA SABEDORIA EM PROVÉRBIOS E JÓ	Revista Paralellus	2,021	cc-by	6,587	RESUMO O artigo analisa as estratégias literárias utilizadas em Provérbios 1,20-33; 8,1-36; 9,1-6 e Jó 28,1-28 comparando como esses diferentes poemas tematizam a sabedoria a fim de caracterizar a disponibilidade e indisponibilidade dela aos seres humanos. A escolha desses textos se dá por critérios objetivos: 1. Neles, a disponibilidade e os caminhos para encontrar a sabedoria são enunciados; 2. Quanto à datação, os textos são, aparentemente, contemporâneos, datados entre o fim do período de dominação persa e o início da dominação grega; 3. Os textos fazem parte da edição final dos respectivos livros. No caso de Provérbios, no conjunto de poemas em Pv 1-9. No caso de Jó, como edição entre os diálogos de Jó e seus amigos (3-27) e a intervenção divina (38-41); 4. Os textos desempenham um papel literário-teológico importante em sua localização no livro. Por isso, analisa também a função dessa caracterização na edição dos livros de Provérbios e Jó. Por fim, a análise da dialética literária da disponibilidade e indisponibilidade da sabedoria contribui a para compreensão da mesma dialética na experiência humana. Palavras-chave: Bíblia Hebraica; Literatura sapiencial; Análise literário-teológica. THE AVAILABILITY AND UNAVAILABILITY OF WISDOM: THE DIALECTICS OF WISDOM IN PROVERBS AND JOB Luiz Alexandre Solano Rossi* Lucas Merlo Nascimento** * Doutor em Ciências da Religião pela Universidade Metodista de São Paulo. Professor da Pontifícia Universidade Católica do Paraná e do Centro Universitário Internacional (UNINTER). ** Doutorando em Teologia pela PUC-PR. Mestre em Ciências da Religião, com ênfase em Literatura e religião no mundo bíblico, pela Universidade Metodista de São Paulo (UMESP). Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 Esta obra está licenciada sob uma licença Creative Commons Enviado: 05-2021 * Aprovado: 08-2021 Universidade Católica do Paraná e do Centro Universitário Internacional (UNINTER). ** Doutorando em Teologia pela PUC-PR. Mestre em Ciências da Religião, com ênfase em Literatura e religião no mundo bíblico, pela Universidade Metodista de São Paulo (UMESP). ABSTRACT 433 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 Esta obra está licenciada sob uma licença Creative Commons Enviado: 05-2021 * Aprovado: 08-2021 * Doutor em Ciências da Religião pela Universidade Metodista de São Paulo. Professor da Pontifícia Universidade Católica do Paraná e do Centro Universitário Internacional (UNINTER). ** Doutorando em Teologia pela PUC-PR. Mestre em Ciências da Religião, com ênfase em Literatura e religião no mundo bíblico, pela Universidade Metodista de São Paulo (UMESP). utor em Ciências da Religião pela Universidade Metodista de São Paulo. Professor da Pontifíc versidade Católica do Paraná e do Centro Universitário Internacional (UNINTER). ( ) utorando em Teologia pela PUC-PR. Mestre em Ciências da Religião, com ênfase em Literatu eligião no mundo bíblico, pela Universidade Metodista de São Paulo (UMESP). The article analyzes the literary strategies used in Proverbs 1:20-33; 8:1-36; 9:1-6 and Job 28:1-28 comparing how these different poems thematize wisdom in order to characterize its availability and unavailability to humans. The choice of these texts is based on objective criteria: 1. in them the availability and the ways to find wisdom are enunciated; 2. as for dating the texts are contemporary, dated between the end of the period of Persian domination and the beginning of Greek domination; 3. the texts are part of the final edition of the respective books, in collection of poems in Prov. 1-9 and in Job as an edition between the dialogues of Job and his friends (3-27) and the divine intervention (38-41); 4. The texts play an important literary-theological role in their location in the book. Therefore it also analyzes the function of this characterization in the editing of the books of Proverbs and Job. Finally, the analysis of the literary dialectic of the availability and unavailability of wisdom contributes to an understanding of the same dialectic in human experience. Keywords: Hebrew Bible; wisdom literature; literary-theological analysis. Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 435 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 1 A tradução da obra de Lorenzin para a Língua Portuguesa grafa “Eclesiastes”, provavelmente equívoco do tradutor. 2 O autor se refere especificamente aos poemas encontrados em 1,8-19; 2,1-22; 3,1-12; 3,21-35; 4,1- 9; 4,10-19; 4,20-27; 5,1-23; 6,20-35; 7,1-27. A esses se acrescentam os poemas de personificação da sabedoria (1,20-33; 8,1-36; 9,1-6.13-18) que lhes são complementar (PINTO, 2018, p.13-14) 1. INTRODUÇÃO O presente artigo tem como objetivo esclarecer as estratégias literárias utilizadas nos textos da tradição sapiencial da Bíblia Hebraica em que a sabedoria é apresentada paradoxalmente como disponível e indisponível ao ser humano, a fim de compreender a função dessa caracterização para os livros de Provérbios e Jó. Para tanto, são abordados os textos de Provérbios 1,20-33; 8,1-36; 9,1-6 e Jó 28,1-28. A abordagem aos textos indicados não se dá por meio de uma exegese exaustiva, antes, por meio do levantamento sumário de dados textuais que indicam a disponibilidade e indisponibilidade da sabedoria. A análise fundamenta-se em exegetas de tradição católica e protestante. A escolha desses textos se dá por critérios objetivos: 1. Neles, a disponibilidade e os caminhos para encontrar a sabedoria são tematizados; 2. Do ponto de vista da datação, os textos são, aparentemente, contemporâneos; 3. Os textos fazem parte da edição final dos respectivos livros; 4. Os textos desempenham um papel literário-teológico importante em sua localização no livro. Os capítulos iniciais do livro de Provérbios (1-9), assim como o final (31) são as partes mais recentes do livro. Servem como “introdução” à coletânea dos provérbios populares (10-30), conforme Líndez (2011, p.66-67). Descrevendo de forma geral a formação do livro, Lorenzin (2020, p.39) propõe: Num primeiro momento, no período pré-exílico, existiam as coleções de provérbios populares em relação à família e ao clã. Num segundo momento, no final do século VIII a.C., sob o reinado de Ezequias (25,1), foram encorajadas a coleta, a reprodução, a adaptação e a 434 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 ampliação do material antigo. O terceiro momento diz respeito à redação do editor, provavelmente ocorrida no período pós-exílico (IV- III a.C.), e seguramente antes da formação do livro do Eclesiástico1 (por volta do ano 190 a.C.), que em 47,17 faz referência a Pv 1,6. O redator final completou o livro acrescentando Pv 1-9 como introdução, e 31,10-31 como conclusão. Konrad Schmid (2013, p.231) argumenta que Pv 1-9 pode ser compreendido a partir do pano de fundo da teologia de Isis, propagada pelos ptolomeus, corroborando a datação do texto no período helenista. Para ele, Pv 1-9 apresenta-se como sabedoria em roupagem teológica, já desenvolvida. Os sinais desse desenvolvimento são: 1. O caminho para a sabedoria passa pelo temor a YHWH; 2. A sabedoria encontra-se personificada, inclusive presente na criação do mundo; 3. Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 1. INTRODUÇÃO A visão otimista de retribuição daquele que possui a sabedoria, possivelmente sob influência da teologia deuteronomista (SCHMID, 2013, p.232-233). Konrad Schmid (2013, p.231) argumenta que Pv 1-9 pode ser compreendido a partir do pano de fundo da teologia de Isis, propagada pelos ptolomeus, corroborando a datação do texto no período helenista. Para ele, Pv 1-9 apresenta-se como sabedoria em roupagem teológica, já desenvolvida. Os sinais desse desenvolvimento são: 1. O caminho para a sabedoria passa pelo temor a YHWH; 2. A sabedoria encontra-se personificada, inclusive presente na criação do mundo; 3. A visão otimista de retribuição daquele que possui a sabedoria, possivelmente sob influência da teologia deuteronomista (SCHMID, 2013, p.232-233). Pinto (2018, p. 25-26) identifica Pv 1-9 com o período persa tardio ou helenista. Para ele, apontam nesta direção os paralelos entre a figura da mulher estrangeira em Pv 1- 9 e o perigo do casamento com mulheres não judias em Esdras e Neemias (Ed-10, Ne 13). Também a forma de instrução que se encontra em Pv 1-9, segundo o autor, aproxima-se da retórica greco-romana. Por isso “é plausível pensar que o(s) redator(es) de Pr 1-9 tenha(m) querido redigir um texto impregnado de tradições judaicas, mas, no que se refere à estrutura, em forma helenística2.” O autor indica uma datação “provavelmente por volta de 350 a.C.: no final do período persa e antes, certamente, do escrito de Coélet (cerca de 250 a.C.), do qual ignora as angústias existenciais e religiosas” (PINTO, 2018, p.26). Especificamente sobre a relação dos poemas de Pv 1,8-9,18 e sua relação interna e com o restante do livro, Pinto (2018, p.13) escreve: Ao prólogo segue-se a primeira coleção (1,8-9,18), que apresenta uma divisão interna diferente em relação às outras: não se ilustram simples ditos separados um dos outros e/ou apenas levemente ligados ao contexto do capítulo em que se encontram, mas se oferece uma série 435 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 de poemas em estreita relação com os capítulos do livro da mesma coleção e com o resto do livro. de poemas em estreita relação com os capítulos do livro da mesma coleção e com o resto do livro. Mesmo uma postura exegética mais conservadora, reconhece-se o papel do editor final e sua localização no período persa ou helenista, que teria permitido à obra permanecer debaixo do nome de Salomão, que seria o autor principal (WALTKE, 2011, p.77). Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 3 Cf. ainda HEINEM, 1982, p.91 1. INTRODUÇÃO Kidner (1980, p.22-27) também entende que a atribuição a Salomão em Pv 1,1 seja editorial. Discute o pano de fundo canaanita de Pv 8 e 9, que apontaria para certa antiguidade, contra a tendência de atribuir ao período persa ou grego, mas reconhece que os caps. 1-9 foram colocados por um redator final. Assim, pode-se notar que os poemas contidos em Pv 1-9 fazem parte da edição final do livro e mantém, portanto, uma função estrutural em relação ao mesmo. Konrad Schmid (2013, p.234) localiza Jó 28 no período helenístico, na última edição do livro e reconhece sua “posição peculiar” no livro junto aos discursos de Eliú em Jó 32-37. Para ele, esses textos tenderiam a trazer novamente o livro de Jó a uma “ortodoxia moderada”. Neste sentido também apontam Rossi e Nascimento (2021, p.10). Especificamente “O hino da sabedoria em Jó 28, que, anexado a Jó 27, aparece construído como discurso de Jó, afasta a possível compreensão de que nos diálogos Jó estaria recusando totalmente sabedoria e temor a Deus” (SCHMID, 2013, p.235). Jean Leveque (1987, p.8) aponta na mesma direção quanto a datação, assim como destaca a função editorial do cap. 28: O redator anônimo (IV-III séculos), que o inseriu em seu lugar atual, deu provas de gosto muito seguro. Sem dúvidas, ele quis concluir as discussões de Jó e de seus amigos relativizando ousadamente todo saber do homem e todas as suas afirmações: o homem não conhece o caminho da Sabedoria, e esta não se encontra na terra dos vivos (vv.12s, 20s). O poema de Jó 28 lança, assim, uma ponte entre os diálogos (4-27) e a segunda metade do poema, em que Jó, depois de haver protestado e confirmado sua inocência e depois de ter feito a Deus seu último desafio (29-31), verá, por sua vez, contestados seu poder e sua sabedoria3. 436 Andersen (1984), sem identificar o período histórico, e com ampla discussão sobre a crítica do desenvolvimento histórico do livro, reconhece que Jó 28 não se encaixa bem na continuidade da fala do personagem Jó: O capítulo 28 é melhor explicado como um tipo de coda entre o diálogo principal e a palavra final de Jó, mas não como parte do próprio pensamento de Jó a esta altura. É um comentário do autor, e quem fala é a pessoa contando a história, não uma das personagens da história (ANDERSEN, 1984, p.51). 1. INTRODUÇÃO Desse modo, tem-se certa convergência por parte dos autores entre o período de redação e a função editorial dos textos de Pv 1-9 e Jó 28, de modo que podem ser compreendidos como reflexão teológica ampla sobre o papel da sabedoria na realidade humana em sua função estrutural-literária nos respectivos livros. Passamos, então, à compreensão dos poemas nos quais a (in-)disponibilidade da sabedoria é tematizada. 4 Quanto ao uso plural ḥokmôt como expressão da personificação da sabedoria, Fox (2008, p.96) diz que a forma plural é usada com verbo no singular refere-se a um indivíduo. Kidner (1984, p.58, nota 3) indica ser uma forma fenícia de singular ou plural hebraico. No último caso, seria para expressar intensidade e plenitude, sendo regido pelo verbo no singular. Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 2. A DISPONIBILIDADE DA SABEDORIA: PV 1,20-33; 8,1-36 E 9,1-6 Os poemas de Pv 1,20-33; 8,1-36 e Pv 9,1-6 apresentam-se como textos nos quais a sabedoria é personificada e clama por ser ouvida, e são intercalados aos poemas de instrução iniciados com “Filho meu” encontrados em 1,8-19; 2,1-22; 3,1-12; 3,21-35; 4,1-9; 4,10-19; 4,20-27; 5,1-23; 6,20-35; 7,1-27. O texto de Pv 1,20-33 apresenta-se como um breve poema no qual a sabedoria4 grita a fim de que os ingênuos e insensatos ouçam sua voz. A estrutura do poema e sua linguagem apresentam-se com características de discurso profético de denúncia (v.24-25) e ameaça (v.26-31), recoberto pela linguagem sapiencial (v.23.32-33, cf. ainda LORENZIN, 2020, p.41; PINTO, 2018, p.38-39; MURPHY, 2002, p.10). Tem-se uma divisão em 4 partes do poema, a saber: 1,20-21 um convite no qual a sabedoria apresenta-se personificada e clama por ser ouvida; 1,22-25 questionamento, convite e acusação aos homens; 1,26-31 juízo e 1,32-33 exortação final. 437 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 A publicidade e disponibilidade da sabedoria encontra-se na introdução ao poema, nos v.20-23, em terceira pessoa, que introduz o discurso da sabedoria. Nele podemos encontrar ações que a sabedoria personificada faz: grita (tārōnnâ – v.20), dá sua voz (tittēn qôlāh - v.20), clama (hōmiyyôt – v.21), chama (tiqrā’ - v.21, também no v.24, em primeira pessoa), diz (to’mēr - v.21). Sua publicidade ainda é reforçada pelos locais nos quais a sabedoria personificada faz suas ações, todos eles locais públicos e visíveis: na rua (ḥûṣ - v.20), praça (rəḥōbôt - v.20), alto (rō’š -v.21), entrada das portas (pitḥê šəā’rîm – v.20), na cidade (‘îr - v.21 – v.20). A publicidade da sabedoria pode ser encontrada também no poema de 8,1-36, que pode ser estruturado da seguinte forma: 8,1-3 sabedoria fala em público; 8,4-11 admoestação pública; 8,12-21 autoapresentação da sabedoria; 8,22-31 a sabedoria e a criação; 8,32-36 exortação final em forma de bem-aventurança. Uma vez mais a atenção voltada para a introdução do poema, em terceira pessoa, leva-nos à dimensão pública e disponível da sabedoria. Nos v.1-3 tem-se repetidas as ações que apontam para sua publicidade: chamar (tiqrā’ – v.1), dar sua voz (tittēn qôlāh - v.1), gritar (tārōnnâ - v.3), reforçada ainda pelo início do discurso em primeira pessoa, com o uso do verbo chamar (’eqrā’ - v.4). 2. A DISPONIBILIDADE DA SABEDORIA: PV 1,20-33; 8,1-36 E 9,1-6 Neste segundo poema destacam-se ainda mais os locais públicos a partir dos quais a sabedoria personificada faz ouvir sua voz: no topo das alturas (rô’š mərômîm – v.2), ao lado do caminho (‘ălê-dārek -v.2), nas encruzilhadas, veredas (nətîbôt - v.2), junto às portas (ləyad-šə‘ārîm - v.3), na abertura da cidade (pî-qāret – v.3), no acesso das entradas (məbô’ pətāḥîm - v.3). Esse poema, especificamente, ainda desenvolve o tema da disponibilidade da sabedoria ao apresentá-la como presente na ordem criacional (v.32-36). Nessa parte, não apenas ela é apresentada como antecedente aos demais atos criadores (v.22- 26a), como também presente no restante da criação (v.26b-31), destacando-se dua presença entre os “filhos de adam” (v.31b). Por último, o breve poema de 9,1-6, no qual a sabedoria personificada mostra-se como pública e disponível, estrutura-se em duas partes: 9,1-3 apresenta os atos de 438 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 preparação do banquete da sabedoria; 9,4-6 convite ao banquete. Também neste a publicidade da sabedoria pode ser notada no início do poema, especificamente no v. 3: “envia suas servas a chamar (tiqrā’), sobre os topos dos altos (‘al-gappê-mərōmê) da cidade (qāret)”. Olhando-se os três poemas em conjunto, observa-se um vocabulário comum para a publicidade e disponibilidade da sabedoria, que pode ser classificado em: 1. Verbos e expressões que apontam para a ação oral de chamar e seus cognatos: grita (tārōnnâ, usado predominantemente com significado de “gritar de júbilo”, cf. Lv 9,24; Dt 32,43; Jó 38,7; Sl 20,6, Is 54,1); dá sua voz (tittēn qôlāh, cf. Gn 45,2; Pv 2,3); clama (hōmiyyôt, cf. Sl 46,6; 55,18; Is 17,12); chama (tiqrā’, cf. Pv 2,3; Jó 14,15; Is 55,5; 58,9; Lm 2,22). 1. Verbos e expressões que apontam para a ação oral de chamar e seus cognatos: grita (tārōnnâ, usado predominantemente com significado de “gritar de júbilo”, cf. Lv 9,24; Dt 32,43; Jó 38,7; Sl 20,6, Is 54,1); dá sua voz (tittēn qôlāh, cf. Gn 45,2; Pv 2,3); clama (hōmiyyôt, cf. Sl 46,6; 55,18; Is 17,12); chama (tiqrā’, cf. Pv 2,3; Jó 14,15; Is 55,5; 58,9; Lm 2,22). 2. Substantivos que apontam para lugares públicos e de grande circulação: rua (ḥûṣ, cf. Jó 18,7; Pv 7,12; Ct 8,1; Is 42,2; Na 2,5); praça (rəḥōbôt, cf. Pv 7,12; Lm 2,11-12; Na 2,5; Zc 8,4); cidade (‘îr, cf. Gn 34,28, em oposição ao campo); entrada das portas (pitḥê šəā’rîm, cf. 5 Nota-se o uso tanto de ‘îr como qeret como referência à cidade. O substantivo qeret é usado apenas em textos poéticos sapienciais: Pv 8,3; 9,3.14; 11,11; Jó 29,7. 6 Tradução nossa: “the heart of commerce judicial activity and social exchange ” Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 439 5 Nota-se o uso tanto de ‘îr como qeret como referência à cidade. O substantivo qeret é usado apenas em textos poéticos sapienciais: Pv 8,3; 9,3.14; 11,11; Jó 29,7. 6 Tradução nossa: “the heart of commerce, judicial activity, and social exchange.” 439 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 p p , ; , ; , ; , 6 Tradução nossa: “the heart of commerce, judicial activity, and social exchange.” p p g , , , p 8 De fato, a mudança para o discurso da sabedoria em Pv 9,1-6 só se explicita no v.5, no qual o uso da primeira pessoa torna-se claro. O v.4a pode ser a convocação das servas da sabedoria, e o v.4b, finalmente, a introdução ao discurso direto, que finalmente se inicia no v.5. , ç , q 9 Ao comentar Pv 1,20-33, Waltke (2011, p.272) propõe que a passagem do discurso direto nos ditos “filho meu” para o discurso indireto nos poemas da sabedoria personificada apontariam para um afastamento no tempo do julgamento, no qual “a distância tomará o lugar da intimidade; a separação substituirá o envolvimento”. Porém, deve-se levar em conta que o discurso indireto serve para introduzir o discurso da sabedoria personificada, diferenciando-a da voz daquele que ensina ao “filho meu”. 2. A DISPONIBILIDADE DA SABEDORIA: PV 1,20-33; 8,1-36 E 9,1-6 Is 26,2; Ez 21,27); ao lado do caminho (‘ălê-dārek, cf. Gn 49,17), nas encruzilhadas, veredas (nətîbôt, cf. Jó 38,20; Is 58,12; Jr 18,15), junto às portas (ləyad-šə‘ārîm, cf. expressão aproximada “junto ao caminho” - Sl 140,6), na abertura da cidade (pî-qāret), no acesso das entradas (məbô’ pətāḥîm, cf. expressão aproximada “acesso da cidade” Jz 1,24-25). Conforme Pinto (2018, p.38.93), alguns desses são lugares para tomar decisões, administrar a justiça e governar a cidade (Jr 17,19-20; Rt 4,10-11; Jó 29,7-25). Descrevem também uma realidade urbana de grande circulação5. Em concordância, Murphy (2002, p.49) indica a “abertura ou portões da cidade” como “o coração do comércio, da atividade jurídica e do intercâmbio social”6. Waltke (2011, p.274) aponta para as “praças públicas” como lugar de transações comerciais e reuniões públicas, em contraste com as ruas estreitas e Fox (2008, p.266) declara que os portões da cidade, são o centro da vida social, comercia e legal. 439 3. Substantivo que apontam lugares destacados e visíveis devido à altitude: alto (rô’š, cf. Ex 24,17; Nm 20,28; Jz 9,7); no topo das alturas (rô’š mərômîm, cf. Is 33,16); sobre os topos dos altos (‘al-gappê-mərōmê, cf. 2Rs 19,23, par. Is 37,24) da cidade (qāret). 3. Substantivo que apontam lugares destacados e visíveis devido à altitude: alto (rô’š, cf. Ex 24,17; Nm 20,28; Jz 9,7); no topo das alturas (rô’š mərômîm, cf. Is 33,16); sobre os topos dos altos (‘al-gappê-mərōmê, cf. 2Rs 19,23, par. Is 37,24) da cidade (qāret). 3. Substantivo que apontam lugares destacados e visíveis devido à altitude: alto (rô’š, cf. Ex 24,17; Nm 20,28; Jz 9,7); no topo das alturas (rô’š mərômîm, cf. Is 33,16); sobre os topos dos altos (‘al-gappê-mərōmê, cf. 2Rs 19,23, par. Is 37,24) da cidade (qāret). Tendo observado essas características textuais, pode-se compreender ao menos quatro estratégias literárias que os poemas usam para indicar a publicidade e disponibilidade da sabedoria: o uso da personificação, a indicação de ações de chamar, a localização pública da sabedoria e o discurso direto no qual a própria sabedoria toma a palavra. Sobre a personificação, ainda que possa ser visto não apenas como estratégia literária, mas também como uma carga teológica, o uso teológico não elimina sua função poética. Sobre isso, Líndez afirma (2011, p.55): Parece que a pura abstração poética é muito pouco, pois o recurso à personificação da sabedoria não é um mero jogo da fantasia do artista, cujo conteúdo permanece em sua imaginação. 7 O v.22 dirige-se à segunda pessoal do plural, os néscios, corroborando-se como fala da própria sabedoria na primeira pessoa do singular, no v.23. Cf. PINTO, 2018, p.38 7 O v.22 dirige-se à segunda pessoal do plural, os néscios, corroborando-se como fala da própria sabedoria na primeira pessoa do singular, no v.23. Cf. PINTO, 2018, p.38 8 De fato, a mudança para o discurso da sabedoria em Pv 9,1-6 só se explicita no v.5, no qual o uso da primeira pessoa torna-se claro. O v.4a pode ser a convocação das servas da sabedoria, e o v.4b, finalmente, a introdução ao discurso direto, que finalmente se inicia no v.5. 9 Ao comentar Pv 1,20-33, Waltke (2011, p.272) propõe que a passagem do discurso direto nos ditos “filho meu” para o discurso indireto nos poemas da sabedoria personificada apontariam para um afastamento no tempo do julgamento, no qual “a distância tomará o lugar da intimidade; a separação substituirá o envolvimento”. Porém, deve-se levar em conta que o discurso indireto serve para introduzir o discurso da sabedoria personificada, diferenciando-a da voz daquele que ensina ao “filho meu”. 2. A DISPONIBILIDADE DA SABEDORIA: PV 1,20-33; 8,1-36 E 9,1-6 A subsistência própria, mesmo que dependente de Deus, vai longe demais, pois na realidade não se tem em vista uma pessoa. Devemos, pois, entender por personificação da sabedoria um termo médio entre a pura fantasia poética e a realidade de uma verdadeira pessoa. As estratégias da indicação das ações e a localização pública da sabedoria ficam claras pelo vocabulário utilizado, tanto os verbos e locuções verbais, quanto nos substantivos e expressões nominais, acima explorados. Por último, note-se o uso do discurso direto. O discurso direto pode ser visto pela mudança de pessoas, passando da terceira pessoa, nas introduções aos poemas (1,20-21; 8,1-3; 9,1-3) para a primeira pessoa, na qual a própria sabedoria passa a falar (1,22-23s7; 8,4s; 9,4s8). O uso do discurso direto nesses poemas reforça a proximidade da sabedoria, que, no conjunto de Pv 1-9, soma-se ao uso da expressão “filho meu” nos poemas de “instrução”, na qual o mestre/pai conduz o aprendiz/filho nos caminhos da sabedoria9. 440 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 y p p 11 A tradução grega corrigiu “preço” (‘rk) por “caminho” (drk). Sobre isso, conferir BIBLIA DE JERUSALÉM, 2006, p.836, nota. d.; BHS, 2009, p.1255, aparato crítico. Ainda que a correção faça sentido, o texto hebraico parece seguir a lógica dos versos seguintes (v.14-19), cf. REYBURN, 1992, p.508; ANDERSEN, 1984, p.226. Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 10 Tradução nossa. Original: “En Pv 8 el punto de partida es el opuesto, la sabiduría toma la i 10 Tradução nossa. Original: “En Pv 8 el punto de partida es el opuesto, la sabiduría toma la iniciativa de buscar al hombre, y por ello el hombre la puede encontrar” 10 Tradução nossa. Original: En Pv 8 el punto de partida es el opuesto, la sabiduría toma la i de buscar al hombre, y por ello el hombre la puede encontrar” 10 Tradução nossa. Original: “En Pv 8 el punto de partida es el opuesto, la sabiduría toma la iniciativa de buscar al hombre, y por ello el hombre la puede encontrar” 11 A tradução grega corrigiu “preço” (‘rk) por “caminho” (drk). Sobre isso, conferir BIBLIA DE JERUSALÉM, 2006, p.836, nota. d.; BHS, 2009, p.1255, aparato crítico. Ainda que a correção faça sentido, o texto hebraico parece seguir a lógica dos versos seguintes (v.14-19), cf. REYBURN, 1992, p.508; ANDERSEN, 1984, p.226. 3. A INDISPONIBILIDADE DA SABEDORIA: JÓ 28,1-28 Conforme já visto, é possível que o poema de Jó 28 seja contemporâneo aos textos de Pv 1-9. Chama a atenção que também tematize o tema da disponibilidade da sabedoria. Porém, em sentido inverso a Provérbios, em Jó a sabedoria é descrita como oculta e indisponível. A estrutura do poema baseia-se na repetição da pergunta nos v. 12 e 20 (LEVEQUE, 1987; HEINEM, 1982, p.91; ANDERSEN, 1984, p.222-223): “E a sabedoria onde é encontrada? / e onde está o lugar do discernimento?”. Essa pergunta divide o poema em 3 partes: 28,1-11 a capacidade humana consegue explorar a pedra preciosa mais oculta, porém... 28,12-19 a capacidade humana desconhece a sabedoria; 28,20-28 Deus conhece a sabedoria e, para o ser humano, sabedoria é temor a Deus. A descrição explicita da ocultação e indisponibilidade da sabedoria concentra-se na segunda e terceira parte do poema, especificamente após a pergunta dos v.12 e 20, apresentando-se como resposta imediata. A própria pergunta apresenta-se como estratégia literária que tematiza a indisponibilidade da sabedoria: pergunta-se por aquilo que não está disponível. Interessante notar que, enquanto em Pv 1-9 a própria sabedoria clama, grita e chama, em Jó 28 é o narrador que pergunta pela sabedoria. Comparando Jó 28 com Pv 8, Schökel (1971, p.132) propõe que “em Pv 8 o ponto de partida é o oposto, a sabedoria toma a inciativa de buscar o ser humano e por ele o ser humano a pode encontrar”10 Segue-se à pergunta as respostas. No v. 13, explicita-se o duplo desconhecimento e ocultação da sabedoria em relação ao ser humano: este desconhece seu preço (‘erkāh11) e não está disponível na terra dos viventes (’ereṣ haḥayyîm). O tema do “preço da sabedoria” desenvolve-se nos v.15-19, enquanto sua indisponibilidade geográfica no v.14. Após a segunda pergunta, no v.20, a resposta do v. 21 reforça a 441 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 indisponibilidade da sabedoria às pessoas: encoberta (ne‘elmâ) aos olhos de todo vivente (kol-ḥāy), escondida (nistārâ) às aves dos céus (‘ôp haššāmayim). A indisponibilidade da sabedoria pode ser notada também por meio da comparação entre as duas primeiras partes do poema (v.1-11+12-19). Entre essas, é comparada a ação de buscar minérios (v.3-5.7-9), com a busca pela sabedoria. Enquanto aqueles, com muito esforço, podem ser encontrados, esta é indisponível ao ser humano. ( , p ) g q p , primevos e elementares poderes da natureza. Tehom e Yam, o original caos aquático da criação.” 14 Martin-Achard (2015, p.55.60) levanta as seguintes referências nas quais as expressões são usadas associadas aos šə’ôl: para təhôm: Sl 42,7; 71,20; 77,17; Ex 15,5.8 etc; para yām: Sl 46,3; Lm 2,13; Jó 26,5ss; ’ăbaddôn Sl 88,12; Jó 26,6; 28,22; 31,12; Pv 15,11. Também G. Gerleman associa o šə’ôl ao ’ăbaddôn e māwet (JENNI; WESTERMANN, 1985, Tomo II, col.1056). Andersen (1984, p.226) comenta: “Se Abadom e a Morte são poderes elementares que correspondem aos caos, então não devemos achar no v. 22 uma contradição da idéia de que aquilo que não está disponível “na terra dos viventes” (v. 13, que é outro contrapeso ao v. 22) talvez seja encontrado no Seol, sem mencionar uma polêmica contra a necromancia.” 3. A INDISPONIBILIDADE DA SABEDORIA: JÓ 28,1-28 Além da comparação da ação de buscar, reforça-se ainda por meio da comparação do valor incalculável da sabedoria em relação aos minérios de alto valor (v.1-2.5.15-19). Assim, não apenas encontrar a sabedoria seria humanamente impossível, como também precificá-la. Interessante estratégia literária utilizada em Jó 28 é o uso da personificação, quando comparada a Pv 1-9. Em Jó 28 a sabedoria não é personificada, antes, os lugares nos quais ela poderia estar, mas não se encontra, são (cf. CLINES,2006, p. 917). A personificação pode ser notada pelos discursos diretos. São personificados o abismo (təhôm - v.14), o mar (yām - v.14), a perdição (’ăbaddôn - v.22) e a morte (māwet12 - v.22). São lugares ocultos ao ser humano13. Essas referências transcendem a “localização geográfica”, sendo associados ao “mundo dos mortos” (šə’ôl – MARTIN- ACHARD, 2015, p.55.60) e possuindo um caráter simbólico-religioso14. Terrien (1994, p.219-220) explica: No ambiente cultural do Oriente clássico, essas alusões não podem ser tomadas como simples adornos poéticos. O poeta faz alusões às práticas rituais dos egípcios, dos fenícios e dos mesopotâmicos. Eis o homem à procura do segredo de seu destino. Ele oferece o preço mais exorbitante para adquiri-lo. [...] O Abismo e a Morte, símbolos cósmicos, são tão estranhos à sabedoria como a tecnologia do homem e o rito dos sacerdotes. No começo da terceira estrofe ele põe em cena 442 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 as divindades da morte, assim como no começo da segunda estrofe, por um instante, fez falar os deuses do abismo aquoso... as divindades da morte, assim como no começo da segunda estrofe, por um instante, fez falar os deuses do abismo aquoso... Nessas referências geográfico-simbólicas, em comparação com os poemas de Pv 1- 9, os lugares na qual a sabedoria está ou é procurada apresentam-se em contraste flagrante: de lugares públicos, de grande circulação, conhecidos e visíveis, a lugares ocultos, difíceis e simbólico-religiosos nos quais, nem lá, a sabedoria se encontra. A última estratégia literária para compreender a descrição da indisponibilidade da sabedoria, em comparação a Pv 1-9, especificamente Pv 8,1-36, é a tematização da sabedoria em relação à criação. 15 Sobre essa relação entre Jó 28 e o derradeiro discurso divino, Terrien (1994, p.29) escreve: “Entretanto, a sua linguagem e o seu estilo apresentam afinidades estreitas com os discursos de Iahweh [...] Ele parece ser, pois, uma composição do próprio poeta, colocada por um membro da escola jobiana em sua posição atual, na forma de ‘gradual’ ou de preparação para a teofania (caps. 38ss).” 3. A INDISPONIBILIDADE DA SABEDORIA: JÓ 28,1-28 Enquanto em Pv 8 o tema apresenta-se em função da disponibilidade e explicitação da sabedoria na criação, seja como sua antecessora, seja como presente na mesma, em Jó o próprio Deus é quem perscruta a sabedoria na criação (28,23-27), tema que é desenvolvido nos discursos finais de Deus em Jó 38-4115. Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 4. DISPONIBILIDADE E INDISPONIBILIDADE: A FUNÇÃO LITERÁRIA DA SABEDORIA EM PROVÉRBIOS E JÓ E A DIALÉTICA DA VIDA HUMANA Após notar as aproximações temporais e temáticas de Pv 1-9 e Jó 28, perguntamos sobre as diferentes e antagônicas caracterizações da sabedoria em ambos e sua relação com a composição com os respectivos livros. Já foi observado que tais textos fazem parte da composição final dos livros. Portanto, partimos de sua forma final para compreender sua função estrutural e, ao comparar ambas, deduzir do conjunto do corpus literário sapiencial as percepções literário-teológicos sobre a sabedoria. Já se destacou a função de Pv 1-9 como introdução ao livro de Provérbios. A personificação da sabedoria em discurso direto, assim como a retórica instrucional presente nos poemas de Pv 1-9 servem para introduzir as coletâneas de sabedoria popular presentes em Pv 10,1-31,9. 443 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 Exatamente na compreensão desse nexo encontra-se a resposta literário-teológica para a disponibilidade e explicitação da sabedoria em Pv 1-9. Isso porque em Pv 10- 31,9 tem-se a coletânea de ditos e sentenças proverbiais populares, fruto da experiência cotidiana e comum, da qual deriva a sabedoria experiencial. Assim, por meio da compreensão dessa sabedoria prática, empírica, observável, deduz-se a explicitação da sabedoria. Uma vez que a sabedoria pode ser observada na experiência comum do ser humano, é como se ela clamasse constantemente aos seres humanos para ser observada e seguida. A explicitação do conjunto de comportamento humano, conforme encontrado nos provérbios práticos de Pv 10-31,9, equivale a publicidade da sabedoria personificada, que grita constantemente em lugares públicos. Já em Jó 28 o contexto literário muda, alterando-se, também, a percepção quanto à sabedoria. Jó 28 serve como invólucro dos diálogos entre Jó e seus amigos sobre a natureza e razão do sofrimento de Jó. Sobre isso, Leveque (1987, p.51) esclarece: O papel de Jó 28 não se reduz, pois, ao de simples interlúdio. Ao inserir o poema neste lugar, o autor quis, certamente, concluir o diálogo de Jó com seus amigos; mas fê-lo propondo, por sua vez, uma tese radical, que rejeita definitivamente a teologia demasiado curta na visão dos visitantes. Esta crítica das pretensões da sabedoria atinge, aliás, o próprio Jó, que, por seu turno, tentou abrir para si caminho a nível das expressões humanas. (LEVEQUE, 1987, p.51) O papel de Jó 28 não se reduz, pois, ao de simples interlúdio. Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 4. DISPONIBILIDADE E INDISPONIBILIDADE: A FUNÇÃO LITERÁRIA DA SABEDORIA EM PROVÉRBIOS E JÓ E A DIALÉTICA DA VIDA HUMANA Ao inserir o poema neste lugar, o autor quis, certamente, concluir o diálogo de Jó com seus amigos; mas fê-lo propondo, por sua vez, uma tese radical, que rejeita definitivamente a teologia demasiado curta na visão dos visitantes. Esta crítica das pretensões da sabedoria atinge, aliás, o próprio Jó, que, por seu turno, tentou abrir para si caminho a nível das expressões humanas. (LEVEQUE, 1987, p.51) Assim, a indisponibilidade da sabedoria em Jó 28 serve como certa resposta editorial ao debate entre Jó e seus amigos sobre a suposta culpa de Jó por seu sofrimento (Jó 3-27). Não apenas isso, como também introduz os discursos finais de Deus (38-41). Enquanto resposta ao debate, Jó 28 tira da “terra dos viventes” a possibilidade de encontrar a razão total pelos dissabores da vida humana. Nesta perspectiva, Schwienhorst-Schönberger (2011, p.142) escreve: “Quem pode penetrar filosoficamente de tal maneira na horrível situação de Jó, a ponto de mostrar um caminho de compreensão e interpretação?”. Essa perspectiva condiz com o prólogo narrativo do livro (Jó 1-2), no qual encontra-se a razão do sofrimento nas cenas celestiais (1,6-12; 2,1-6) e não nas terrenas (1,1-5.13-20; 2,7-13). Desse modo, a indisponibilidade da sabedoria em Jó 28 vincula-se à tentativa humana frustrada de dar resposta final ao sofrimento humano (Jó 3-27), diante do mistério da vida (Jó 38- 41). 41). 444 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 Desse modo, a dialética da disponibilidade e indisponibilidade da sabedoria presente nos usos literários em Provérbios e Jó vincula-se às diferentes formas de tematizar a sabedoria na experiência humana. Se por um lado o livro de Provérbios apresenta uma sabedoria prática, empírica, funcional e essa pode ser apreendida pela observação, pelo cotidiano, pelo conselho dos mais velhos, sendo sabedoria disponível, por outro, em Jó, a sabedoria enquanto tentativa de domesticação da realidade, de explicação dos mistérios da vida, apresenta-se como indisponível e oculta ao ser humano, cabendo a este, no mistério de sua existência, abrir mão das tentativas de elaborar sistemas totais de explicação de sua condição finita. Na trilha dessas sapiencialidades encontra-se o temor a Deus (yir’at ’ădōnāy/YHWH Pv 1,29; 8,13; Jó 28,2816), o respeito pelo mistério do dom da existência, que nos conduz a observar e cuidar da vida em sabedoria prática, sem, contudo, arrogar-se ao domínio da mesma, em sabedoria total17. 17 Sobre o temor a Deus como sabedora acessível ao homem e a sabedoria inacessível enquanto controle da vida, cf. HEINEM, 1982, p.93 16 Conferir ainda Pv 1,7; 9,10; 10,27; 14,27; 15,33; 19,23; 22,4; 31,30. Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 5. CONSIDERAÇÕES FINAIS O desenvolvimento literário-teológico da sabedoria na Bíblia Hebraica passa pela compreensão não apenas das diferentes formas literárias que a sapiência bíblica possui, ou da variedade de fenômenos que investiga, desde a arte da prática e do bem viver a partir das observações cotidianas até as especulações sobre o sentido da vida e do sofrimento, como também pela forma com que tematiza a própria sabedoria, como uma espécie de “metadiscurso”. Nos textos de Pv 1,20-33; 8,1-36 e 9,1-6 e Jó 28,1-28, esse metadiscurso em forma poética apresenta a sabedoria como disponível (Provérbios) e indisponível (Jó) ao ser humano. Essa dialética da sabedoria dá-se pela função que esses textos desempenham nos livros, enquanto pertencentes à edição final destes. No conjunto da sabedoria bíblica, essa dialética também aponta para a experiência humana que ora conhece e desenvolve sabedoria prática, ora não pode lidar com os mistérios da vida. 445 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 REFERÊNCIAS ANDERSEN, Francis I. Jó: introdução e comentário. São Paulo: Vida Nova, 1984. (Série Cultura Bíblica) BHS. Bíblia Hebraica Stuttgartensia. Stuttgart: Barueri: Deutsch Bibelgesellschaft: Sociedade Bíblica do Brasil, 2009. BIBLEWORKS. BibleWorks 10. 2015. [software] BÍBLIA DE JERUSALÉM. Bíblia de Jerusalém: nova edição revista e ampliada. São Paulo: Paulus, 2006. CLINES, D. J. A. Word Biblical Commentary: Job 21-37. Nashville: Thomas Nelson Publishers, 2006. FOX, M. V. Proverbs 1-9: A new translation with introduction and commentary. New Haven; London: Yale University Press, 2008. (Anchor Yale Bible Commentary) HEINEM, Karl. O Deus indisponível: o livro de Jó. São Paulo: Paulinas, 1982. JENNI, Ernst; WESTERMANN, Claus. Diccionario Teologico Manual del Antiguo Testamento. Madrid: Cristiandad, 1985. [Tomo II] KIDNER, Derek. Provérbios: introdução e comentário. São Paulo: Vida Nova, 1980. (Série Cultura Bíblica) EVEQUE, Jean. Jó: o livro e a mensagem. São Paulo: Paulinas, 1987. (Cadernos Bíblico LÍNDEZ, José Vílchez. Sabedoria e sábios em Israel. São Paulo: Loyola, 2011. (Coleção bíblica Loyola). LORENZIN, Tiziano. Livros Sapienciais e Poéticos. Petrópolis: Vozes, 2020. (Introdução aos Estudos Bíblicos). MARTIN-ACHARD, Robert. Da morte à ressurreição segundo o Antigo Testamento. Santo André, SP: Academia Cristã, 2015. MURPHY, R. E. Word Biblical Commentary: Proverbs. Dallas: Word Inc., 2002. ORTIZ, Pedro. Dicionário de hebraico e aramaico bíblicos. São Paulo: Loyola, 2010. PINTO, Sebastiano (org). Provérbios: introdução, tradução, comentário. São Paulo: Loyola, 2018. (Coleção bíblica Loyola). POPE, M. H. Job: Introduction, translation, and notes. New Haven; London: Yale University Press, 2008. (Anchor Yale Bible Commentary) REYBURN, William D. A handbook on the book of Job. New York: United Bible Society, 1992. (UBS Handbook Series). ROSSI, Luiz Alexandre Solano; NASCIMENTO, Lucas Merlo (orgs). O livro de Jó: leituras e reflexões. São Paulo: Recriar, 2021. SCHMID, Konrad. História da Literatura do Antigo Testamento: uma introdução. São Paulo: Loyola, 2013. (Coleção bíblica Loyola). 446 446 Paralellus, Recife, v. 12, n. 30, mai./ago. 2021, p. 433-447 SCHWIENHORST-SCHÖNBERGER, Ludger. Um caminho através do sofrimento: o livro de Jó. São Paulo: Paulinas, 2011. WALTKE, Bruce K. Provérbios: volume 1, capítulos 1 a 15. São Paulo: Cultura Cristã, 2011. (Comentários do Antigo Testamento) ERRIEN, Samuel. Jó. São Paulo: Paulus, 1993. (Grande Comentário Bíblico) SCHWIENHORST-SCHÖNBERGER, Ludger. Um caminho através do sofrimento: o livro de Jó. São Paulo: Paulinas, 2011. TERRIEN, Samuel. Jó. São Paulo: Paulus, 1993. (Grande Comentário Bíblico) WALTKE, Bruce K. Provérbios: volume 1, capítulos 1 a 15. São Paulo: Cultura Cristã, 2011. (Comentários do Antigo Testamento) SCHWIENHORST-SCHÖNBERGER, Ludger. Um caminho através do sofrimento: o livro de Jó. São Paulo: Paulinas, 2011. TERRIEN, Samuel. Jó. São Paulo: Paulus, 1993. (Grande Comentário Bíblico) 447
https://openalex.org/W2018419967	https://www.pure.ed.ac.uk/ws/files/7743202/Fatigue_after_stroke_baseline_predictors_and_influence_on_survival._Analysis_of_data_from_UK_patients_recruited_in_the_International_Stroke_Trial.pdf	English	null	Fatigue after Stroke: Baseline Predictors and Influence on Survival. Analysis of Data from UK Patients Recruited in the International Stroke Trial	PloS one	2,011	cc-by	5,639	Digital Object Identifier (DOI): 10.1371/journal.pone.0016988 Link: Link to publication record in Edinburgh Research Explorer Document Version: Publisher's PDF, also known as Version of record General rights C i h f h General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Edinburgh Research Explorer Fatigue after Stroke: Baseline Predictors and Influence on Survival. Analysis of Data from UK Patients Recruited in the International Stroke Trial Citation for published version: Mead, GE, Graham, C, Dorman, P, Bruins, SK, Lewis, SC, Dennis, MS & Sandercock, PAG 2011, 'Fatigue after Stroke: Baseline Predictors and Influence on Survival. Analysis of Data from UK Patients Recruited in the International Stroke Trial', PLoS ONE, vol. 6, no. 3, e16988. https://doi.org/10.1371/journal.pone.0016988 Citation for published version: Mead, GE, Graham, C, Dorman, P, Bruins, SK, Lewis, SC, Dennis, MS & Sandercock, PAG 2011, 'Fatigue after Stroke: Baseline Predictors and Influence on Survival. Analysis of Data from UK Patients Recruited in the International Stroke Trial', PLoS ONE, vol. 6, no. 3, e16988. https://doi.org/10.1371/journal.pone.0016988 Published In: PLoS ONE Publisher Rights Statement: Publisher Rights Statement: Publisher Rights Statement: Copyright: © 2011 Mead et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Take down policy Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact openaccess@ed.ac.uk providing details, and we will remove access to the work immediately and investigate your claim. Download date: 24. Oct. 2024 Fatigue after Stroke: Baseline Predictors and Influence on Survival. Analysis of Data from UK Patients Recruited in the International Stroke Trial Abstract Funding: This analysis was funded by Chief Scientist Office Scottish Government CZG/2 analysis, decision to publish or preparation of the manuscript. Funding: This analysis was funded by Chief Scientist Office Scottish Government CZG/2/387. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. Competing Interests: The authors have declared that no competing interests exist. * E-mail: Gillian.e.mead@ed.ac.uk Gillian E. Mead1, Catriona Graham2, Paul Dorman3, Karsten Bruins Slot4, Steff C. Lewis1, Martin S. Dennis1, Peter A. G. Sandercock1, on behalf of the UK Collaborators of IST Gillian E. Mead1, Catriona Graham2, Paul Dorman3, Karsten Bruins Slot4, Steff C. S. Dennis1, Peter A. G. Sandercock1, on behalf of the UK Collaborators of IST 1 University of Edinburgh, Edinburgh, United Kingdom, 2 Wellcome Trust Clinical Research Facility, University of Edinburgh, Edinburgh, United Kingdom, 3 Department of Neurology, Newcastle Acute Hospitals NHS Foundation Trust, Newcastle-upon-Tyne, United Kingdom, 4 Division of Medicine, Oslo University Hospital-Ulleva˚l, Oslo, Norway Abstract Background and Purpose: Little is known about the associations of post-stroke fatigue or its influence on survival. The vitality component of the Short Form 36 (SF-36) is a valid and reliable measure of post-stroke fatigue. We sought to identify associates of post-stroke fatigue and determine whether fatigue predicted survival. Methods: We used SF-36 vitality scores obtained by postal questionnaires from 1080 UK patients randomised in the International Stroke Trial, at a mean of 64 weeks after stroke onset. We used logistic regression to explore factors at randomisation which predicted SF-36 vitality at follow-up, and the relationship between SF-36 vitality and both SF-36 mental health and SF-36 emotional role function at follow-up. We used Cox proportional hazards to explore the influence of SF-36 vitality at follow-up on subsequent survival, using four different statistical models for handling missing data. Results: Female sex, increasing age, lower mental health and lower emotional role function scores were associated with greater degrees of fatigue after stroke (i.e. lower vitality scores) but these factors explained ,30% of the variance (R2) in fatigue. In two models, fatigue at follow-up was associated with shorter subsequent survival. Conclusion: Increasing age, female sex, emotional role function and mental health were associated with increased fatigue at a mean of 64 weeks after stroke onset, but explained less than 30% of the variance. Fatigue was associated with reduced subsequent long-term survival in 2/4 models. Further work is needed to identify the biological substrate of fatigue and to clarify its influence on survival. Citation: Mead GE, Graham C, Dorman P, Bruins Slot K, Lewis SC, et al. (2011) Fatigue after Stroke: Baseline Predictors and Influence on Survival. Analysis of Data from UK Patients Recruited in the International Stroke Trial. PLoS ONE 6(3): e16988. doi:10.1371/journal.pone.0016988 Editor: German Malaga, Universidad Peruana Cayetano Heredia, Peru Received September 22, 2010; Accepted January 18, 2011; Published March 18, 2011 Received September 22, 2010; Accepted January 18, 2011; Published March 18, 2011 ad et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits tion, and reproduction in any medium, provided the original author and source are credited. Copyright: 2011 Mead et al. This is an open-access article distributed under the terms of the Creative Commons Attributi unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. PLoS ONE \| www.plosone.org Fatigue after Stroke: Baseline Predictors and Influence on Survival. Analysis of Data from UK Patients Recruited in the International Stroke Trial Gillian E. Mead1, Catriona Graham2, Paul Dorman3, Karsten Bruins Slot4, Steff C. Lewis1, Martin S. Dennis1, Peter A. G. Sandercock1, on behalf of the UK Collaborators of IST Collection of survival data Where the initial model generated contained variables which did not reach statistical significance at the 5% level, we removed the least significant variable and generated a new model. We repeated this process until the only variables remaining in the model were those which reached statistical significance. We calculated four different models to explore robustness to differing assumptions about missing variables. All UK patients in IST were ‘flagged’ at the Office for National Statistics central registry of deaths, enabling us to collect data on all deaths occurring in this cohort up to November 2000 giving us survival data for between 5 and 7 years after the assessment of SF- 36 [18]. Collection of SF-36 data A subset of the patients recruited by the UK centres in the IST trial participated in a study of quality of life after stroke. Surviving patients (n = 2253) were sent either a SF-36 or EuroQUOL (another scale to measure quality of life) by postal questionnaire [14-16] at a mean of 64 (SD 30) weeks following randomisation. Between 2nd March 1993 and 31st May 1995, 1400 patients were sent a SF-36, of whom 1080 responded (77%) [14-16]. The SF-36 includes the following domains: vitality, physical functioning, physical role function, social functioning, bodily pain, mental health, emotional role functioning and general health [13-16]. Each domain, including the mental health and emotional role function domains, includes several questions. We changed the question ‘Did you feel full of pep?’ to ‘Did you feel full of life?’ to ensure its cultural relevance [13-15]. The responses to the questions allow a total score for each domain to be calculated using the standardised SF-36 scoring manual [13]. Scores for each domain range from 0-100, with higher scores indicating better health. We produced multiple regression models which included the variables at randomisation shown to have an effect on vitality score (i.e. the final models generated in the previous section) and the SF- 36 mental health and SF-36 emotional role scores. Missing scores for emotional role function and mental health were treated in the same way as missing scores for vitality. The International Stroke Trial The IST was a randomised controlled trial of the effects of aspirin, subcutaneous heparin, both or neither, among 19,435 patients with acute ischaemic stroke, recruited within 48 hours of stroke onset [17]. To receive the patient’s ran- domised treatment allocation, clinicians telephoned a central randomisation service. During this telephone call, after data on age, gender, pathological subtype of stroke, stroke subtype, neurological deficits, blood pressure and cardiac rhythm had been recorded and checked, the system generated the treatment allocation [17]. We performed four sensitivity analyses to determine the robustness of the estimates to different assumptions about missing data, in which we: i) removed those with missing vitality data ii) imputed the minimum value recorded iii) imputed the maximum value recorded iv) imputed the mean value observed. Post-Stroke Fatigue in International Stroke Trial Post-Stroke Fatigue in International Stroke Trial Our aims were to determine: a) what factors at stroke onset predicted SF-36 vitality scores [13] measured a few months later, b) the relationship between SF-36 vitality measured in survivors a few months after stroke and mood (SF-36 mental health domain and emotional role function domain) and c) influence of SF-36 vitality in survivors on subsequent survival. anterior circulation syndrome (PACS), lacunar syndrome (LACS), posterior circulation syndrome (POCS) or ‘other’), presence of a visible infarct on brain imaging, cardiac rhythm (atrial fibrillation or sinus rhythm) and systolic blood pressure (,140, 140-159, 160- 79, .180mmHg. Results were reported in the form of regression equations. For a simple linear regression, with one predictor variable, the equation would be of the form Y = a+bX i.e. the Y variable can be expressed in terms of a constant (a) and a slope (b) multiplied by the X variable. For multiple linear regression analysis when there is more than one independent variable (as in this study), the regression line cannot be visualized in the two dimensional space, but can be computed using a linear equation containing all those variables. Multiple regression procedures estimate a linear equation of the form: Y = a+b1* X1+b2X2+…+bpXp. Statistical analysis of factors at randomisation associated with fatigue on follow-up Using the threshold value of at least ‘10 patients per variable’ rule for the analyses of regression, we had sufficient power for our analyses [19]. We analysed fatigue severity (SF-36 vitality) as a continuous variable. We performed multiple linear regression analyses to determine the relationship between fatigue and the following variables measured at randomisation: age, gender, pathological subtype of stroke (ischaemic, haemorrhagic or indeterminate), stroke subtype (total anterior circulation syndrome (TACS), partial Statistical analysis of association between mood and fatigue At follow-up, patients completed the emotional role score and the mental health scores of the SF-36. The responses were used to calculate a total score for each domain [13], ranging from 0-100, with higher score indicating better health. Materials and Methods We used data from patients in the International Stroke Trial (IST), recruited by UK Centres, who had participated in a substudy on quality of life after stroke [14-17]. Statistical analysis of the influence of fatigue at follow-up on subsequent survival We generated a Cox proportional hazards model containing the following variables: age, sex, pathological subtype, stroke subtype, visible infarct, atrial fibrillation (at randomisation), systolic blood pressure (at randomisation) and all component parts of the SF-36 at follow-up (vitality, role-emotional, mental health, physical function, bodily pain, general health and social function) with the exception of role-physical. We did not include role-physical as this measures similar constructs to physical function and there were more missing values for role physical than for physical function. For these analyses, survival was measured from the date on which SF-36 vitality was measured. These data allowed us to explore the relationship between SF- 36 vitality a few months after stroke, patient characteristics at randomisation and the SF-36 role emotional and SF-36 mental health a few month after stroke. Introduction All previous studies of fatigue after stroke used generic fatigue scales which had not been tested for validity or reliability in stroke. The vitality domain of the Short-Form 36 (SF-36) has face validity and is reliable after stroke [12], and correlates with other fatigue scales [11]. Stroke patients fulfilling a case definition for clinically significant fatigue have lower SF-36 vitality scores than those without clinically significant fatigue [12]. Fatigue is common after stroke, with estimates of prevalence ranging from 16% [1] to 70% [2]. Fatigue is one of the most distressing symptoms after stroke [3], yet its aetiology remains uncertain. In the general non-stroke population, fatigue is often a symptom of depression. Several cross-sectional studies [1,3-8] and one longitudinal study [9] have sought associations of fatigue after stroke using different generic fatigue scales. Most [1,3,4,6,9] but not all [7,8] have found associations with depression. One previous study found an association with variability in blood pressure [10]. Early studies of fatigue after stroke suggested that it is more common in brain stem lesions [7], but other studies have not shown such an association [1,4,6,8,9,11,12]. The single previous study that investigated associations between fatigue and survival found that it was independent predictor of shorter survival [1]. The vitality domain in the original version of the SF-36 includes four questions: In the past four weeks: ‘Did you feel full of pep?’, ‘Did you have a lot of energy?’, ‘Did you feel worn out?’ and ‘Did you feel tired?’ Responses to each question are: ‘All of the time’, ‘Most of the time’, ‘a Good bit of the time’, ‘Some of the time’ ‘A little of the time’ and ‘None of the time’ [13]. After the responses to the individual questions were recorded, a total score for vitality can be calculated according to the SF-36 scoring manual [13]. The range of possible scores is from 0 to 100, with higher scores indicating higher vitality (i.e. less fatigue). PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e16988 March 2011 \| Volume 6 \| Issue 3 \| e16988 1 Results The mean age of the participants was 71.1 (SD 10.8) years and 602 (55.7%) were men. Other demographic details are PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e16988 March 2011 \| Volume 6 \| Issue 3 \| e16988 2 Post-Stroke Fatigue in International Stroke Trial distribution of the vitality scores for men and women is shown in figure 1. Table 1. Baseline characteristics of the 1080 participants. distribution of the vitality scores for men and women is shown in figure 1. Characteristic Number % Sex Male 602 55.7 Stroke type Ischaemic 918 85.0 Haemorrhagic 82 7.6 Indeterminate 53 4.9 Not stroke 27 2.5 Stroke sub-type LACS 285 26.4 PACS 446 41.3 POCS 124 11.5 TACS 222 20.6 Other 3 0.28 Infarct visible Yes 265 24.5 Atrial fibrillation Yes 166 15.4 Systolic blood pressure ,140 200 18.5 140 - 159 306 28.3 160 - 179 273 25.3 . = 179 301 27.9 doi:10.1371/journal.pone.0016988.t001 Association between mood and fatigue We determined association between mood and fatigue, using SF- 36 mental health scores, SF-36 emotional role scores and vitality scores. There were missing data from 6.7% of mental health scores and 18.3% of emotional role scores. Figure 2 shows the univariate relationship between SF-36 mental health and vitality (Pearson correlation coefficient 0.20, p,0.001, n = 1004). Figure 3 shows the univariate relationship between SF-36 role emotional and vitality (Pearson correlation coefficient 0.38, p,0.001, n = 863). To our knowledge, this is the first study to explore not only the factors at stroke onset that predict post-stroke fatigue, but also the relationship between fatigue and mood, and the influence of fatigue at follow-up on subsequent survival. Increasing age, worse emotional role function and worse mental health score measured at follow-up were significantly associated with fatigue. Fatigue on follow-up was associated with significantly reduced subsequent survival in two models. Although we did not specifically aim to compare SF-36 vitality scores after stroke with a general population, we noted that the median SF-36 vitality score of 37.5 was substantially lower than the mean SF-36 vitality score of 55.8 from the general population aged 70-74 years [20]. For the multiple logistical regression models, gender, age, mental health score and emotional role function were significant predictors of vitality in all four models. The amount of variability in the vitality score explained by variation in the models is small [adjusted R2 for the 4 models are: 0.18, 0.29, 0.27, 0.16 respectively]. However, these models explain much more of the variance than those models which did not include mental health and emotional role scores. The study had several strengths: it was a very large, prospective and conducted in a single, well characterised cohort of stroke patients recruited at a uniform time point (stroke onset); it used a valid and reliable measure of fatigue after stroke (SF-36 vitality score) [11,12]. It explored baseline predictors of fatigue, relationship between fatigue and mood, and the influence of fatigue on survival in the same large group of patients. The regression equations for all 4 models are shown below. a) Missing values excluded Increasing age, lower emotional role function and lower mental health score were significantly associated with fatigue in all four models and female sex was also significant in three models. Influence of fatigue on subsequent survival Survival data to November 2000 were available for 1072/1080 (99.3%) participants, of whom 420/1072 (39.2%) had died. The median number of days’ follow-up from randomisation to November 2000 was 5.9 years (IQR 3.8 to 6.6). The median length of follow-up from when SF-36 was measured, in the 1016 participants with this information recorded, was 5.2 years (IQR 2.4 to 5.3). SF-36 vitality~62:74{0:22 (age)z4:29(if male){ 11:89(if ischaemic){9:79(if haemorrhagic){ 9:15(if indeterminate)z1:13(if POCS){ 5:17(if TACS){0:93(if LACS){ 10:78(if other stroke subtype) SF-36 vitality~62:74{0:22 (age)z4:29(if male){ 11:89(if ischaemic){9:79(if haemorrhagic){ 9:15(if indeterminate)z1:13(if POCS){ 5:17(if TACS){0:93(if LACS){ 10:78(if other stroke subtype) Table 2 presents multivariate analyses showing independent predictors of survival. Increasing age, atrial fibrillation (at randomisation) and ‘other’ stroke subtype predicted shorter survival in all four models. In two models, higher SF-36 vitality at follow-up predicted longer subsequent survival. Note: ‘Other’ stroke syndrome is defined as those individuals whose clinical syndromes could not be assigned to one of the four OCSP syndromes. Baseline predictors of fatigue We analysed SF-36 data from 1080 patients, of whom 1006 (93%) had a vitality score recorded. The amount of variability in the vitality score explained by the variables we included was very small [adjusted R2 values for the 4 models were: 0.037, 0.054, 0.004, 0.034 respectively]. Three of the four final multiple regression models contained the same variables although the parameter estimates varied slightly. However, when the missing values were set to the maximum recorded value, the model was very different, with only gender predicting fatigue. These equations allow us to estimate SF-36 vitality score if we have data for age, gender, pathological type (ischaemic, haemorrhagic or indeterminate), and subtype (TACS, PACS, LACS, POCS or ‘other’ stroke subtype). Each of models is specified in full below, according to how missing data were dealt with: a) Missing values excluded a) Missing values excluded SF-36 vitality~64:21{0:22 (age)z4:71(if male){ 12:48(if ischaemic){10:09(if haemorrhagic){ 9:53(if indeterminate)z1:31(if POCS){ 5:65(if TACS){0:97(if LACS){ 11:13(if other stroke subtype) SF-36 vitality~64:21{0:22 (age)z4:71(if male){ 12:48(if ischaemic){10:09(if haemorrhagic){ 9:53(if indeterminate)z1:31(if POCS){ 5:65(if TACS){0:97(if LACS){ 11:13(if other stroke subtype) shown in table 1. Median SF-36 vitality score (inter- quartile range) was 37.5 (20.0, 55) for the entire group, 40 (IQR 25, 55) for men and 35 (20,50) for women. The shown in table 1. Median SF-36 vitality score (inter- quartile range) was 37.5 (20.0, 55) for the entire group, 40 (IQR 25, 55) for men and 35 (20,50) for women. The Figure 1. Distribution of SF-36 vitality scores, according to gender. Higher vitality scores indicate less fatigue. doi:10.1371/journal.pone.0016988.g001 Figure 1. Distribution of SF-36 vitality scores, according to gender. Higher vitality scores indicate less fatigue. doi:10.1371/journal.pone.0016988.g001 Figure 1. Distribution of SF-36 vitality scores, according to gender. Higher vitality scores indicate less fatigue. doi:10.1371/journal.pone.0016988.g001 March 2011 \| Volume 6 \| Issue 3 \| e16988 PLoS ONE \| www.plosone.org 3 c) Missing values set to maximum b) Missing values set to minimum c) Missing values set to maximum b) Missing values set to minimum SF-36 vitality~69:65{0:32 (age)z4:79(if male){ 13:76(if ischaemic){10:59(if haemorrhagic){ 8:97(if indeterminate)z1:16(if POCS){ 6:44(if TACS)z0:73(if LACS){ 9:34(if other stroke subtype) c) Missing values set to maximum SF-36 vitality~40:23z3:65(if male) d) Missing values imputed at mean d) Missing values imputed at mean d) Missing values imputed at mean d) Missing values imputed at mean SF-36 vitality~30:52-0:21 (age)z3:14(if male)z SF-36 vitality~30:52-0:21 (age)z3:14(if male)z 0:24 (mental health score)z0:18 (emotional role score) SF-36 vitality~40:23z3:65(if male) (‘*’ means multiply) Association between mood and fatigue Importantly, the factors we identified accounted for only a small amount of variance in fatigue, suggesting that other factors, for which we did not collect data, must also be associated with fatigue. SF-3SF-36 vitality~32:24{0:23 agez3:79(if male)z SF-3SF-36 vitality~32:24{0:23 agez3:79(if male)z 0:24 (mental health score)z0:18 (emotional role score) Post-Stroke Fatigue in International Stroke Trial Post-Stroke Fatigue in International Stroke Trial fatigue. In the general population, SF-36 vitality scores are lower Table 2. Multivariate analyses showing independent predictors of survival, according to the method of dealing with missing SF-36 data. Parameter Hazard ratio Lower 95% CI Upper 95% CI Missing data set to mean Age 1.060 1.048 1.072 LACS 0.78 0.585 1.039 ‘Other’ stroke type 4.435 1.083 18.169 PACS 0.875 0.676 1.132 POCS 0.68 0.466 0.992 Atrial fibrillation (at randomisation) 0.708 0.555 0.904 SF-36 vitality 0.993 0.987 0.999 SF-36 social role 0.996 0.992 1.000 Missing data set to maximum Age 1.061 1.049 1.073 LACS 0.778 0.584 1.036 ‘Other’ stroke type 4.334 1.057 17.767 PACS 0.856 0.662 1.107 POCS 0.669 0.458 0.975 Atrial fibrillation (at randomisation) 0.694 0.544 0.885 SF-36 social role 0.994 0.991 0.997 SF-36 vitality 1.001 0.996 1.005 Missing data set to minimum Age 1.058 1.046 1.070 LACS 0.784 0.589 1.044 ‘Other’ stroke type 4.662 1.139 19.084 PACS 0.873 0.675 1.129 POCS 0.677 0.464 0.988 Atrial fibrillation (at randomisation) 0.710 0.556 0.905 SF-36 vitality 0.988 0.984 0.993 Missing data excluded Age 1.065 1.052 1.078 Atrial fibrillation (at randomisation) 0.689 0.530 0.897 SF-36 general health 0.989 0.984 0.994 SF-36 vitality 0.994 0.988 1.001 LACS = lacunar syndrome, PACS = partial anterior circulation syndrome, TACS = total anterior circulation syndrome, POCS = posterior circulation syndrome. Note that hazard ratios and 95% CI for SF-36 vitality are provided for all four models, even though the hazard ratio reached statistical significance in only two of the models. doi:10.1371/journal.pone.0016988.t002 pre-stroke fatigue [4], suggesting that some fatigue was present before the stroke and that some develops after the stroke. After correction for role emotional and mental health, we found no evidence of a relationship between OCSP stroke subtype, as an indicator of stroke lesion location and size, and fatigue. Ours is the largest study to date to explore associations between fatigue and stroke lesion location and size, and so this is an important negative finding. Two previous studies have suggested that fatigue is more common in brain stem strokes [5,7], although other studies did not find an association between brain stem strokes and fatigue. Although fatigue may not be related to lesion location, it is, nevertheless, possible that fatigue may have a ‘central’ origin i.e. it might be a direct consequence of a cerebral infarct or cerebral haemorrhage [21]. Post-Stroke Fatigue in International Stroke Trial The observation that fatigue after stroke is unlike anything ever experienced before by stroke survivors [22] supports the idea that fatigue might be directly caused by the brain lesion. One study has suggested that fatigue may be related to dysregulation of blood pressure after stroke, perhaps as a result of antihypertensive drugs [10]. We did not find an association between baseline blood pressure and fatigue, but were unable to explore relationship between fatigue and current blood pressure. Intriguingly, lower SF-36 vitality scores (i.e. more fatigue) in our study were associated with shorter survival. However, after correction for pain, social role functioning, physical function and general health, fatigue remained a significant predictor in only two models. Nonetheless, if confirmed, this would strengthen the case for further studies into the biology of post-stroke fatigue. Our study has some weaknesses. Our data were from a randomised trial rather than from a population-based study, although the inclusion criteria for IST were broad. The SF-36 scores were obtained at a median of 64 weeks after stroke onset which perhaps somewhat limits the generalisability of the findings. The 77% response rate leads to some degree of uncertainty in the analyses, though it is similar to the previous postal survey of fatigue after stroke [1]. There were missing data items in some of the SF- 36 forms. We could have simplified data analysis by performing only one analysis after excluding those with missing data, but this might have biased our results, because non-responders may have been systematically different from those who did respond e.g. more depressed, more fatigued. There is no consensus about how to deal with missing data in this type of study. Each possible method has its own potential biases, so we dealt with these missing data items in four different ways statistically. Reassuringly, final regression models for SF-36 vitality were similar, and importantly, all explained only a small amount of variance in fatigue. We were unable to correct our analyses for levels of dependence, which is known to influence survival [18], but we did correct for SF-36 self- reported physical function. There were a few patients who did not have a definite diagnosis of stroke; we decided to include these as the physician recruiting them to IST felt that they had had a stroke. PLoS ONE \| www.plosone.org b) Missing values set to minimum b) Missing values set to minimum Previous cross-sectional studies have explored the relationship between fatigue and mood disorders after stroke, with most but not all finding associations. Thus, the evidence points to an association between fatigue and mood disorders. However, we cannot determine direction of causality from current evidence i.e. whether fatigue causes mood disorders or whether mood disorders cause SF-36 vitality~30:95{0:25 agez3:34(if male)- SF-36 vitality~30:95{0:25 agez3:34(if male)- 7:58(if ischaemic)-5:44(if haemorrhagic)- 7:58(if ischaemic)-5:44(if haemorrhagic)- 2:60(if indeterminate)z0:40 (mental health score)z 0:17 (emotional role score) 0:17 (emotional role score) PLoS ONE \| www.plosone.org PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e16988 4 Post-Stroke Fatigue in International Stroke Trial Figure 2. Relationship between SF-36 mental health and vitality. Higher scores indicate better health status. As the number of data points are limited, a small random component (‘jitter’) has been added to each variable to better demonstrate the number of patients at a single co- ordinate. (Pearson correlation coefficient 0.20, p,0.001, n = 1004). doi:10.1371/journal.pone.0016988.g002 Figure 2. Relationship between SF-36 mental health and vitality. Higher scores indicate better health status. As the number of data points are limited, a small random component (‘jitter’) has been added to each variable to better demonstrate the number of patients at a single co- ordinate. (Pearson correlation coefficient 0.20, p,0.001, n = 1004). doi:10.1371/journal.pone.0016988.g002 Figure 3. Relationship between SF-36 role emotional and vitality. Higher scores indicate better health status. As the number of data points are limited, a small random component (‘jitter’) has been added to each variable to better demonstrate the number of patients at a single co- ordinate. (Pearson correlation coefficient 0.38, p,0.001, n = 863). doi:10.1371/journal.pone.0016988.g003 Figure 3. Relationship between SF-36 role emotional and vitality. Higher scores indicate better health status. As the number of data points are limited, a small random component (‘jitter’) has been added to each variable to better demonstrate the number of patients at a single co- ordinate. (Pearson correlation coefficient 0.38, p,0.001, n = 863). doi:10.1371/journal.pone.0016988.g003 March 2011 \| Volume 6 \| Issue 3 \| e16988 PLoS ONE \| www.plosone.org 5 March 2011 \| Volume 6 \| Issue 3 \| e16988 References 14. Dorman PJ, Slattery J, Farrell B, Dennis MS, Sandercock PAG, the United Kingdom Collaborators in the International Stroke Trial (1997) A randomised comparison of the Euroquol and Short Form-36 after stroke. BMJ 315: 461. 1. Glader E-L, Stegmayr B, Asplund K (2002) Post-stroke fatigue. A 2 year follow- up study of stroke patients in Sweden. Stroke 33: 1327–1333. p y p 2. Leegard OF (1983) Diffuse cerebral symtoms in convalescents from cerebral infarction and myocardial infarction. Acta Neurol Scand 67: 348–355. infarction and myocardial infarction. Acta Neurol Scand 67: 348 15. Dorman PJ, Dennis MS, Sandercock PAG for the United Kingdom Collaborators in the International Stroke Trial (1999) How do scores on the Euroquol Relate to Scores on the SF-36 after stroke? Stroke 30: 2146–2151. 3. van der Werf, van den Broek HLP, Anten HWM, Bleijenberg G (2001) Experience of Severe Fatigue long after stroke and its relation to depressive symptoms and disease characteristics. Eur Neurol 45: 28–33. 16. Dorman PJ, Slattery J, Farrell B, Dennis MS, Sandercock P (1998) Qualitative comparison of the reliability of Health Status assessments with the Euroquol and SF-36 questionnaires after stroke. Stroke 29: 63–68. 4. Choi-Kwan S, Han SW, Kwon SU, Kim JS (2005) Post-stroke fatigue: Characteristics and Related factors. Cerebrovasc Dis 19: 84–90. 5. Naess H, Nyland HI, Thomassen L, Aarseth J, Myhr K-M (2005) Fatigue at long-term follow-up in young adults with cerebral infarction. Cerebrovasc Dis 20: 245–250. 17. Anonymous (1997) The International Stroke Trial (IST): a randomised trial of aspirin, subcutaneous heparin, both, or neither among 19435 patients with acute ischaemic stroke. International Stroke Trial Collaborative Group. Lancet 349: 1569–81. 6. Morley W, Jackson K, Mead G (2005) Post-Stroke fatigue: an important yet neglected synptom. Age Ageing 34: 313. 18. Slot KB, Berge E, Dorman P, Lewis S, Dennis M, et al. (2008) , Oxfordshire Community Stroke Project, the International Stroke Trial (UK); Lothian Stroke Register (2008) Impact of functional outcome at six months on long term survival in patients with ischaemic stroke: prospective cohort studies. BMJ 336: 376–379. 7. Staub F, Bogousslavsky J (2001) Fatigue after stroke: a major but neglected issue. Cerebrovasc Dis 12: 75–81. 8. Ingles JL, Eskes GA, Phillips SJ (1999) Fatigue after stroke. Arch Phys Med Rehabil 80: 173–178. 9. Schepers VP, Wisser-Meily AM, Ketelaar A, Lindeman E (2006) Post-stroke fatigue: course and its relationship to personal and stroke-related factors. Post-Stroke Fatigue in International Stroke Trial Although fatigue is a complex experience for patients, in large studies such as this, it is necessary to use a quick, simple tool, such as the SF-36 vitality component, that has been validated against a case definition for clinically significant fatigue after stroke [11,12]. Despite these weaknesses, this very large study represents an important new contribution to the sparse literature on fatigue after stroke. fatigue. In the general population, SF-36 vitality scores are lower in women than men, and fall with increasing age [20], and fatigue is a well-recognised symptom of mood disorders. Thus, these findings are consistent with findings from the general population. What are the implications of this study? Physicians should consider mood disorders in patients with post-stroke fatigue. Further research is required to determine the relationship between pre-stroke and post-stroke fatigue, the natural history of fatigue after stroke, the direction of causality between mood disorders and fatigue and to determine whether fatigue is an independent The extent to which the post-stroke vitality scores might have been explained by pre-stroke scores is uncertain. One previous study found that 62% of patients with post-stroke fatigue had had PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e16988 6 Post-Stroke Fatigue in International Stroke Trial Post-Stroke Fatigue in International Stroke Trial Post-Stroke Fatigue in International Stroke Trial Post-Stroke Fatigue in International Stroke Trial predictor of survival and, if it is, the biological mechanism that accounts for the survival disadvantage that fatigue confers. There is clearly also a need to develop interventions for fatigue after stroke and test them in appropriate trials [23]. Author Contributions Conceived and designed the experiments: GEM KBS PD SCL MSD PAGS. Performed the experiments: KBS PD MSD PAGS. Analyzed the data: CG SCL. Contributed reagents/materials/analysis tools: SCL. Wrote the paper: GEM CG PD SCL MSD KBS PAGS. Conceived and designed the experiments: GEM KBS PD SCL MSD PAGS. Performed the experiments: KBS PD MSD PAGS. Analyzed the data: CG SCL. Contributed reagents/materials/analysis tools: SCL. Wrote the paper: GEM CG PD SCL MSD KBS PAGS. PLoS ONE \| www.plosone.org References Arch Phys Med Rehabil 87: 184–8. 19. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR (1996) A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 49: 1373–1379. 10. Harbison JA, Walsh S, Kenny RA (2009) Hypertension and daytime hypotension on ambulatory blood pressure is associated with fatigue following stroke and TIA. Q J Med 102: 109–115. 20. Walters SJ, Munro JF, Brazier JE (2001) Using the SF-36 with older adults: a cross-sectional community based survey. Age and Ageing 30: 337–343. stroke and TIA. Q J Med 102: 109–115. 21. Winward C, Sackley C, Metha S, Rothwell PM (2009) A Population-Based Study of the Prevalence of Fatigue After Transient Ischemic Attack and Minor Stroke. Stroke 40: 757–761. 11. Lynch J, Mead GE, Greig CA, Young A, Lewis S et al (2007) Development and evaluation of a case definition for fatigue after stroke. J Psychosom Res 63: 539–544. 22. Flinn NA, Stube JE (2009) Post-stroke fatigue: qualitative study of three focus groups. Occup Ther Int 17: 81–91. 12. Mead GE, Lynch J, Greig CA, Young A, Lewis S et al (2007) Evaluation of fatigue scales in stroke patients. Stroke 38: 2090–5. 23. McGeough E, Pollock A, Smith LN, Dennis M, Sharpe M, et al. (2009) Interventions for post-stroke fatigue. Cochrane Database of Systematic Reviews, Issue 3. Art. No.:CD007030.DOI: 10.1002/14651858.CD007030.pub2. 13. Ware JE, Snow KK, Kosinski M, Gandek B (1993) SF-36 Health Survey: Manual and Interpretation Guide. Boston Mass: Health Institute, New England Medical Centre. PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e16988 7 7
https://openalex.org/W1848738154	https://cdr.lib.unc.edu/downloads/t148fq21r	English	null	How do persistent infections with hepatitis C virus cause liver cancer?	Current opinion in virology	2,015	cc-by	8,258	Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Abstract Persistent infection with hepatitis C virus (HCV) is associated with an increased risk of hepatocellular carcinoma (HCC). Cancer typically develops in a setting of chronic hepatic inflammation and advanced fibrosis or cirrhosis, and such tissue represents a pre-neoplastic “cancer field”. However, not all persistent infections progress to HCC and a combination of viral and host immune factors likely to contribute to carcinogenesis. HCV may disrupt cellular pathways involved in detecting and responding to DNA damage, potentially adding to the risk of cancer. Efforts to unravel how HCV promotes HCC are hindered by lack of a robust small animal model, but a better understanding of molecular mechanisms could identify novel biomarkers for early detection and allow for development of improved therapies. Author Manuscript Author Manuscript Corresponding author: Stanley M. Lemon (smlemon@med.unc.edu), 8.034 Burnett-Womack CB #7292, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7292 USA, Tel: 1-919-843-1848; Fax: 1-919-843-7240. How do persistent infections with hepatitis C virus cause liver cancer? Jonathan Mitchell, Stanley M. Lemon, and David R. McGivern Jonathan Mitchell, Stanley M. Lemon, and David R. McGivern Division of Infectious Diseases, Department of Medicine, and Lineberger Comprehensive Cancer Center, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA. Author Manuscript Corresponding author: Stanley M. Lemon (smlemon@med.unc.edu), 8.034 Burnett-Womack CB #7292, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7292 USA, Tel: 1-919-843-1848; Fax: 1-919-843-7240. Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript The manuscript will undergo copyediting typesetting and review of Published in final edited form as: Published in final edited form as: Curr Opin Virol. 2015 October ; 14: 101–108. doi:10.1016/j.coviro.2015.09.003. HHS Public Access Author Manuscript INTRODUCTION However, the absence of a tractable small animal model of HCV infection and hepatocellular carcinogenesis has slowed progress in this field, such that many questions concerning how HCV causes liver cancer remain unresolved. Author Manuscript Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. INTRODUCTION Approximately 3.5 million persons in the USA are persistently infected with hepatitis C virus (HCV) [1]. Chronic hepatitis C is a slowly progressing disease and many infected persons remain asymptomatic for decades after initial infection. However, the long-term complications of infection are substantial, and include hepatic fibrosis, cirrhosis, and hepatocellular carcinoma (HCC). Many currently asymptomatic individuals acquired infection prior to the identification of HCV as the etiologic agent of chronic ‘non-A non-B’ hepatitis in 1989 [2]. Because the likelihood of disease increases with the length of infection, the incidence of HCV-associated cirrhosis and HCC has been increasing as this cohort ages and is predicted to reach a peak within the next decade [3]. In addition, a recent surge in injection drug use among young, largely white, non-urban Americans is leading to worrisome increases in the incidence of new HCV infections, and is likely predictive of future increases in liver cancer. Compounding the impact of such increases in HCC incidence, the five-year survival rate of patients with liver cancer has remained low (~15%) within the U.S. Author Manuscript Author Manuscript Mitchell et al. Page 2 As a single-stranded, positive-sense RNA virus (Figure 1), replicating primarily if not exclusively within hepatocytes, HCV is unique among cancer-causing viruses. Its intracellular replication cycle is entirely cytoplasmic and without potential for integration of viral genome into host chromosomal DNA. Similar to infection with other oncogenic viruses, only a minority of chronically infected persons develop cancer, suggesting that viral elements co-operate with host and environmental factors to promote tumorigenesis. Transgenic mice with liver-specific expression of HCV proteins are at increased risk for hepatocellular carcinoma [4,5]. However, the absence of a tractable small animal model of HCV infection and hepatocellular carcinogenesis has slowed progress in this field, such that many questions concerning how HCV causes liver cancer remain unresolved. As a single-stranded, positive-sense RNA virus (Figure 1), replicating primarily if not exclusively within hepatocytes, HCV is unique among cancer-causing viruses. Its intracellular replication cycle is entirely cytoplasmic and without potential for integration of viral genome into host chromosomal DNA. Similar to infection with other oncogenic viruses, only a minority of chronically infected persons develop cancer, suggesting that viral elements co-operate with host and environmental factors to promote tumorigenesis. Transgenic mice with liver-specific expression of HCV proteins are at increased risk for hepatocellular carcinoma [4,5]. Indirect effects of chronic inflammation: a pro-carcinogenic environment Within the infected liver, double-stranded viral RNA replication intermediates are sensed by host pathogen-associated molecular pattern (PAMP) receptors resulting in the activation of the transcription factors IRF3/7 and NF-κB, and the induction of interferons and related interferon-stimulated genes (ISGs). HCV has evolved several mechanisms that antagonize these responses (reviewed in [6]), and these may contribute to viral persistence by attenuating cell-intrinsic innate immune responses. Despite this virus antagonism, antiviral signaling pathways are activated during persistent HCV infection resulting in increased intrahepatic ISG transcription [7]. Subsequent adaptive immune responses, in particular virus-specific CD8+ cytotoxic T lymphocytes [8,9], while critical for resolving the infection, succeed in clearing the virus only in a minority (~30-40%) of cases [10]. Thus, most infected persons develop lifelong, persistent infection. Such individuals are at risk for protracted but ultimately ineffective immune responses leading to chronic immune-mediated inflammation with repeated cycles of hepatocyte destruction and regeneration. Author Manuscript What are the effects of chronic inflammation?—Persistent immune-mediated hepatic inflammation and associated fibrogenic wound-healing responses are likely to be important drivers of liver cancer in chronic hepatitis C (Figure 2). Activated inflammatory cells promote a pro-carcinogenic microenvironment by releasing reactive oxygen (ROS) and nitrogen (RNS) species and induce lipid peroxidation [11]. The expression of some HCV proteins, in particular core and NS5A, may also add directly to the induction of oxidative stress (see below). The HCV RNA replicase complex possesses a unique ‘sensor’ that responds to lipid peroxidation by down-regulating viral RNA synthesis, thereby maintaining replication at low levels and minimizing oxidative damage [12]. While this may facilitate HCV persistence, hepatic oxidative DNA damage is nonetheless common in chronic hepatitis C [13-15]. Such DNA damage is likely not heritable in terminally differentiated cells. However, in the context of inflammatory liver disease and hepatocellular necrosis associated with chronic hepatitis C, regenerative pathways may be activated allowing dedifferentiation and proliferation to replace damaged tissue. Author Manuscript Without treatment, approximately 20% of persons with chronic hepatitis C develop severe scarring of the liver, or cirrhosis. Most (perhaps 90%) HCV-associated cancers arise in a background of advanced fibrosis or cirrhosis. A progression of lesions (large regenerative nodules, low-grade and high-grade dysplastic nodules) have been identified in the cirrhotic Page 3 Page 3 Mitchell et al. liver that likely represent HCC precursors [16]. Indirect effects of chronic inflammation: a pro-carcinogenic environment Thus, during the development of hepatocellular cancer, the cirrhotic liver may be considered a pre-neoplastic “cancer field” comprised of genetically abnormal but non-neoplastic tissue that is at high risk for malignant transformation (Figure 2) [17]. Accordingly, HCC is often multifocal in origin in chronic hepatitis C [18]. Cells within this precancerous field already contain mutations that predispose to the cancerous phenotype (discussed below). Author Manuscript Au Author Manuscript What is the impact of chronic hepatocyte turnover?—Over decades of chronic HCV infection, apoptosis of infected hepatocytes (either immune- or virus-mediated) with compensatory hepatocellular proliferation may be strongly pro-carcinogenic. Tumor- promoting effects of apoptosis have been demonstrated in transgenic mouse models where liver-specific knockout of anti-apoptotic proteins such as Mcl-1 promote cancer [19]. Conversely, knockout of the proapoptotic p53 up-regulated modulator of apoptosis (PUMA, or Bbc3) inhibits tumor development following exposure to the chemical carcinogen, diethylnitrosamine [20]. Author Manuscript HCV-associated HCC may be considered, like other cancers, a disease of somatic evolution where individual cells are the unit of reproduction [21]. In chronic hepatitis C, the generation of ROS by either viral or immune-mediated mechanisms could create a pro- mutagenic environment, while chronic hepatocellular turnover could select for cells with genetic or epigenetic changes that confer growth advantages allowing clonal expansion (Figure 2). Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Direct effects of HCV on the infected hepatocyte Surprisingly, only a minority of hepatocytes contain detectable viral RNA or proteins within the HCV-infected liver [22,23], suggesting that most cells remain uninfected. The factors restricting replication are not well understood, and whether cancer arises from an infected hepatocyte or an uninfected bystander cell remains an open question. Author Manuscript Does HCV infection render a cell prone to accumulate genetic alterations?— Does HCV infection render a cell prone to accumulate genetic alterations?— HCCs exhibit a high degree of genetic heterogeneity indicative of a fundamental loss of genomic stability during hepatocellular carcinogenesis [24]. HCV infection is likely to contribute to this genetic instability directly by inducing the generation of reactive oxygen species (ROS) with consequent DNA damage (Figure 2). Although immune-mediated inflammation is at least partly responsible, as discussed above, expression of either the HCV core protein or NS5A, a component of the viral replicase, increases ROS levels and promotes oxidative stress in transgenic mice as well as cultured cells [25-29]. These effects have been attributed to mitochondrial and ER stress initiated by core and NS5A, respectively [25,27,28], although it is uncertain whether these proteins are expressed in sufficient abundance for such direct effects to occur within the infected liver. Author Manuscript In addition to promoting DNA damage, some evidence suggests HCV infection compromises the ability of host cells to detect and repair damaged DNA. The ataxia telangiectasia mutated (ATM) kinase plays a central role in initiating cellular responses to double-strand DNA breaks [30]. Overexpression studies suggest that core, NS3/4A (the major viral protease), and NS5B (RNA-dependent RNA polymerase) are capable of Page 4 Mitchell et al. Page 4 interacting with components of the ATM-driven response, thereby interfering with host DNA repair [31-33]. Although these interactions await confirmation in the context of viral replication, HCV infection impairs phosphorylation of an ATM substrate, histone 2A.X (H2AX), in Huh-7.5 cells [34]. Phosphorylated H2AX (γ-H2AX) normally acts as a platform for the recruitment of DNA repair factors to the sites of DNA damage [35]. As such, inhibition of γ-H2AX might disrupt DNA repair and contribute to genomic instability during HCV infection. Unfortunately, the small proportion of hepatocytes infected with HCV in vivo coupled with low-level expression of viral proteins makes it extraordinarily difficult to confirm or refute such effects in primary human tissues. Author Manuscript Author Manuscript The p53 protein is a critical tumor suppressor that coordinates cell-cycle arrest, senescence, and apoptosis in response to DNA damage and other stresses [36]. Mutations that disrupt p53 function are associated with the vast majority of human cancers, including HCC [37,38]. Numerous studies have indicated that HCV proteins, including core, NS3, and NS5A, can interact with p53 when overexpressed in cell culture (for an extensive review see [39]). Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Does HCV infection render a cell prone to accumulate genetic alterations?— However, these studies have yielded often-conflicting results as to the impact of HCV proteins on p53 activity, and interactions between HCV proteins and p53 have yet to be reported in the context of infection. The overall impact of HCV infection on p53 function remains undefined, largely because the cell-lines that are most permissive for HCV (Huh-7 hepatoma cells and their derivatives) express a mutated, inactive form of p53 [40,41]. Author Manuscript Author Manuscript A Does HCV infection promote hepatocellular proliferation?—A common feature of oncogenic viruses is their ability to increase cell proliferation, particularly through the inactivation of host tumor suppressors [42]. The retinoblastoma (Rb) protein restricts cell proliferation by repressing the activation of E2F transcription factors necessary for S-phase entry in the cell cycle [43]. HCV infection negatively regulates Rb abundance in cell culture [44-46]. This effect is mediated by the NS5B polymerase, which binds Rb via an LxCxE motif bearing sequence homology to the Rb-binding motifs of DNA virus oncoproteins [46]. NS5B mediates Rb relocalization, promotes proteasomal degradation of Rb in association with the host ubiquitin ligase E6AP, and activates E2F-responsive promoters [45,46]. Rb is essential for optimal innate immune responses in hepatocytes [47], explaining why an RNA virus such as HCV (that does not depend upon DNA synthesis for replication) would have evolved such a mechanism. Despite E2F activation, however, HCV infection reduces cell proliferation and triggers cell- cycle arrest in cultured cells [48,49]. These findings point towards opposing pro- and anti- proliferative effects during HCV infection. Consistent with this hypothesis, normally quiescent hepatocytes show increased entry into the cell cycle within HCV-infected livers, but markers of G1 arrest are also elevated [50]. However, it is unclear to what extent cell- cycle perturbations observed within the liver represent direct effects of HCV infection since only a minority of hepatocytes are typically infected in vivo [7,22]. It is possible that G1 arrest may be triggered by endogenous or therapeutically administered interferon [51]. Author Manuscript HCV may also deregulate the Wnt/β-catenin signaling pathway. This pathway is normally activated upon binding of Wnt ligands to the Frizzled receptor and ultimately triggers activation of the transcription factor β-catenin and consequent upregulation of cellular pro- Mitchell et al. Page 5 Page 5 survival genes. Mutations within this pathway, including activating mutations within β- catenin, occur frequently in HCC, indicating an important role for this pathway in hepatocellular carcinogenesis [37]. Does HCV infection render a cell prone to accumulate genetic alterations?— Both core and NS5A are capable of activating β-catenin when overexpressed in cell culture [52-55]. Likewise, β-catenin is activated in HCV transgenic mice, wherein it drives overexpression of the oncogene c-myc and thereby contributes to oxidative DNA damage and impaired cell-cycle arrest [26]. Author Manuscript Author Manuscript Do HCV-infected cells evade apoptosis?—The role of apoptosis in HCV-related HCC is likely complex. On one hand, apoptosis is expected to exert tumor suppressive effects through the removal of aberrant and infected cells. On the other hand, successive rounds of apoptotic cell death with subsequent regenerative proliferation could also contribute to carcinogenesis as discussed above, particularly in an environment of oxidative stress and genomic instability. Author Manuscript Apoptosis can function as a potent antiviral defense, and many viruses thus suppress this host response [56,57]. Whether HCV similarly restricts apoptosis remains unclear. Individual HCV proteins, including core, E2, NS2, NS3, and NS5A, inhibit apoptosis when overexpressed (reviewed in [39]), whereas HCV infection can induce apoptosis in Huh-7 cells and further sensitize them to death receptor-mediated apoptosis [48,49,58-60]. However, most studies of HCV-induced apoptosis have utilized an atypical, robustly replicating, infectious molecular clone of HCV, JFH1 or its derivatives. JFH1 replication is not suppressed by lipid peroxidation [12], and it may trigger apoptosis through artificially high-levels of replication and viral protein expression that do not reflect HCV infections in vivo [7,22]. An alternate molecular clone, H77S.3, replicates at lower levels in cell culture and elicits little apoptosis [48]. Virus produced in cell culture from a closely related clone, H77S.2, established persistent infection in a chimpanzee, whereas cell culture-derived JFH1 virus infects chimpanzees but does not establish persistence [61]. Moreover, passage of JFH1 in chimpanzees resulted in adaptive mutations that rendered it less pro-apoptotic than wild-type JFH1 [62]. Although only a small number of chimpanzees were involved in these studies, they suggest evasion of apoptosis may be important for establishing persistent infection. Author Manuscript Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. What genetic changes contribute to the “cancer field”? What can we learn from studying the genetics of HCC?—Although it is possible to identify driver genes in HCC by whole exome sequencing [37,63,64], no unique driver genes have yet been linked specifically to HCV infection [64]. A better understanding of genetic and epigenetic changes that drive HCC progression may ultimately allow for improved, personalized therapies, but the genetic diversity of HCC makes such studies very challenging [24]. Author Manuscript What are the earliest genetic changes in HCC?—The stepwise accumulation of mutations in HCC is poorly understood. Nonetheless, mutations identified in abnormal, pre- neoplastic tissue may be predictive of progression to HCC. Mutations in the telomerase reverse transcriptase (TERT) promoter are frequently present in high-grade dysplastic nodules (19%) and early HCC (61%) in cirrhosis of several different etiologies, including Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Page 6 Page 6 Mitchell et al. HCV infection [65,66]. Thus TERT promoter mutations represent one of the earliest recognized genetic changes in hepatocellular carcinogenesis, and could contribute to the cancer field effect in cirrhosis. Author Manuscript Author Manuscript What is the cellular origin of HCV-associated HCC?—The precise nature of the founder cell in HCC is also poorly understood. Proliferation is essential for a mutation arising in a pre-neoplastic cell to spread within a population of such cells. Both mature hepatocytes and hepatic progenitor cells are long-lived cells that are capable of repopulation following liver injury, and either could be the source of HCC founder cells in chronic HCV infection [67]. As discussed above, it is not clear whether HCV-associated HCC arises from infected or uninfected cells [68]. An interesting but unanswered question is whether HCV is able to replicate within abnormal hepatocytes in dysplastic nodules, and thus might act co-operatively with mutations that arise early in the progression to cancer. Expression of the liver-specific micro-RNA, miR-122, is often lost during progression to HCC, but is relatively conserved in HCV- associated versus hepatitis B virus (HBV)-associated cancer [69]. Since miR-122 is a critical pro-viral host factor for HCV, this suggests that HCV infection of founder cells may be important at some stage of this process. Author Manuscript Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. How will new HCV antivirals impact the incidence of HCC? How will new HCV antivirals impact the incidence of HCC? Over the last few years, the introduction of potent direct-acting antiviral (DAA) drugs targeting the NS3/4A protease, NS5A, or the NS5B polymerase has revolutionized the treatment of chronic hepatitis C. All oral combinations of these DAAs are well tolerated, and typically achieve a sustained virologic response (“SVR”) in 90-95% of patients with genotype 1 infections after only 12 weeks of therapy [78]. Patients with SVR are truly cured of the infection, as HCV does not archive its genome like HIV or HBV. These new drugs thus have the potential to substantially reduce HCC incidence over the coming decades. Several cohort studies demonstrate that SVR achieved with older interferon-based regimens reduces, but does not completely eliminate the risk of HCC in cirrhotic patients (reviewed in [79]). It is likely, but not yet proven, that successful DAA therapy will match these results. Nonetheless, HCC typically grows slowly and may not become clinically evident for many months [80]. Studies are needed to determine how long and to what extent the risk of HCC will persist in successfully treated patients. Until these questions are answered, screening for HCC will remain important for those patients achieving SVR. Author Manuscript What factors enhance the risk of HCC development in chronic hepatitis C? Host and environmental factors—While the progression of liver disease is highly variable amongst chronically-infected persons, 15-25% progress to cirrhosis by 25-30 years [70]. HCC occurs primarily in persons with cirrhosis, and approximately 20% of cirrhotic individuals ultimately progress to cancer [71]. Such progression is exacerbated by a number of risk factors, including alcohol consumption, older age, male gender, obesity, diabetes, and co-infection with human immunodeficiency virus (HIV) or HBV [72]. Author Manuscript Genetic risk factors for HCC development are less well defined. However, genome-wide association studies (GWAS) have revealed several single nucleotide polymorphisms (SNPs) associated with HCC development in HCV-infected persons [73,74]. A SNP located upstream of the MHC class I polypeptide-related sequence A (MICA) gene was associated with a slightly elevated risk for HCC [73]. This SNP appears to influence MICA expression levels, but it is not clear whether it increases the risk of cancer per se, or acts indirectly by influencing progression to cirrhosis. A second GWAS identified an intronic SNP within the DEP domain containing 5 (DEPDC5) locus that nearly doubled the risk of HCC [74]. Both GWAS studies were conducted within Japanese populations, raising the question of whether genetic determinants of HCC vary between ethnic groups. Indeed, an independent study of a Caucasian cohort revealed a reverse correlation between the MICA SNP and HCC risk, and also identified a SNP within the HLA complex P5 (HCP5) locus upstream of MICA associated with an increased risk of HCC in both Caucasian and Japanese populations [75]. The DEPDC5 SNP was not associated with cancer in the Caucasian cohort, further highlighting differences in genetic determinants of HCC across ethnicities. Author Manuscript Viral factors—There are seven different genotypes of HCV that vary by over 30% at the nucleotide level. Epidemiological studies suggest that HCC risk may differ amongst these Mitchell et al. Page 7 Page 7 genotypes, as genotypes 1b and 3 are associated more commonly with HCC than other genotypes [76,77]. In contrast, genotype 2 may be associated with a lower risk of HCC [77]. The mechanisms underlying these potential genotype-specific differences in HCC risk are not clear, nor is it proven that such differences do not arise from genetic or other differences in the infected host populations. Author Manuscript What does the future hold? Thus, the elimination of HCV- associated liver cancer seems a far-off and potentially elusive goal, and one that is unlikely to be achieved without development of an effective HCV vaccine. Author Manuscript Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. What does the future hold? A better understanding of the molecular drivers of HCV-associated HCC is likely to result in improved and increasingly personalized therapies for this cancer with attendant increases in five-year survival rates. Improved screening and early diagnostic modalities are also likely, along with reductions in the numbers of individuals reaching advanced stages of HCV- related liver disease due to continued improvements in antiviral therapy. Despite this, however, HCV-associated liver cancer will remain an important public health problem for the foreseeable future. There are several reasons for this. First, many infected persons with chronic hepatitis C have no access to new DAAs, either because of their high cost or lack of medical insurance. Second, because of coexisting conditions, the potential for drug-drug interactions, and/or HCV genotype-specific differences in antiviral response, a substantial proportion of patients remain difficult to treat with DAAs. Additionally, a large fraction of individuals are simply unaware of their infection until they develop symptoms of advanced liver disease [1]. And, not to be overlooked, persons fortunate enough to achieve SVR lack protective immunity and are at risk for re-infection. Thus, the elimination of HCV- associated liver cancer seems a far-off and potentially elusive goal, and one that is unlikely to be achieved without development of an effective HCV vaccine. A better understanding of the molecular drivers of HCV-associated HCC is likely to result in improved and increasingly personalized therapies for this cancer with attendant increases in five-year survival rates. Improved screening and early diagnostic modalities are also likely, along with reductions in the numbers of individuals reaching advanced stages of HCV- Author Manuscript related liver disease due to continued improvements in antiviral therapy. Despite this, however, HCV-associated liver cancer will remain an important public health problem for the foreseeable future. There are several reasons for this. First, many infected persons with chronic hepatitis C have no access to new DAAs, either because of their high cost or lack of medical insurance. Second, because of coexisting conditions, the potential for drug-drug interactions, and/or HCV genotype-specific differences in antiviral response, a substantial proportion of patients remain difficult to treat with DAAs. Additionally, a large fraction of individuals are simply unaware of their infection until they develop symptoms of advanced liver disease [1]. And, not to be overlooked, persons fortunate enough to achieve SVR lack protective immunity and are at risk for re-infection. REFERENCES Author Manuscript 1. Holmberg SD, Spradling PR, Moorman AC, Denniston MM. Hepatitis C in the United States. N Engl J Med. 2013; 368:1859–1861. [PubMed: 23675657] 2. Choo QL, Kuo G, Weiner AJ, Overby LR, Bradley DW, Houghton M. Isolation of a cDNA clone derived from a blood-borne non-A, non-B viral hepatitis genome. Science. 1989; 244:359–362. [PubMed: 2523562] 3. Davis GL, Alter MJ, El-Serag H, Poynard T, Jennings LW. Aging of hepatitis C virus (HCV)- infected persons in the United States: a multiple cohort model of HCV prevalence and disease progression. Gastroenterology. 2010; 138:513–521. 521, e511–516. [PubMed: 19861128] 4. Moriya K, Fujie H, Shintani Y, Yotsuyanagi H, Tsutsumi T, Ishibashi K, Matsuura Y, Kimura S, Miyamura T, Koike K. The core protein of hepatitis C virus induces hepatocellular carcinoma in transgenic mice. Nat.Med. 1998; 4:1065–1067. [PubMed: 9734402] [First report of HCC in transgenic mice expressing the HCV core protein.] 5. Lerat H, Honda M, Beard MR, Loesch K, Sun J, Yang Y, Okuda M, Gosert R, Xiao SY, Weinman SA, et al. Steatosis and liver cancer in transgenic mice expressing the structural and nonstructural proteins of hepatitis C virus. Gastroenterology. 2002; 122:352–365. [PubMed: 11832450] [Transgenic mice expressing the entire HCV polyprotein develop liver cancer in the absence of antiviral immune responses and inflammation.] Author Manuscript 6. Li K, Lemon SM. Innate immune responses in hepatitis C virus infection. Semin Immunopathol. 2013; 35:53–72. [PubMed: 22868377] 7. Wieland S, Makowska Z, Campana B, Calabrese D, Dill MT, Chung J, Chisari FV, Heim MH. Simultaneous detection of hepatitis C virus and interferon stimulated gene expression in infected human liver. Hepatology. 2014; 59:2121–2130. [PubMed: 24122862] 8. Cooper S, Erickson AL, Adams EJ, Kansopon J, Weiner AJ, Chien DY, Houghton M, Parham P, Walker CM. Analysis of a successful immune response against hepatitis C virus. Immunity. 1999; 10:439–449. [PubMed: 10229187] 9. Weiner A, Erickson AL, Kansopon J, Crawford K, Muchmore E, Hughes AL, Houghton M, Walker CM. Persistent hepatitis C virus infection in a chimpanzee is associated with emergence of a cytotoxic T lymphocyte escape variant. Proc Natl Acad Sci U S A. 1995; 92:2755–2759. [PubMed: 7708719] 10. Hajarizadeh B, Grebely J, Dore GJ. Epidemiology and natural history of HCV infection. Nat Rev Gastroenterol Hepatol. 2013; 10:553–562. [PubMed: 23817321] 11. Bartsch H, Nair J. Chronic inflammation and oxidative stress in the genesis and perpetuation of cancer: role of lipid peroxidation, DNA damage, and repair. Acknowledgements Supported in part by grants from the National Institutes of Health, F32-CA192633 (JKM), R21-AI115207 (DRM), and R01-CA164029 and R01-AI095690 (SML), and the University Cancer Research Fund of the University of North Carolina. Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Page 8 Page 8 Mitchell et al. Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. REFERENCES Kandathil AJ, Graw F, Quinn J, Hwang HS, Torbenson M, Perelson AS, Ray SC, Thomas DL, Ribeiro RM, Balagopal A. Use of laser capture microdissection to map hepatitis C virus-positive hepatocytes in human liver. Gastroenterology. 2013; 145:1404–1413. [PubMed: 23973767] Author Manuscript 24. Thorgeirsson SS, Grisham JW. Molecular pathogenesis of human hepatocellular carcinoma. Nat Genet. 2002; 31:339–346. [PubMed: 12149612] 25. Gong G, Waris G, Tanveer R, Siddiqui A. Human hepatitis C virus NS5A protein alters intracellular calcium levels, induces oxidative stress, and activates STAT-3 and NF-kappa B. Proc Natl Acad Sci U S A. 2001; 98:9599–9604. [PubMed: 11481452] 26. Higgs MR, Lerat H, Pawlotsky JM. Hepatitis C virus-induced activation of beta-catenin promotes c-Myc expression and a cascade of pro-carcinogenetic events. Oncogene. 2013; 32:4683–4693. [PubMed: 23108410] [Activation of β-catenin in transgenic mice expressing the entire HCV polyprotein suggests that this cellular signaling pathway may be directly modulated by viral proteins.] 27. Korenaga M, Wang T, Li Y, Showalter LA, Chan T, Sun J, Weinman SA. Hepatitis C virus core protein inhibits mitochondrial electron transport and increases reactive oxygen species (ROS) production. J Biol Chem. 2005; 280:37481–37488. [PubMed: 16150732] [Direct induction of ROS by HCV core protein contributes to oxidative stress and potentially DNA damage in the HCV-infected liver. (See also reference 28).] 28. Okuda M, Li K, Beard MR, Showalter LA, Scholle F, Lemon SM, Weinman SA. Mitochondrial injury, oxidative stress, and antioxidant gene expression are induced by hepatitis C virus core protein. Gastroenterology. 2002; 122:366–375. [PubMed: 11832451] 29. Wang AG, Lee DS, Moon HB, Kim JM, Cho KH, Choi SH, Ha HL, Han YH, Kim DG, Hwang SB, et al. Non-structural 5A protein of hepatitis C virus induces a range of liver pathology in transgenic mice. J Pathol. 2009; 219:253–262. [PubMed: 19621337] 30. Shiloh Y, Ziv Y. The ATM protein kinase: regulating the cellular response to genotoxic stress, and more. Nat Rev Mol Cell Biol. 2013; 14:197–210. 31. Ariumi Y, Kuroki M, Dansako H, Abe K, Ikeda M, Wakita T, Kato N. The DNA damage sensors ataxia-telangiectasia mutated kinase and checkpoint kinase 2 are required for hepatitis C virus RNA replication. J Virol. 2008; 82:9639–9646. [PubMed: 18667510] 32. Lai CK, Jeng KS, Machida K, Cheng YS, Lai MM. Hepatitis C virus NS3/4A protein interacts with ATM, impairs DNA repair and enhances sensitivity to ionizing radiation. Virology. 2008; 370:295–309. [PubMed: 17931678] Author Manuscript 33. REFERENCES Langenbecks Arch Surg. 2006; 391:499–510. [PubMed: 16909291] 12. Yamane D, McGivern DR, Wauthier E, Yi M, Madden VJ, Welsch C, Antes I, Wen Y, Chugh PE, McGee CE, et al. Regulation of the hepatitis C virus RNA replicase by endogenous lipid peroxidation. Nat Med. 2014; 20:927–935. [PubMed: 25064127] 13. Fujita N, Sugimoto R, Ma N, Tanaka H, Iwasa M, Kobayashi Y, Kawanishi S, Watanabe S, Kaito M, Takei Y. Comparison of hepatic oxidative DNA damage in patients with chronic hepatitis B and C. J Viral Hepat. 2008; 15:498–507. [PubMed: 18331251] 14. Shimoda R, Nagashima M, Sakamoto M, Yamaguchi N, Hirohashi S, Yokota J, Kasai H. Increased formation of oxidative DNA damage, 8-hydroxydeoxyguanosine, in human livers with chronic hepatitis. Cancer Res. 1994; 54:3171–3172. [PubMed: 8205535] Author Manuscript 15. Tanaka H, Fujita N, Sugimoto R, Urawa N, Horiike S, Kobayashi Y, Iwasa M, Ma N, Kawanishi S, Watanabe S, et al. Hepatic oxidative DNA damage is associated with increased risk for hepatocellular carcinoma in chronic hepatitis C. Br J Cancer. 2008; 98:580–586. [PubMed: 18231107] 16. Pathologic diagnosis of early hepatocellular carcinoma: a report of the international consensus group for hepatocellular neoplasia. Hepatology. 2009; 49:658–664. [PubMed: 19177576] 17. Braakhuis BJ, Tabor MP, Kummer JA, Leemans CR, Brakenhoff RH. A genetic explanation of Slaughter's concept of field cancerization: evidence and clinical implications. Cancer Res. 2003; 63:1727–1730. [PubMed: 12702551] Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Mitchell et al. Page 9 Page 9 18. Colombo M. Natural history and pathogenesis of hepatitis C virus related hepatocellular carcinoma. J Hepatol. 1999; 31(Suppl 1):25–30. [PubMed: 10622556] Author Manuscript 19. Weber A, Boger R, Vick B, Urbanik T, Haybaeck J, Zoller S, Teufel A, Krammer PH, Opferman JT, Galle PR, et al. Hepatocyte-specific deletion of the antiapoptotic protein myeloid cell leukemia-1 triggers proliferation and hepatocarcinogenesis in mice. Hepatology. 2010; 51:1226– 1236. [PubMed: 20099303] 20. Qiu W, Wang X, Leibowitz B, Yang W, Zhang L, Yu J. PUMA-mediated apoptosis drives chemical hepatocarcinogenesis in mice. Hepatology. 2011; 54:1249–1258. [PubMed: 21725994] 21. Salk JJ, Fox EJ, Loeb LA. Mutational heterogeneity in human cancers: origin and consequences. Annu Rev Pathol. 2010; 5:51–75. [PubMed: 19743960] 22. Liang Y, Shilagard T, Xiao SY, Snyder N, Lau D, Cicalese L, Weiss H, Vargas G, Lemon SM. Visualizing hepatitis C virus infections in human liver by two-photon microscopy. Gastroenterology. 2009; 137:1448–1458. [PubMed: 19632233] 23. REFERENCES Machida K, McNamara G, Cheng KT, Huang J, Wang CH, Comai L, Ou JH, Lai MM. Hepatitis C virus inhibits DNA damage repair through reactive oxygen and nitrogen species and by interfering with the ATM-NBS1/Mre11/Rad50 DNA repair pathway in monocytes and hepatocytes. J Immunol. 2010; 185:6985–6998. [PubMed: 20974981] 34. Duong FH, Christen V, Lin S, Heim MH. Hepatitis C virus-induced up-regulation of protein phosphatase 2A inhibits histone modification and DNA damage repair. Hepatology. 2010; 51:741– 751. [PubMed: 20043320] 35. van Attikum H, Gasser SM. Crosstalk between histone modifications during the DNA damage response. Trends Cell Biol. 2009; 19:207–217. [PubMed: 19342239] Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Mitchell et al. Page 10 36. Bieging KT, Mello SS, Attardi LD. Unravelling mechanisms of p53-mediated tumour suppression. Nat Rev Cancer. 2014; 14:359–370. [PubMed: 24739573] Author Manuscript 37. Guichard C, Amaddeo G, Imbeaud S, Ladeiro Y, Pelletier L, Maad IB, Calderaro J, Bioulac-Sage P, Letexier M, Degos F, et al. Integrated analysis of somatic mutations and focal copy-number changes identifies key genes and pathways in hepatocellular carcinoma. Nat Genet. 2012; 44:694– 698. [PubMed: 22561517] 38. Hussain SP, Schwank J, Staib F, Wang XW, Harris CC. TP53 mutations and hepatocellular carcinoma: insights into the etiology and pathogenesis of liver cancer. Oncogene. 2007; 26:2166– 2176. [PubMed: 17401425] 39. McGivern DR, Lemon SM. Virus-specific mechanisms of carcinogenesis in hepatitis C virus associated liver cancer. Oncogene. 2011; 30:1969–1983. [PubMed: 21258404] 40. Bressac B, Galvin KM, Liang TJ, Isselbacher KJ, Wands JR, Ozturk M. Abnormal structure and expression of p53 gene in human hepatocellular carcinoma. Proc Natl Acad Sci U S A. 1990; 87:1973–1977. [PubMed: 2155427] 41. Hsu IC, Tokiwa T, Bennett W, Metcalf RA, Welsh JA, Sun T, Harris CC. p53 gene mutation and integrated hepatitis B viral DNA sequences in human liver cancer cell lines. Carcinogenesis. 1993; 14:987–992. [PubMed: 8389256] Author Manuscript 42. Mesri EA, Feitelson MA, Munger K. Human viral oncogenesis: a cancer hallmarks analysis. Cell Host Microbe. 2014; 15:266–282. [PubMed: 24629334] 43. Classon M, Harlow E. The retinoblastoma tumour suppressor in development and cancer. Nat Rev Cancer. 2002; 2:910–917. [PubMed: 12459729] 44. McGivern DR, Villanueva RA, Chinnaswamy S, Kao CC, Lemon SM. Impaired replication of hepatitis C virus containing mutations in a conserved NS5B retinoblastoma protein-binding motif. J Virol. 2009; 83:7422–7433. [PubMed: 19458004] 45. Munakata T, Liang Y, Kim S, McGivern DR, Huibregtse J, Nomoto A, Lemon SM. REFERENCES [PubMed: 14709551] Author Manuscript 55. Street A, Macdonald A, McCormick C, Harris M. Hepatitis C virus NS5A-mediated activation of phosphoinositide 3-kinase results in stabilization of cellular beta-catenin and stimulation of beta- catenin-responsive transcription. J Virol. 2005; 79:5006–5016. [PubMed: 15795286] 56. Best SM. Viral subversion of apoptotic enzymes: escape from death row. Annu Rev Microbiol. 2008; 62:171–192. [PubMed: 18729734] 57. Fuentes-Gonzalez AM, Contreras-Paredes A, Manzo-Merino J, Lizano M. The modulation of apoptosis by oncogenic viruses. Virol J. 2013; 10:182. [PubMed: 23741982] 58. Deng L, Adachi T, Kitayama K, Bungyoku Y, Kitazawa S, Ishido S, Shoji I, Hotta H. Hepatitis C virus infection induces apoptosis through a Bax-triggered, mitochondrion-mediated, caspase 3- dependent pathway. J Virol. 2008; 82:10375–10385. [PubMed: 18768989] 59. Lan L, Gorke S, Rau SJ, Zeisel MB, Hildt E, Himmelsbach K, Carvajal-Yepes M, Huber R, Wakita T, Schmitt-Graeff A, et al. Hepatitis C virus infection sensitizes human hepatocytes to TRAIL-induced apoptosis in a caspase 9-dependent manner. J Immunol. 2008; 181:4926–4935. [PubMed: 18802096] Author Manuscript 60. Mateu G, Donis RO, Wakita T, Bukh J, Grakoui A. Intragenotypic JFH1 based recombinant hepatitis C virus produces high levels of infectious particles but causes increased cell death. Virology. 2008; 376:397–407. [PubMed: 18455749] 61. Yi M, Hu F, Joyce M, Saxena V, Welsch C, Chavez D, Guerra B, Yamane D, Veselenak R, Pyles R, et al. Evolution of a cell culture-derived genotype 1a hepatitis C virus (H77S.2) during persistent infection with chronic hepatitis in a chimpanzee. J Virol. 2014; 88:3678–3694. [PubMed: 24429362] 62. Saeed M, Shiina M, Date T, Akazawa D, Watanabe N, Murayama A, Suzuki T, Watanabe H, Hiraga N, Imamura M, et al. In vivo adaptation of hepatitis C virus in chimpanzees for efficient virus production and evasion of apoptosis. Hepatology. 2011; 54:425–433. [PubMed: 21538444] 63. Cleary SP, Jeck WR, Zhao X, Kuichen, Selitsky SR, Savich GL, Tan TX, Wu MC, Getz G, Lawrence MS, et al. Identification of driver genes in hepatocellular carcinoma by exome sequencing. Hepatology. 2013; 58:1693–702. [PubMed: 23728943] 64. Schulze K, Imbeaud S, Letouze E, Alexandrov LB, Calderaro J, Rebouissou S, Couchy G, Meiller C, Shinde J, Soysouvanh F, et al. Exome sequencing of hepatocellular carcinomas identifies new mutational signatures and potential therapeutic targets. Nat Genet. 2015; 47:505–511. [PubMed: 25822088] 65. Nault JC, Calderaro J, Di Tommaso L, Balabaud C, Zafrani ES, Bioulac-Sage P, Roncalli M, Zucman-Rossi J. REFERENCES Hepatitis C virus induces E6AP-dependent degradation of the retinoblastoma protein. PLoS Pathog. 2007; 3:1335–1347. [PubMed: 17907805] [HCV NS5B protein targets the Rb tumor suppressor protein for proteasomal degradation in infected cell cultures, suggesting that disruption of the Rb pathway may be a mechanism by which HCV directly increases hepatocellular proliferation and promotes cancer. (see also reference 46).] 46. Munakata T, Nakamura M, Liang Y, Li K, Lemon SM. Down-regulation of the retinoblastoma tumor suppressor by the hepatitis C virus NS5B RNA-dependent RNA polymerase. Proc Natl Acad Sci U S A. 2005; 102:18159–18164. [PubMed: 16332962] 47. Hutcheson J, Bourgo RJ, Balaji U, Ertel A, Witkiewicz AK, Knudsen ES. Retinoblastoma protein potentiates the innate immune response in hepatocytes: significance for hepatocellular carcinoma. Hepatology. 2014; 60:1231–1240. [PubMed: 24824777] 48. Kannan RP, Hensley LL, Evers LE, Lemon SM, McGivern DR. Hepatitis C virus infection causes cell cycle arrest at the level of initiation of mitosis. J Virol. 2011; 85:7989–8001. [PubMed: 21680513] 49. Walters KA, Syder AJ, Lederer SL, Diamond DL, Paeper B, Rice CM, Katze MG. Genomic analysis reveals a potential role for cell cycle perturbation in HCV-mediated apoptosis of cultured hepatocytes. PLoS Pathog. 2009; 5:e1000269. [PubMed: 19148281] 50. Marshall A, Rushbrook S, Davies SE, Morris LS, Scott IS, Vowler SL, Coleman N, Alexander G. Relation between hepatocyte G1 arrest, impaired hepatic regeneration, and fibrosis in chronic hepatitis C virus infection. Gastroenterology. 2005; 128:33–42. [PubMed: 15633121] Author Manuscript 51. Chin YE, Kitagawa M, Su WC, You ZH, Iwamoto Y, Fu XY. Cell growth arrest and induction of cyclin-dependent kinase inhibitor p21 WAF1/CIP1 mediated by STAT1. Science. 1996; 272:719– 722. [PubMed: 8614832] 52. Liu J, Ding X, Tang J, Cao Y, Hu P, Zhou F, Shan X, Cai X, Chen Q, Ling N, et al. Enhancement of canonical Wnt/beta-catenin signaling activity by HCV core protein promotes cell growth of hepatocellular carcinoma cells. PLoS One. 2011; 6:e27496. [PubMed: 22110662] 53. Park CY, Choi SH, Kang SM, Kang JI, Ahn BY, Kim H, Jung G, Choi KY, Hwang SB. Nonstructural 5A protein activates beta-catenin signaling cascades: implication of hepatitis C virus-induced liver pathogenesis. J Hepatol. 2009; 51:853–864. [PubMed: 19726098] Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Page 11 Page 11 Mitchell et al. 54. Street A, Macdonald A, Crowder K, Harris M. The Hepatitis C virus NS5A protein activates a phosphoinositide 3-kinase-dependent survival signaling cascade. J Biol Chem. 2004; 279:12232– 12241. Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. REFERENCES Telomerase reverse transcriptase promoter mutation is an early somatic genetic alteration in the transformation of premalignant nodules in hepatocellular carcinoma on cirrhosis. Hepatology. 2014; 60:1983–1992. [PubMed: 25123086] [The identification of TERT promoter mutations as an early genetic change in abnormal but pre-neoplastic tissue suggests a genetic basis for increased cancer risk in patients with cirrhosis. (Also see reference 66).] 66. Nault JC, Mallet M, Pilati C, Calderaro J, Bioulac-Sage P, Laurent C, Laurent A, Cherqui D, Balabaud C, Zucman-Rossi J. High frequency of telomerase reverse-transcriptase promoter somatic mutations in hepatocellular carcinoma and preneoplastic lesions. Nat Commun. 2013; 4:2218. [PubMed: 23887712] 67. Alison MR. Liver stem cells: implications for hepatocarcinogenesis. Stem Cell Rev. 2005; 1:253– 260. [PubMed: 17142862] Author Manuscript 68. Lemon SM, McGivern DR. Is hepatitis C virus carcinogenic? Gastroenterology. 2012; 142:1274– 1278. [PubMed: 22537433] 69. Spaniel C, Honda M, Selitsky SR, Yamane D, Shimakami T, Kaneko S, Lanford RE, Lemon SM. microRNA-122 abundance in hepatocellular carcinoma and non-tumor liver tissue from Japanese patients with persistent HCV versus HBV infection. PLoS One. 2013; 8:e76867. [PubMed: 24130799] Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Mitchell et al. Page 12 Page 12 70. Freeman AJ, Dore GJ, Law MG, Thorpe M, Von Overbeck J, Lloyd AR, Marinos G, Kaldor JM. Estimating progression to cirrhosis in chronic hepatitis C virus infection. Hepatology. 2001; 34:809–816. [PubMed: 11584380] Author Manuscript 71. Seeff LB. The history of the “natural history” of hepatitis C (1968-2009). Liver Int. 2009; 29(Suppl 1):89–99. [PubMed: 19207971] 72. El-Serag HB. Epidemiology of viral hepatitis and hepatocellular carcinoma. Gastroenterology. 2012; 142:1264–1273. e1261. [PubMed: 22537432] 73**. Kumar V, Kato N, Urabe Y, Takahashi A, Muroyama R, Hosono N, Otsuka M, Tateishi R, Omata M, Nakagawa H, et al. Genome-wide association study identifies a susceptibility locus for HCV-induced hepatocellular carcinoma. Nat Genet. 2011; 43:455–458. [PubMed: 21499248] [GWAS studies may identify host genetic variants that differ between chronic hepatitis C patients who either progress or do not progress to HCC. Although a potentially powerful technique, identification of patients at higher risk remains elusive. (Also see references 74 and 75).] 74. Miki D, Ochi H, Hayes CN, Abe H, Yoshima T, Aikata H, Ikeda K, Kumada H, Toyota J, Morizono T, et al. Variation in the DEPDC5 locus is associated with progression to hepatocellular carcinoma in chronic hepatitis C virus carriers. Nat Genet. 2011; 43:797–800. [PubMed: 21725309] Author Manuscript Author Manuscript 75. Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. REFERENCES Lange CM, Bibert S, Dufour JF, Cellerai C, Cerny A, Heim MH, Kaiser L, Malinverni R, Mullhaupt B, Negro F, et al. Comparative genetic analyses point to HCP5 as susceptibility locus for HCV-associated hepatocellular carcinoma. J Hepatol. 2013; 59:504–509. [PubMed: 23665287] 76. Raimondi S, Bruno S, Mondelli MU, Maisonneuve P. Hepatitis C virus genotype 1b as a risk factor for hepatocellular carcinoma development: a meta-analysis. J Hepatol. 2009; 50:1142–1154. [PubMed: 19395111] 77. Kanwal F, Kramer JR, Ilyas J, Duan Z, El-Serag HB. HCV genotype 3 is associated with an increased risk of cirrhosis and hepatocellular cancer in a national sample of U.S. Veterans with HCV. Hepatology. 2014; 60:98–105. [PubMed: 24615981] 78. Kohli A, Shaffer A, Sherman A, Kottilil S. Treatment of hepatitis C: a systematic review. Jama. 2014; 312:631–640. [PubMed: 25117132] 79. Kwon H, Lok AS. Does antiviral therapy prevent hepatocellular carcinoma? Antivir Ther. 2011; 16:787–795. [PubMed: 21900710] Author Manuscript 80. Lok AS, Everhart JE, Wright EC, Di Bisceglie AM, Kim HY, Sterling RK, Everson GT, Lindsay KL, Lee WM, Bonkovsky HL, et al. Maintenance peginterferon therapy and other factors associated with hepatocellular carcinoma in patients with advanced hepatitis C. Gastroenterology. 2011; 140:840–849. [PubMed: 21129375] Author Manuscript Mitchell et al. Page 13 HIGHLIGHTS • The incidence of HCV-associated liver cancer will rise over the next decade. • Chronic inflammation and hepatic fibrosis are important drivers of liver cancer. • HCV-infected cirrhotic liver represents a “field of cancerization”. • Genetic studies highlight a role for host genome instability in carcinogenesis. • Virus disruption of host responses to DNA damage may contribute to HCC. Author Manuscript Author Manuscript Author Manuscript Author Manuscript Mitchell et al. Page 14 Figure 1. Organization of the 9.7 kb positive-sense HCV RNA genome 5’ and 3’ untranslated regions (UTRs) contain cis-acting elements essential for virus replication, including a 5’ internal ribosome entry site (IRES) that drives cap-independent expression of a polyprotein (box) that is processed into 10 mature viral proteins by a combination of host and viral proteases. Their functions are highlighted above the genome. The core (C) protein and envelope glycoproteins E1 and E2 are structural components of the virion. Nonstructural (NS) proteins possess functions necessary for replication, including helicase (NS3), protease (NS2 and NS3/4A), and RNA-dependent RNA polymerase (NS5B) activities. NS4B and NS5A drive formation of the ‘membranous web’, a cytoplasmic structure where these proteins accumulate to direct viral RNA synthesis. REFERENCES NS2 and NS5A also function in virion assembly, whereas p7 is essential for egress. Virus-host interactions that may contribute to HCC development are highlighted below the genome. Core and NS5A expression have been linked to the generation of ROS that may contribute to host DNA damage. Multiple HCV proteins interact with and modulate host pathways to facilitate virus replication and may in theory promote carcinogenesis. i 1 O i ti f th 9 7 kb iti HCV RNA Author Manuscript Figure 1. Organization of the 9.7 kb positive-sense HCV RNA genome Figure 1. Organization of the 9.7 kb positive-sense HCV RNA genome 5’ and 3’ untranslated regions (UTRs) contain cis-acting elements essential for virus replication, including a 5’ internal ribosome entry site (IRES) that drives cap-independent expression of a polyprotein (box) that is processed into 10 mature viral proteins by a combination of host and viral proteases. Their functions are highlighted above the genome. The core (C) protein and envelope glycoproteins E1 and E2 are structural components of the virion. Nonstructural (NS) proteins possess functions necessary for replication, including helicase (NS3), protease (NS2 and NS3/4A), and RNA-dependent RNA polymerase (NS5B) activities. NS4B and NS5A drive formation of the ‘membranous web’, a cytoplasmic structure where these proteins accumulate to direct viral RNA synthesis. NS2 and NS5A also function in virion assembly, whereas p7 is essential for egress. Virus-host interactions that may contribute to HCC development are highlighted below the genome. Core and NS5A expression have been linked to the generation of ROS that may contribute to host DNA damage. Multiple HCV proteins interact with and modulate host pathways to facilitate virus replication and may in theory promote carcinogenesis. gu e . O ga at o o t e 9. b pos t ve se se CV N ge o e 5’ and 3’ untranslated regions (UTRs) contain cis-acting elements essential for virus replication, including a 5’ internal ribosome entry site (IRES) that drives cap-independent expression of a polyprotein (box) that is processed into 10 mature viral proteins by a combination of host and viral proteases. Their functions are highlighted above the genome. The core (C) protein and envelope glycoproteins E1 and E2 are structural components of the virion. Nonstructural (NS) proteins possess functions necessary for replication, including helicase (NS3), protease (NS2 and NS3/4A), and RNA-dependent RNA polymerase (NS5B) activities. Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. Curr Opin Virol. Author manuscript; available in PMC 2016 October 01. REFERENCES NS4B and NS5A drive formation of the ‘membranous web’, a cytoplasmic structure where these proteins accumulate to direct viral RNA synthesis. NS2 and NS5A also function in virion assembly, whereas p7 is essential for egress. Virus-host interactions that may contribute to HCC development are highlighted below the genome. Core and NS5A expression have been linked to the generation of ROS that may contribute to host DNA damage. Multiple HCV proteins interact with and modulate host pathways to facilitate virus replication and may in theory promote carcinogenesis. Author Manuscript Author Author Manuscript Author Manuscript Aut Author Manuscript Author Manuscript Mitchell et al. Page 15 Figure 2. Model for HCV-associated carcinogenesis Hepatocellular carcinoma (HCC) likely results from a combination of indirect host- and direct HCV-mediated mechanisms. Persistent immune-mediated inflammation, coupled with expression of core and NS5A, generates ROS that trigger oxidative DNA damage. HCV infection further compromises host genome stability by impairing DNA repair pathways. Repeated cycles of hepatocellular destruction with regenerative proliferation and progressive fibrosis within this pro-mutagenic environment result in a “cancer field” comprised of pre- neoplastic but genetically-altered hepatocytes. Continued hepatocellular turnover may, in turn, select for aberrant hepatocytes with growth advantages. Whether HCC arises directly from HCV-infected hepatocytes remains unclear, although HCV infection may enhance survival of abnormal hepatocytes by promoting cell proliferation and inhibiting apoptosis. Ultimately, these combined mechanisms may select for transformed cells, culminating in development of HCC over 2-3 decades of persistent HCV infection. Page 15 Author Manuscript Figure 2. Model for HCV-associated carcinogenesis Hepatocellular carcinoma (HCC) likely results from a combination of indirect host- and direct HCV-mediated mechanisms. Persistent immune-mediated inflammation, coupled with expression of core and NS5A, generates ROS that trigger oxidative DNA damage. HCV infection further compromises host genome stability by impairing DNA repair pathways. Repeated cycles of hepatocellular destruction with regenerative proliferation and progressive fibrosis within this pro-mutagenic environment result in a “cancer field” comprised of pre- neoplastic but genetically-altered hepatocytes. Continued hepatocellular turnover may, in turn, select for aberrant hepatocytes with growth advantages. Whether HCC arises directly from HCV-infected hepatocytes remains unclear, although HCV infection may enhance survival of abnormal hepatocytes by promoting cell proliferation and inhibiting apoptosis. Ultimately, these combined mechanisms may select for transformed cells, culminating in development of HCC over 2-3 decades of persistent HCV infection. Hepatocellular carcinoma (HCC) likely results from a combination of indirect host- and direct HCV-mediated mechanisms. REFERENCES Persistent immune-mediated inflammation, coupled with expression of core and NS5A, generates ROS that trigger oxidative DNA damage. HCV infection further compromises host genome stability by impairing DNA repair pathways. Repeated cycles of hepatocellular destruction with regenerative proliferation and progressive fibrosis within this pro-mutagenic environment result in a “cancer field” comprised of pre- neoplastic but genetically-altered hepatocytes. Continued hepatocellular turnover may, in turn, select for aberrant hepatocytes with growth advantages. Whether HCC arises directly from HCV-infected hepatocytes remains unclear, although HCV infection may enhance survival of abnormal hepatocytes by promoting cell proliferation and inhibiting apoptosis. Ultimately, these combined mechanisms may select for transformed cells, culminating in development of HCC over 2-3 decades of persistent HCV infection. Author Manuscript Author Manuscript Auth Author Manuscript Author Manuscript
https://openalex.org/W2069493887	https://europepmc.org/articles/pmc3489572?pdf=render	English	null	Intergenerational enrollment and expenditure changes in Medicaid: trends from 1991 to 2005	BMC health services research	2,012	cc-by	4,476	* Correspondence: stepatri@med.umich.edu 1The Department of Pediatrics and Communicable Diseases, University of Michigan Health System, Ann Arbor, MI 48109, USA 2Robert Wood Johnson Foundation Clinical Scholars Program, University of Michigan Health System, Ann Arbor, MI 48109, USA Full list of author information is available at the end of the article Abstract Background: From its inception, Medicaid was aimed at providing insurance coverage for low income children, elderly, and disabled. Since this time, children have become a smaller proportion of the US population and Medicaid has expanded to additional eligibility groups. We sought to evaluate relative growth in spending in the Medicaid program between children and adults from 1991-2005. We hypothesize that this shifting demographic will result in fewer resources being allocated to children in the Medicaid program. Methods: We utilized retrospective enrollment and expenditure data for children, adults and the elderly from 1991 to 2005 for both Medicaid and Children’s Health Insurance Program Medicaid expansion programs. Data were obtained from the Centers for Medicare and Medicaid Services using their Medicaid Statistical Information System. Results: From 1991 to 2005, the number of enrollees increased by 83% to 58.7 million. This includes increases of 33% for children, 100% for adults and 50% for the elderly. Concurrently, total expenditures nationwide rose 150% to $273 billion. Expenditures for children increased from $23.4 to $65.7 billion, adults from $46.2 to $123.6 billion, and elderly from $39.2 to $71.3 billion. From 1999 to 2005, Medicaid spending on long-term care increased by 31% to $84.3 billion. Expenditures on the disabled grew by 61% to $119 billion. In total, the disabled account for 43% and long-term care 31%, of the total Medicaid budget. Conclusion: Our study did not find an absolute decrease in the overall resources being directed toward children. However, increased spending on adults on a per-capita and absolute basis, particularly disabled adults, is responsible for much of the growth in spending over the past 15 years Medicaid expenditures have grown faster Conclusion: Our study did not find an absolute decrease in the overall resources being directed toward children. However, increased spending on adults on a per-capita and absolute basis, particularly disabled adults, is responsible for much of the growth in spending over the past 15 years. Medicaid expenditures have grown faster than inflation and overall national health expenditures. A national strategy is needed to ensure adequate coverage for Medicaid recipients while dealing with the ongoing constraints of state and federal budgets. responsible for much of the growth in spending over the past 15 years. Medicaid expenditures have grown faster than inflation and overall national health expenditures. Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Open Access © 2012 Patrick and Freed; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Intergenerational enrollment and expenditure changes in Medicaid: trends from 1991 to 2005 Stephen W Patrick1,2,3,5* and Gary L Freed1,4 Stephen W Patrick1,2,3,5* and Gary L Freed1,4 Abstract A national strategy is needed to ensure adequate coverage for Medicaid recipients while dealing with the ongoing constraints of state and federal budgets. Keywords: Medicaid, CHIP, Long-term care, Disabled Keywords: Medicaid, CHIP, Long-term care, Disabled Background in covering a greater proportion of poor children and pregnant women. Medicaid has also expanded coverage beyond what could have been conceptualized in 1965 to include such things as coverage for breast cancer and AIDS treatment. Furthermore, all states now cover many “optional” services -including prescription drugs and long-term care. When Medicaid, Title XIX of the Social Security Act, began in 1965, it was aimed at providing insurance coverage for low income children, elderly, and disabled. Medicaid, an entitlement program, relies on a unique federal-state partnership for financing and is adminis- tered at the state level. Because of this, Medicaid has been able to “formulate creative structural solutions and implement reforms” [1] tailored to state needs. Through the years the program has seen incremental expansions The Children’s Health Insurance Program (CHIP), Title XXI of the Social Security Act, began in 1998 and aimed to provide health insurance coverage to low in- come, uninsured children not eligible for Medicaid. Un- like Medicaid, CHIP is a block grant, not an entitlement program. As with Medicaid, CHIP is financed with both federal and state dollars and is administered at the state level. In administering the program, states are given the Page 2 of 7 Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 option of having CHIP combined with their Medicaid program, as a stand-alone program or as an expansion to their Medicaid program. States that choose to have a Medicaid expansion CHIP program must meet the same mandated benefit packages as Medicaid [2]. each group we evaluated their expenditures for long- term care, and also enrollment and expenditures for those whose basis for eligibility in the Medicaid program was due to being classified as “blind/disabled.” Long- term care includes expenditures on Home Health Ser- vices, Intermediate Care Facility Services for the Mentally Retarded, Mental Health Facility Services and Personal Support Services. All findings were adjusted for inflation based upon the year 2005, unless otherwise specified. For simplicity, we refer to both Medicaid and Medic- aid Expansion CHIP programs as “Medicaid.” Historically, because of Medicaid and CHIP’s partial dependence upon state general revenues, when state rev- enues dropped and healthcare costs rose, programs were often cut or curtailed [1]. During harsh economic times earlier this decade, states were faced with growing enrollment, medical inflation and plummeting state revenues [3]. Background In response, many states implemented bar- riers to secure and retain coverage in their Medicaid and CHIP programs [4]. Recently, as the U.S. economy has faltered, enrollment has increased, medical inflation has outpaced overall inflation [5] and state budgets are again in crisis. The data do not include Disproportionate Share Hos- pital (DSH) Payments and only include enrollment and expenditures for Medicaid Expansion CHIP. The cat- egory “adjustments” are lump sum payments made for Medicaid enrollees. These payments include add-on pro- grams including those for inpatient, outpatient, psychi- atric, pediatric, critical care and Medicaid high-volume providers. An example would be the Illinois Rural Crit- ical Hospital Adjustment Payment Process, which is a quarterly payment program that provides rural Illinois hospitals with additional Medicaid payments based upon either their obstetrical care or general care admissions from a pre-determined base period [10]. Because all data were aggregate, de-identified data obtained from public sources, this study was exempt from human subjects consideration. While Medicaid has continued to expand in scope, one of the program’s original target populations, children, have become a smaller proportion of the United States population. In 1966 those under age 18 accounted for 37 percent of the total population [6]; whereas, in 2005 they accounted for just 25 percent [7]. It has been speculated that as America ages, resources will be preferentially allo- cated to the growing elderly population [8]. Previous studies have evaluated total United States social spending and found that spending per capita for children, particu- larly in challenging economic times, has not kept pace with per capita spending for the elderly [9,10]. We hypothesize that this shifting demographic will result in fewer resources being allocated to children in the Medic- aid program. Trends in Medicaid enrollment From 1991 to 2005, the absolute number of Medicaid enrollees nearly doubled from 32.2 to 58.7 million. Chil- dren increased by more than 33 percent to 31.8 million, adults increased by 100 percent to 20.7 million, and the elderly increased by approximately 50 percent from 4 to 6 million (Figure 1). Methods Data from all 50 states were obtained from the Centers for Medicare and Medicaid Services (CMS) using their Medicaid Statistical Information System (MSIS). Prior to 1999, data were collected using the HCFA-2082 hard- copy reporting process. Beginning with FY 1999, the Balanced Budget Act (BBA) of 1997 requires states to submit all eligibility and claims data to CMS on a quar- terly basis through the MSIS. Because of this reporting change, data after 1999 were available to examine specific age groups in specific service categories such as long- term care. We evaluated data from the MSIS from 1991 to 2005, looking at trends in enrollment and expendi- tures focusing on children, adults and the elderly. Because of the constraints of the data sources, for ana- lyses which include data from 1991 to 1998, children are considered aged birth to 20 years and adults are 21 to 64 years; whereas, for analyses which includes data only after 1999 children are considered birth to 18 years and adults 19 to 64 years. Throughout all analyses, the elderly are always categorized as aged 65 years and greater. For Overall, children as a proportion of total Medicaid en- rollment held steady at approximately 55 percent. Enroll- ment for non-elderly adults increased from 31.3 percent to 35.4 percent of all beneficiaries, and the proportion of elderly decreased from 12.5 percent to 10.3 percent. Trends in Medicaid expenditures Over the same time period, total annual Medicaid expen- ditures nationwide rose from $110 to $273 billion, an increase of almost 150 percent. This includes increases in expenditures for children from $23.4 to $65.7 billion, adults from $46.2 to $123.6 billion, and elderly from $39.2 to $71.3 billion (Figure 2). From 1991 to 2005, there were increased per capita expenditures across all groups, averaging increases of $763 for children, $1,366 for adults and $2,133 for the elderly. Average cost per enrollee in the entire program Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Page 3 of 7 Figure 1 Total Medicaid enrollees, by age, from 1991-2005. Figure 1 Total Medicaid enrollees, by age, from 1991-2005. increased by $1,219, with the highest cost per enrollee for the elderly and the lowest for children (Figure 3). Long-term care Long-term care remains a major proportion of the over- all Medicaid budget. In 1999, long-term care accounted for 37.3 percent of total Medicaid expenditures, decreas- ing to 30.9 percent in 2005. For the elderly, long-term care continues to account for the majority of their Medicaid expenditures – 70 percent in 1999 and 62.9 percent in 2005. Expenditures on long-term care have steadily increased, but have not grown as rapidly as overall Medicaid expenditures. Since 1991, expenditures for children and adults as a proportion of overall Medicaid expenditures increased from 21.2 to 24.1 percent and 41.8 to 45.3 percent, re- spectively; whereas, the overall proportion of expendi- tures for the elderly decreased from 35.5 percent to 26.1 percent. From 1991 to 2005, children accounted for 52.2 per- cent of growth in enrollment and 26.0 percent of growth in expenditures. This is in contrast to adults who accounted for 40.3 percent of enrollment growth and 47.6 percent of growth in expenditures, and the elderly which accounted for 7.5 percent enrollment growth and 19.7 percent expenditures growth (Table 1). From 1999 to 2005, Medicaid spending on long-term care increased by 31 percent or $19.9 billion; from $64.4 to $84.3 billion. This increase was consistent in all age groups, with an increase for children of $2.5 billion, from $6.4 to $8.9 billion, for adults of $8.4 billion, from Figure 2 Total Medicaid expenditures, by age, from 1991 to 2005. Figure 2 Total Medicaid expenditures, by age, from 1991 to 2005. Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Page 4 of 7 Figure 3 Medicaid expenditures per enrollee, by age, 1991-2005. Figure 3 Medicaid expenditures per enrollee, by age, 1991-2005. $20.7 to $29.1 billion, and for the elderly of $10 billion, from $36.7 to $44.9 billion (Figure 4). comparison, over the same period, national health expen- ditures grew by approximately 165 percent and overall inflation increased by 43 percent [5,11,12]. We hypothe- sized that the decreasing proportion of children in the United States would result in decreased resources being allocated to children; however, this did not occur during our study period. It is likely that child enrollment and expenditures in Medicaid were bolstered by new pro- grams targeted at children, such as the Children’s Health Insurance Program. Our study does find, however, that most of the new spending in the program has been tar- geted toward adults. The disabled Al h h h Although the disabled account for fifteen percent of Medicaid enrollees, they account for over forty percent of total Medicaid expenditures - disabled adults alone now accounting for over a third of total Medicaid expen- ditures. The absolute number of disabled beneficiaries has grown for several reasons. While states have borne the responsibility of caring for the disabled through the years, their emergence in the Medicaid program has been more recent. Over the last several decades, care for the disabled has been gradually deinstitutionalized and in- creasingly “medicalized” through an effort to “convert service programs for disabled children, developmentally disabled and the chronically ill to Medicaid-based finan- cing” [13]. In addition, since Medicaid’s enactment, there have been incremental expansions in eligibility, and the group now encompasses “disabled children, physically disabled but cognitively intact nonelderly adults, the de- velopmentally disabled and people with severe and Disabled The majority of those whose basis of eligibility for Medicaid is “disabled” have been, and continue to be, aged 19-64 years. The number of disabled adults enrolled in Medicaid grew steadily from 1999 to 2005, while the numbers of disabled children and the elderly did not change markedly. From 1999 to 2005, adult disabled enrollees increased from 5.4 million to 6.8 million, while the number of disabled children on Medicaid grew from 1.25 million to 1.3 million and the elderly from 641,000 to 699,000. Disabled adults continue to be the most costly among the disabled, with expenditures increasing from $58 bil- lion in 1999 to $95.3 billion in 2005. In 2005, this group alone accounted for greater than 35 percent of total Medicaid expenditures. Disabled children and elderly experienced smaller increases from 1999 to 2005 - disabled children increasing from $10.6 to $15 billion and disabled elderly increasing from $5.3 to $8.4 billion (Figure 5). Long-term care We find that expenditures for adults in Medicaid are growing faster than children or the elderly. Discussion Without adjusting for inflation, from 1991 to 2005, Medicaid expenditures grew by almost 250 percent. By Table 1 Growth in Enrollment and Expenditures by Age, 1991-2005 Enrollment Expenditures Children 52.2% 26.0% Adults 40.3% 47.6% Elderly 7.5% 19.7% Table 1 Growth in Enrollment and Expenditures by Age, 1991-2005 Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Page 5 of 7 Figure 4 Medicaid Expenditures on Long-Term Care, by age, 1999-2005. (Long-Term Care includes expenditures on Home Health Services, Intermediate Care Facility Services for the Mentally Retarded, Mental health Facility Services and Personal Support Services). Figure 4 Medicaid Expenditures on Long-Term Care, by age, 1999-2005. (Long-Term Care includes expenditures on Home Health Services, Intermediate Care Facility Services for the Mentally Retarded, Mental health Facility Services and Personal Support Services). Authors’ contribution ll h ( ) All authors (SP, GF) contributed to the analysis, interpretations of the data, drafting and revision of the manuscript. All authors approved the final manuscript. Funding source h h ld The authors would like to thank the Robert Wood Johnson Foundation for their support of this work. Conclusions By 2005, Medicaid provided health insurance coverage more than a quarter of US children and nearly one in ten US adults and elderly [23]. Growth in the Medicaid pro- gram outpaces both inflation and national health expen- ditures. In our current economic climate, where both budgetary cuts and eligibility expansions are being pro- posed, it is fair to ask where the breaking point of the Medicaid program might be. If Medicaid continues to grow at this pace it will continue to be a target for state and federal legislators looking to reduce overall state expenditures. The challenge for our nation’s future will be how to provide needed health care to each of Medi- caid’s target populations within the constraints of the current US economy. We believe providing health care to our nation’s safety net, including our children, must continue to be a national priority. Innovative financing and structural changes to Medicaid which would ensure access to health care while being fiscally prudent must be a research priority for health services researchers in the coming years. Author details 1 1The Department of Pediatrics and Communicable Diseases, University of Michigan Health System, Ann Arbor, MI 48109, USA. 2Robert Wood Johnson Foundation Clinical Scholars Program, University of Michigan Health System, Ann Arbor, MI 48109, USA. 3Division of Neonatal-Perinatal Medicine, University of Michigan Health System, Ann Arbor, MI 48109, USA. 4The Department of Pediatrics and Communicable Diseases, Child Health Evaluation and Research Unit, University of Michigan Health System, Ann Arbor, MI 48109, USA. 56312 Medical Science Building 1, 1150 W. Medical Center Drive SPC 5604, Ann Arbor, MI 48109-5604, USA. Future implications Health reform A key component of the Patient Protection and Affordable Care Act includes expansion of Medicaid eligibility to pro- vide coverage for all Americans up to 133 percent of the federal poverty level [21]. This will result in millions of new enrollees to the Medicaid program. These enrollees, par- ticularly childless adults, are not currently eligible in many states. To encourage states to enrollee newly eligible popu- lations, beginning in 2014, the federal government will bear 100 percent of the costs of these new enrollees, decreasing to 90 percent in 2020. Governors, particularly in states with high levels of poverty, are expressing concern that such an expansion could financially cripple states [22]. Support The Robert Wood Johnson Foundation Clinical Scholars Program provided support for this project. Conflicts Th h The authors have no conflicts of interests or corporate sponsors to disclose. State effects Th i The strain of the rapid growth of the Medicaid program is clearly felt by the states. In 1991, elementary and sec- ondary education accounted for an average of 22.4 per- cent of state budgets compared to 13.6 percent for Medicaid [17]. By 2007, Medicaid tied elementary and secondary education as the most expensive line item in state budgets, accounting for an average of 21.2 percent of the total; nine states reported spending more than a quarter of their state budget on their Medicaid program, with one state, Missouri, spending 35 percent [18]. States adopted several strategies to contain costs, includ- ing: changes in income verification requirements and en- rollment procedures [4], changes to eligibility rules, changes in eligibility standards (such as reducing coverage for par- ents) provider payment rate reductions or freezes, increased controls on pharmacy utilization, optional benefit elimin- ation for adults, increasing premiums and copayments where allowed, increasing managed care (including auto- matically assigning enrollees), instituting disease and case management programs, reducing payments to long-term care facilities, and investigations into fraud and abuse [19]. Given the broad nature of these cost-containment measures employed earlier this decade, there is concern that states are left with few opportunities for further cost savings [20]. Long-term care persistent mental illness,” and low-income people with AIDS [13]. Lastly, there have been greatly reduced mor- tality rates among certain categories of disabled adults, resulting in longer life spans. Enrollment and expendi- tures for the disabled are likely to increase as medical technology continues to improve, extending life for those who would not have survived even just a few years ago. Thus, proportional increases in spending for disabled adults are likely to continue. Medicaid continues to be the largest purchaser of long- term care in the United States [14] and this category accounts for nearly one-third of total Medicaid expendi- tures. A significant proportion of Medicaid enrollees that require long-term care are also enrolled in Medicare – termed “dual-eligible beneficiaries.” Medicaid assists these low income Medicare beneficiaries by paying Medicare premiums and providing Medicare uncovered services Figure 5 Total Medicaid expenditures for the Blind/Disabled, by age, 1991-2005. Figure 5 Total Medicaid expenditures for the Blind/Disabled, by age, 1991-2005. Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Page 6 of 7 government. In the process of reporting, both errors of omission and commission are possible. Additionally, while we aggregate all states together to make infer- ences at the national level, this approach could overlook effects at the state level. such as long-term care [15]. Expenditures on long-term care rose faster than inflation, but not as fast as total Medicaid expenditures. This is because of a trend to- wards community-based instead of institution based ser- vice delivery [13]. More emphasis has been placed on community-based treatment after the Supreme Court decision of June of 1999 in the case of Olmstead vs. LC, which requires states to provide community-based treatment where appropriate [16]. Abbreviations (CHIP): Children’s health insurance program; (CMS): Centers for medicare and Medicaid services; (MSIS): Medicaid statistical information system; (BBA): Balanced budget act; (DSH): Disproportionate share hospital; (AIDS): Acquired immune deficiency syndrome. Competing interests Both authors declare that they have no competing interests. Competing interests Both authors declare that they have no competing interests. Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 Patrick and Freed BMC Health Services Research 2012, 12:327 http://www.biomedcentral.com/1472-6963/12/327 References 1. Brown L, Sparer M: Poor Program’s Progress: The Unanticipated Politics of Medicaid Policy. Health Aff 2003, 22(1):31–44. y 2. Kaiser Family Foundation: State Children’s Health Insurance Program (SCHIP) at a Glance. 2007:1–2. Available at: http://www.kff.org/medicaid/upload/ 7610.pdf. 3. Weil A: There’s Something About Medicaid. Health Aff 2003, 22(1):13–30. 4. Ross D, Cox L: Beneath the Surface: Barriers Threatened to Slow Pro Expanding Health Coverage of Children and Families. 2004:1–62. 5. Catlin A, Cowan C, Heffler S, Washington B and the National Health Expenditure Accounts Team: National Health Spending in 2005: The Slowdown Continues. Health Aff 2007, 26(1):142–153. 5. Catlin A, Cowan C, Heffler S, Washington B and the National Health Expenditure Accounts Team: National Health Spending in 2005: The Slowdown Continues. Health Aff 2007, 26(1):142–153. 6. United States Census Bureau: Resident Population plus Armed Forces Overseas--Estimates by Age, Sex, and Race: July 1, 1966. Retrieved September 19, 2012. Available from: http://www.census.gov/popest/data/national/asrh/ pre-1980/tables/PE-11-1966.pdf. 6. United States Census Bureau: Resident Population plus Armed Forces Overseas--Estimates by Age, Sex, and Race: July 1, 1966. Retrieved September 19, 2012. Available from: http://www.census.gov/popest/data/national/asrh/ pre-1980/tables/PE-11-1966.pdf. 7. United States Census Bureau: Resident Population by Race, Hispanic Origin, and Age. Retrieved September 19, 2012. Available from: http://www.census. gov/compendia/statab/2012/tables/12s0010.xls. 7. United States Census Bureau: Resident Population by Race, Hispanic Origin, and Age. Retrieved September 19, 2012. Available from: http://www.census. gov/compendia/statab/2012/tables/12s0010.xls. g p 8. Freed G, Fant K: The Impact of the ‘Aging Of America’ on Children. Health Aff 2004, 23(2):168–174. 8. Freed G, Fant K: The Impact of the ‘Aging Of America’ on Children. Health Aff 2004, 23(2):168–174. 9. Pati S, Keren R, Alessandrini E, Schwarz D: Generational Differences In US Public Spending, 1980-2000. Health Aff 2004, 23(5):131–141. 9. Pati S, Keren R, Alessandrini E, Schwarz D: Generational Differences In US Public Spending, 1980-2000. Health Aff 2004, 23(5):131–141. p g 10. Benjamin A, Newacheck P, Wolfe H: Intergenerational Equity and Public Spending. Pediatrics 1991, 88:75–83. 11. Letsch S: National Health Care Spending In 1991. Health Aff 1993, 12(1):94–110. 12. United States Department of Labor, Bureau of Labor Statistics: Consumer Price Index, all Urban Consumers - (CPI-U). Retrieved September 19, 2012. Available from: ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt. 13. Vladeck B: Where the Action Really Is: Medicaid and the Disabled. Health Aff 2003, 22(1):90–100. 14. Kaiser Family Foundation: Medicaid and Long-Term Care Services and Supports: 2009. Available from: www.kff.org/medicaid/upload/2186_06.pdf. pp g p p 15. Kaiser Commission on Medicaid and the Uninsured: Medicaid: A Primer. References 2009:1–21. Available from: www.kff.org/medicaid/upload/416 7334-03.pdf. 16. Olmstead v. L. C. (98-536) 527 U.S. 581: 1999. Retrieved September 19, 2012. Available at: http://www.law.cornell.edu/supct/html/98-536.ZS.html. 17. National Association of State Budget Officers: State Expenditure Report, 1991. 1991:1–82. Available from: http://www.nasbo.org/sites/default/files/ER_1991. PDF. 18. National Association of State Budget Officers: State Expenditure Report, Fiscal Year 2007. 2008:1–112. Available from: http://www.nasbo.org/sites/default/ files/ER_2007.pdf. 18. National Association of State Budget Officers: State Expenditure Report, Fiscal Year 2007. 2008:1–112. Available from: http://www.nasbo.org/sites/default/ files/ER_2007.pdf. 19. Smith V, et al: The Continuing Medicaid Budget Challenge: State Medicaid Spending Growth and Cost Containment in Fiscal Years 2004 and 2005. 2004:1–85. 19. Smith V, et al: The Continuing Medicaid Budget Challenge: State Medicaid Spending Growth and Cost Containment in Fiscal Years 2004 and 2005. 2004:1–85. 20. Dorn S, Garrett B, Holahan J, Williams A: Medicaid, SCHIP, and Economic Downturn: Policy Challenges and Policy Responses: 2008. 20. Dorn S, Garrett B, Holahan J, Williams A: Medicaid, SCHIP, and Economic Downturn: Policy Challenges and Policy Responses: 2008. 21. Kaiser Family Foundation: Side-By-Side Comparison of Major Health Care Reform Proposals. 2009. [cited 2009 September 22, 2009]; Available from: www.kff.org/healthreform/upload/healthreform_sbs_full.pdf. 22. Pettus E: Governors Worry Federal Health Reform Could Strain Budgets In States with many Poor, Jobless. New York, NY: Associated Press; 2009. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit 23. United States Census Bureau: Current Population Survey, 1988 to 2006 Annual Social and Economic Supplements, Table HI-2. Limitations d Our study does have some clear limitations. First, data from MSIS is reported from states to the federal Page 7 of 7 Page 7 of 7 References Health Insurance Coverage Status and Type of Coverage--All People by Age and Sex: 1987 to 2005. Accessed 2012 July 22; Available from: http://www.census.gov/hhes/hlthins/ data/historical/orghihistt2.html. Submit your next manuscript to BioMed Central and take full advantage of: Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission doi:10.1186/1472-6963-12-327 Cite this article as: Patrick and Freed: Intergenerational enrollment and expenditure changes in Medicaid: trends from 1991 to 2005. BMC Health Services Research 2012 12:327. doi:10.1186/1472-6963-12-327 Cite this article as: Patrick and Freed: Intergenerational enrollment and expenditure changes in Medicaid: trends from 1991 to 2005. BMC Health Services Research 2012 12:327.
https://openalex.org/W3153897674	https://popups.uliege.be/esaform21/pdf.php?id=4641	English	null	Raster analysis of Fused Filament Fabrication process	ESAFORM 2021	2,021	cc-by	3,733	RRast aster anal er analyysis of F sis of Fused Filament F used Filament Fabrication pr abrication process ocess Roberto Spina. Dip. di Meccanica, Matematica e Management, Politecnico di Bari, Bari, Italy Istituto Nazionale di Fisica Nucleare (INFN) - Sezione di Bari - Sezione di Bari, Bari, Italy Consiglio Nazionale delle Ricerche - Istituto di Fotonica e Nanotecnologie (CNR-IFN), Via Amendola 173,70126 Bari, Italy Corresponding author: Roberto SPINA. E-mail address: roberto.spina@poliba.it Bruno Cavalcante. Istituto Nazionale di Fisica Nucleare (INFN) - Sezione di Bari - Sezione di Bari, Bari, Italy Roberto Spina. Dip. di Meccanica, Matematica e Management, Politecnico di Bari, Bari, Italy Istituto Nazionale di Fisica Nucleare (INFN) - Sezione di Bari - Sezione di Bari, Bari, Italy Consiglio Nazionale delle Ricerche - Istituto di Fotonica e Nanotecnologie (CNR-IFN), Via Amendola 173,70126 Bari, Italy Corresponding author: Roberto SPINA. E-mail address: roberto.spina@poliba.it Bruno Cavalcante. Istituto Nazionale di Fisica Nucleare (INFN) - Sezione di Bari - Sezione di Bari, Bari, Italy AAbstr bstract act.. The objective of the present work is to study the raster generation to realize Fused Filament Fabrication parts. The research in this paper focused on the evaluation of the deposition of a simple geometry with a FFF machine, supported by an analytical model to compute the build time, also evaluating the geometrical variations caused by changes in process parameters. The main parameters were the print temperature and speed as a function of the thermal and rheological properties of the PLA filament. The study identified essential correlations between process parameters, raster dimensions, and filament properties. An experimental procedure, supported by an analytical model, was implemented for computing raster time and material dimensions. KKeeyw ywor ords ds. Rapid Prototyping, Fused Filament Fabrication, Polymer Filament 1 Intr 1 Introduction oduction Fused Filament Fabrication (FFF) is the most widespread AM processes used for thermoplastic prototypes, tools, and low- volume products. It is a new name for Fused Deposition Modeling, an extrusion-based process in which products are realized by melting polymer-based filaments and depositing molten materials on a horizontal build platform. These polymers soften over a wide range of temperature, forming a high-viscosity material ideal for material extrusion through a 0.2–0.5 mm diameter nozzle [1]. The calculation of layer contours from a digital model of the part (slicing) is the basis for the generation of extrusion trajectories by appropriate programming software [2], useful to produce free- defects components. A comprehensive review of the FFF process parameters categorizes them into slicing parameters, building orientation and temperature conditions, all affecting the filament (inter-roads and intra-layers) bonding, and thus influencing the mechanical properties of final parts. [3]. Some studies link the printing parameters and physical polymer properties to the viscosity and shape of the filament immediately before deposition in the FFF process. Focusing on the specific polymer such as Polylactic Acid (PLA), this is characterized by a higher cost of production compared to petroleum-derived, non-biodegradable materials whereas it is inherently brittle, despite its tunable tacticity, microstructure, and mechanical properties. These features limit the use of PLA in a wide range of short-term applications [4]. The printed strand shape to form the raster has a direct impact on the part surface roughness, cooling process after printing, and bond formation between adjoining raster and overlaid layers [5]. For this reason, the dimensional accuracy and the mechanical properties of the final part were indirectly influenced, although they mainly depend on other processing parameters [6]. The deposition became complex in the presence of fibers because it is necessary to evaluate the Newtonian extrudate swell at the nozzle exit. The effects of the nozzle convergence zone and the extrudate swell at the nozzle exit have important effects on the fiber orientation state and, therefore, the resulting material stiffness of the extruded polymer [7]. An analytical model can be used to predict the flow within the nozzle, as well as the melting rate in an FFF process. The model can also be used to optimize the melting within a material extrusion additive manufacturing process [8]. ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 Raster analysis of Fused Filament Fabrication process parameters were evaluated, also considering the thermal and rheological properties of the PLA filament. The study identified essential correlations between process parameters, raster dimensions, and filament properties. parameters were evaluated, also considering the thermal and rheological properties of the PLA filament. The study identified essential correlations between process parameters, raster dimensions, and filament properties. 1 Intr 1 Introduction oduction The research in this paper focused on the evaluation of the deposition of a simple raster geometry with a FFF machine, supported by an analytical model to compute the build time. The geometrical variations caused by changes in process 4641/1 2 R 2 Rast aster design er design The experimental activity consisted of a deposition of a non-complex geometrical raster with a series of horizontal/ vertical segments extruded in fast sequence (Figure 1). The focus was on the extrusion of this geometry intending to evaluate processing potentials in terms of part production time, quality in the repeatability of product dimensions with a variation of processing parameters such as print speed and temperature, evaluating the surface dimensions. The filaments, made of PLA, had a 1.75 mm diameter with a dimensional diameter tolerance ±25 μm. The main physical and mechanical properties, collected from material data sheets (M3D LLC in Fulton, USA), are reported in Table 1. TTable 1. Mat able 1. Material data. erial data. 4641/2 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 Fig 1 R Fig 1 Rast aster geometry er geometry Fig. 1. R Fig. 1. Rast aster geometry er geometry 3 Multi-ph 3 Multi-phyysical model sical model A multi-physics model was implemented to predict the heat transfer, stresses and, deformations during filament extrusion using COMSOL Multiphysics (Comsol Inc., Burlington, USA). The filament was deposited on a glass plate (100×100×3.4 mm3), following the design path. Coupled thermal and mechanical analyses were carried out by using a segregated solver splitting the Jacobian matrix into smaller sub-problems, usually one for each degree of freedom (temperature and displacement). The segregated approach treated each physics sequentially, using the results of the previously solved physics to evaluate the loads and material properties for the next physics being solved. The advantages of this approach are several such as the use of an optimal iterative solver in each linear sub-step, a more memory-efficient faster solver and, less time for solving iterations than a fully coupled approach. The temperature fields were initially computed during the thermal analysis and then recomputed during the mechanical one to evaluate the induced strain and stress fields at each simulation step. The FE mesh consisted of about 210,000 elements (quadratic hexahedron) with a constant element length of 100 μm and an average mesh quality of 0.82. A high number of small- sized elements were used for the deposition path, whereas large-sized elements were employed for the glass plate. The numerical model reproduced the filament deposition onto the substrate concerning time, using the element birth-death technique [12]. The main assumptions of the model were: 1) The temperature of the glass plate was 50°C at the start of the process. 2) A Gaussian heat flux distribution was applied to the top surfaces of the deposition path. 3) Free convections and radiations were applied to all external surfaces. 4) A contact surface with a temperature continuity boundary was defined between the path and the glass plate. 5) No external loads except heat fluxes were applied to the mechanical model. 6) The mechanical boundary conditions constrained movements along X, Y, and Z axes in the glass plate's bottom region. All elements of the deposition path were active (birth) or inactive (death). The dead elements were kept in the model with the material properties equal to those of the air [13]. The activation sequence strictly followed the physical deposition. The nodes at the nozzle travel were activated and remained active, restoring the original material thermo-mechanical properties. On the contrary, the nodes before the nozzle position were inactive, giving no contribution to the heat transfer process. Fig. 1. R Fig. 1. Rast aster geometry er geometry The print temperature varied between 190°C and 220°C with the platform surface being kept at 50°C. The printing temperature for the PLA should be higher than 190°C to ensure a fully melted filament, avoiding nozzle clogging, but not exceed 220°C to limit severe deformation during printing and polymer degradation caused by an excessively low viscosity. For this reason, Differential Scanning Calorimetry (DSC) tests were performed to determine the main thermal properties of the PLA filament. The measurements were carried out on a Chip-DSC 10 (Linseis Messgeraete GmbH, Selb, Germany), a heat Flux DSC with an integrated heater and temperature sensor. The main processing factors of the DSC analysis were the thermal cycle as well as the heating and cooling rates. Each thermal cycle consisted of a single run with a heating step from 20°C to 220°C, a holding step for 5 minutes and a cooling step from 220°C to 20°C for samples weighing about 8 mg, according to a well-known procedure [9]. The heating and cooling rates varied to investigate the polymorphic aspect of PLA by analyzing the melting process throughout fast heating and cooling, in very near fabrication conditions. Figure 2 shows the DSC measurement results for the different heating/cooling rates from 5 to 12.5°C/min. The glass transition, cold crystallization and melting peaks of the PLA shifted to higher temperatures with an increase in the heating rate. At the same time, the crystallization peak decreased with an increase in the cooling rate. These effects were also confirmed with heating and cooling rates greater than 100°C/min. These latter conditions were prevalent in real fabrication. 4641/3 Raster analysis of Fused Filament Fabrication process Figur Figure 2. DSC thermogr e 2. DSC thermograms of heating (left side) and cooling (rig ams of heating (left side) and cooling (right side). ht side). Figur Figure 2. DSC thermogr e 2. DSC thermograms of heating (left side) and cooling (rig ams of heating (left side) and cooling (right side). ht side). The print tests were carried out using a Funmat HT (Intamsys Technology Co. Ltd., Shanghai, China), a commercial FFF machine used to realize all samples in this experiment. The apparatus has a single extrusion nozzle mounted on a Cartesian axis system controlled by stepper motors. Polymer filaments were mounted on a spool and processed through a gear and pinch wheel system into a heated nozzle of 0.4 mm-diameter and extruded onto the heated platform. Fig. 1. R Fig. 1. Rast aster geometry er geometry From the experimental point of view, a particular device was used to study the geometry of molten polymer material by mounting glass slides onto a platform surface and depositing several lines. This deposition recreated the conditions of the initial layer, taking advantage of a difference in the density for the subsequent microscope scanning analysis, according to a procedure found in the literature [10]. Moreover, the variation in the shear rate along the nozzle diameter was reduced to prevent flow instabilities [11]. 3 Multi-ph 3 Multi-phyysical model sical model The main processing parameters of the FE model 4641/4 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 were deposition power and speed. The speed profile was computed with α equal to 0.5, following the rapid and deposition movements. The time step was small enough to accurately represent the nozzle continuous movement without excessively increasing the solution time for both deposition and cooling steps. Figure 3 shows the temperature fields computed with travel speeds equal to 5 and 50 mm/s, keeping the heat flux constant. The arrow represents the position of the nozzle during its travel. The temperature distribution at lower speed was higher than that at a higher speed because of longer deposition time. This time contributed to an increase in thermal gradients in the deposition path. The low conductivity of the polymer and the heat transfer coefficient between the glass plate and the polymer did not promote temperature diffusion. Figure 4 shows the stress distribution for a deposition speed equal to 50 mm/s. The warpage scale was magnified to highlight shape produced by the filament deposition and induced by the thermal gradient. As a result, deformations were observed at the start of the process and with a variation of the path continuity (acceleration and deceleration). This last effect caused the variation of the path thickness. Figur Figure 3. Simulation r e 3. Simulation results - Thermal field. esults - Thermal field. Figur Figure 3. Simulation r e 3. Simulation results - Thermal field. esults - Thermal field. 4641/5 Raster analysis of Fused Filament Fabrication process Raster analysis of Fused Filament Fabrication process Figur Figure 4. Simulation r e 4. Simulation results - Str esults - Stress field. ess field. Figur Figure 4. Simulation r e 4. Simulation results - Str esults - Stress field. ess field. 4 Experimental t 4 Experimental tests ests The following experimental activity was focused on the realization of several rasters with different print speed u and print temperature T. The lowest values of the print speed and temperature were 5 mm/s and 190°C while the highest values were 60 mm/s and 220°C. A multi-purpose zoom microscope system AZ100M (Nikon Instruments Europe BV, Amsterdam, Netherlands) was used to acquire the images and measure the 2D profiles. Figure 5 shows the measurements took in the same section of the rasters produced with different print speeds and temperatures. The initial segment of the raster was always incomplete because a time gap existed between the deposition start command and the real filament nozzle exit. The measurements of the 2D profiles pointed out the influence of the print temperature on the path width. As a result, an increase in the print temperature caused a decrease in the path width. The measurements were then repeated in all 13 raster sections, confirming the above results (Figure 6). 4641/6 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 Figur Figure 5. R e 5. Rast aster anal er analyysis. sis. Figur Figure 5. R e 5. Rast aster anal er analyysis. sis. 4641/7 Raster analysis of Fused Filament Fabrication process Figur Figure 6. Deposition width v e 6. Deposition width vs Segment s Segment.. Figur Figure 6. Deposition width v e 6. Deposition width vs Segment s Segment.. The average lengths for a print speed equal to 5 mm/s and print temperatures equal to 190°C and 220°C were respectively 1.28 and 1.11 mm. The maximum and minimum standard deviations were equal to 19.56 and 12.31 μm. Thus, the influence of the print temperature was significant since the difference of 0.17 mm between the lowest and highest values of the path width could not be neglected. The study was repeated for a print speed equal to 60 mm/s. The average values for print temperatures equal to 190°C and 220°C were respectively 1.23 and 1.14 mm. The maximum and minimum standard deviations were equal to 21.99 and 10.27 μm. The difference between the lowest and highest values was equal to 0.09 mm. The influence of the print temperature was important. As concern the contribution of the print speed, its increase caused a contraction in width band between the highest and lowest temperature. 4 Experimental t 4 Experimental tests ests A possible hypothesis of this behavior could be a decrease in polymer viscosity with a temperature and speed increase [14]. The rheometer measurements were made to identify the viscosity at the investigated temperatures. All experiments were performed on a HAAKE MARS III rotational rheometer (Thermo Fisher Scientific Inc, Waltham, USA), using nitrogen as a shielding gas to avoid polymer degradation and water as cooling media. Figure 7 reports the viscosity measurements, carried out with a shear rate ranging between 6×10-2 and 2×102 s-1. Shear rate values lower than 6×10-2 s-1 were not used because material degraded [11]. The differences in the viscosity values became smaller, with a shear rate increase. In general, the high viscosity of filament led to an under-extrusion (impossibility to supply the right amount of material), causing missing layers, thin layers, or layers with random dots and holes. For this reason, a low viscosity was highly desired, avoiding shallow values, causing over- extrusion (excess of deposited material). A 3D representation of the raster surface for each condition was then realized by selecting the in-focus area from several 4641/8 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 images with different focal points and producing one all-in-focus image with a scanning step of 5 μm. Such images provided a detailed 3D map for measuring the path height. Figure 8 shows the 3D topology maps of the rasters realized with a print speeds of 5 and 60 mm/s and a print temperature of 190°C and 220°C. Figur Figure 7. Viscosity curv e 7. Viscosity curves. es. Figur Figure 7. Viscosity curv e 7. Viscosity curves. es. 4641/9 Raster analysis of Fused Filament Fabrication process Figur Figure 8. 3D t e 8. 3D topology maps. opology maps. Figur Figure 8. 3D t e 8. 3D topology maps. opology maps. At a first analysis, the realistic display of the 3D surface topographies revealed that the raster produced with the lowest speed and the highest temperature led to the highest height values. On the contrary, the lowest height values were achieved with the lowest speed and temperature. These results were confirmed by the in-depth analysis of four consecutive paths, as Figure 9 reports, extracting 2D profiles from the previous 3D surface topographies. The measure of the average raster heights revealed a value of 0.729 mm for 5 mm/s and 220°C and 0.685 mm for 5 mm/s and 190°C. 4 Experimental t 4 Experimental tests ests Besides, the value of the standard deviation in both conditions was below 8.5 μm, establishing excellent stability during rather deposition. The results of raster width and heights also confirmed that an increase in the raster width caused a reduction in the raster height. Figure 10 shows this last condition by comparing results achieved with the lowest built time computed using the highest print speed equal to 60 mm/s, with the highest built time computed with the lowest print speed equal to 5 mm/s. The variations in width were more important than variations in height. The same figure reports tests carried out with a print speed of 30 mm/s and print temperatures of 190°C and 220°C. The results pointed out a good agreement between the predictions of the analytical model and these experimental tests, with a coefficient of determination R2 greater than 95% for all curves associated with raster height and width. 4641/10 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 Figur Figure 9. Deposition heig e 9. Deposition height v ht vs. Segment path. s. Segment path. Figur Figure 9. Deposition heig e 9. Deposition height v ht vs. Segment path. s. Segment path. 4641/11 4641/11 Raster analysis of Fused Filament Fabrication process Figur Figure 10. R e 10. Rast aster dimension v er dimension vs. Deposition time. s. Deposition time. Figur Figure 10. R e 10. Rast aster dimension v er dimension vs. Deposition time. s. Deposition time. Figur Figure 10. R e 10. Rast aster dimension v er dimension vs. Deposition time. s. Deposition time. 5 Conclusions 5 Conclusions The authors have investigated the raster generation to realize Fused Filament Fabrication parts, focusing on the deposition of the first raster. The geometrical variations have been caused by changes in process parameters such as the print speed and print temperature. The study has identified essential correlations between process parameters, raster dimensions, and filament properties. In particular, the speed variation on raster dimensions has been more important than the temperature variation. Moreover, the raster produced with the lowest speed and the highest temperature has led to the lowest width and the highest height values. On the contrary, the highest width and the lowest height values have been achieved with the lowest speed and temperature. The experimental procedure has been supported by an analytical model, able to compute raster time and dimensions with a coefficient of correlation higher than 95%. In future work, the experimental activities and the analytical model will be expanded to compute the raster deposition after the first layer and the evaluation of section dimensions. Bibliogr Bibliograph aphyy [1] Ngo TD et al. Additive manufacturing (3D printing): A review of materials, methods, applications and challenges. Composites Part B 2018; 143:172-196. [2] Thompson MK et al. Design for Additive Manufacturing: Trends, opportunities, considerations, and constraints. [2] Thompson MK et al. Design for Additive Manufacturing: Trends, opportunities, considerations, and constraints. 4641/12 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 ESAFORM 2021. MS13 (Additive Manufacturing), 10.25518/esaform21.4641 CIRP Ann. 2016;65:737-760. CIRP Ann. 2016;65:737-760. CIRP Ann. 2016;65:737-760. [3] Popescu D et al. FDM process parameters influence over the mechanical properties of T polymer specimens: A review. Polym. Test. 2018;69:157-166. [4] Koh JJ, Zhang X, He C. Fully biodegradable Poly(lactic acid)/Starch blends: A review of toughening strategies. Int. J. Biol. Macromol. 2018;109:99-113. [5] Comminal R, Serdeczny MP, Pedersen DB, Spangenberg J. Numerical modeling of the strand deposition flow in extrusion-based additive manufacturing, Addit. Manuf. 2018;20:68-76. [6] Serdeczny MP, Comminal R, Pedersen DB, Spangenberg J. Experimental validation of a numerical model for the strand shape in material extrusion additive manufacturing. Addit. Manuf. 2018;24:145-153. [7] Heller BP, Smith DE, Jack DA. Effects of extrudate swell and nozzle geometry on fiber orientation in Fused Filament Fabrication nozzle flow. Addit. Manuf.2016;12:252-264. [8] Osswald TA, Puentes J, Kattinger J. Fused filament fabrication melting model. Addit. Manuf. 2018;22:51-59. [9] Spina R, Technological characterization of PE/EVA blends for foam injection molding. Mat. Des. 2015;84:64-71. [10] Hebda M et al. A method for predicting geometric characteristics of polymer deposition during fused-filament- fabrication. Addit. Manuf. 2019;27:99-108. [11] Balani SB, Chabert F, NAssiet V, Cantarel A. Influence of printing parameters on the stability of deposited beads in fused filament fabrication of poly(lactic) acid. Addit. Manuf. 2019;25:112-121. [12] Luo Z, Zhao Y. A survey of finite element analysis of temperature and thermal stress fields in powder bed fusion Additive Manufacturing. Addit. Manuf. 2018;21:318-332. [13] Nain V, Carin M, Engel T, Boisselier D, Cordier C. A Comparative Study for Modelling Filler Material in Metal Additive Manufacturing Processes. COMSOL Conference, Cambridge (UK) 2019. [14] Spina R, Spekowius M, Hopmann C. Simulation of crystallization of isotactic polypropylene with different shear regimes. Thenochi. Acta 2018;659:44-54. PDF automatically generated on 2021-05-20 06:29:15 Article url: https://popups.uliege.be/esaform21/index.php?id=4641 published by ULiège Library in Open Access under the terms and conditions of the CC-BY License (https://creativecommons.org/licenses/by/4.0)
https://openalex.org/W2604557653	https://europepmc.org/articles/pmc5429759?pdf=render	English	null	A hybrid stochastic model of folate-mediated one-carbon metabolism: Effect of the common C677T MTHFR variant on de novo thymidylate biosynthesis	Scientific reports	2,017	cc-by	11,325	A hybrid stochastic model of folate-mediated one-carbon metabolism: Effect of the common C677T MTHFR variant on de novo thymidylate biosynthesis Received: 30 November 2016 Accepted: 13 March 2017 Published: xx xx xxxx Received: 30 November 2016 Accepted: 13 March 2017 Published: xx xx xxxx Karla Misselbeck1,2, Luca Marchetti 1, Martha S. Field3, Marco Scotti4, Corrado Priami1,2 & Patrick J. Stover3 Folate-mediated one-carbon metabolism (FOCM) is an interconnected network of metabolic pathways, including those required for the de novo synthesis of dTMP and purine nucleotides and for remethylation of homocysteine to methionine. Mouse models of folate-responsive neural tube defects (NTDs) indicate that impaired de novo thymidylate (dTMP) synthesis through changes in SHMT expression is causative in folate-responsive NTDs. We have created a hybrid computational model comprised of ordinary differential equations and stochastic simulation. We investigated whether the de novo dTMP synthesis pathway was sensitive to perturbations in FOCM that are known to be associated with human NTDs. This computational model shows that de novo dTMP synthesis is highly sensitive to the common MTHFR C677T polymorphism and that the effect of the polymorphism on FOCM is greater in folate deficiency. Computational simulations indicate that the MTHFR C677T polymorphism and folate deficiency interact to increase the stochastic behavior of the FOCM network, with the greatest instability observed for reactions catalyzed by serine hydroxymethyltransferase (SHMT). Furthermore, we show that de novo dTMP synthesis does not occur in the cytosol at rates sufficient for DNA replication, supporting empirical data indicating that impaired nuclear de novo dTMP synthesis results in uracil misincorporation into DNA. Perturbations in folate-mediated one-carbon metabolism (FOCM) are associated with numerous pathologies including neural tube defects (NTDs)1, stroke2, colorectal and other types of cancer3–5. Furthermore, enzymes in FOCM have been successful targets for the development of antineoplastic pharmaceutical agents including- 0020methotrexate and 5-flurouracil6. FOCM in the cytoplasm is composed of three interconnected biosynthetic pathways, which include de novo thymidylate (dTMP) synthesis, de novo purine synthesis and homocysteine remethylation to methionine (Fig. 1). The FOCM network is sensitive to nutritional status for several vitamins that serve as enzyme cofactors (folate, riboflavin, vitamin B6 and vitamin B12) and genetic factors (coding and expression variants in folate-dependent enzymes) that can alter network outputs, including DNA synthesis, DNA repair and chromatin methylation7–9. Understanding the molecular basis of disease etiology has been limited by the ability to ascribe specific FOCM pathways and their biomarkers to clinical outcomes, because the path- ways of FOCM are tightly interconnected8. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Received: 30 November 2016 Accepted: 13 March 2017 Published: xx xx xxxx A hybrid stochastic model of folate-mediated one-carbon metabolism: Effect of the common C677T MTHFR variant on de novo thymidylate biosynthesis The overall goal is to provide an understanding of function of the entire system in silico that can be used to accelerate discovery and guide the design of biological experimentation.i Early models of FOCM quantified steady-state concentrations of folates, and predicted how methotrexate and 5-fluorouracil affect the rates of de novo purine and thymidylate biosynthesis in L1210 cells12–15. These models considered FOCM reactions as occurring in a common cellular compartment, described the reactions in terms of Michaelis-Menten kinetics, and investigated biological consequences of inhibiting de novo purine and dTMP biosynthesis. Other models have been constructed to simulate the homocysteine remethylation in liver16–18. Using data from experimental animals, non-trivial mathematical functions were adopted to address long-range inhibition and activation processes involving folates and other metabolites that cannot be translated into simple Michaelis-Menten equations17. These models can reproduce changes in metabolite concentrations in response to nutritional deficiencies and the effects of gene variants, thus corroborating human clinical and epidemiological data. These models have also integrated other pathways such as glutathione metabolism and polyamine synthesis, studying the influence of indirect effects beyond FOCM19, 20.hl lf These existing FOCM models describe the continuous flux of metabolite concentrations in terms of ordinary differential equations (ODEs) with time as an independent variable13, 20. Sensitivity analysis is used to summa- rize the effects on metabolite concentrations and reaction velocities in response to changes in inputs or enzyme activities21. Here we present a hybrid stochastic model for simulating the FOCM dynamics where state-of-the art deterministic simulation (based on ODEs) has been coupled with exact stochastic simulation to assess metabolite variabilities in the FOCM network at steady state. A detailed description of this combined approach can be found in Methods. The deterministic approach used in isolation can only provide a rough estimate of FOCM model dynamics, because the deterministic approach is limited when enzyme substrates, such as folate cofactors, are present in the cells at low micromolar concentrations, and because reactions within the network occur randomly at discrete time points. FOCM is expected to exhibit variability (i.e. stochasticity) in its behavior22. Capturing system stochasticity is essential when substrate concentrations are low and limiting, but requires consideration of molecules as discrete entities, rather than describing concentrations as continuous variables through ODEs23. Simulation strategies that combine both deterministic and stochastic approaches can give a more accurate and more detailed understanding of FOCM network functioning and stability. A hybrid stochastic model of folate-mediated one-carbon metabolism: Effect of the common C677T MTHFR variant on de novo thymidylate biosynthesis FOCM complexity is manifest by: (a) competition among the path- ways for a limiting pool of folate cofactors10, (b) long-range and indirect regulatory processes, (c) formation of multi-enzyme complexes, (d) cellular compartmentation, (e) interactions with other metabolic pathways, (f) nutritional status (g) penetrant genetic variants9, 11. Mathematical models have been developed to assess this 1The Microsoft Research - University of Trento Centre for Computational and Systems Biology (COSBI), Piazza Manifattura, 1, 38068, Rovereto (TN), Italy. 2Department of Mathematics, University of Trento, Trento, Italy. 3Division of Nutritional Sciences, Cornell University, Ithaca, New York, 14853, USA. 4GEOMAR Helmholtz Centre for Ocean Research Kiel, Düsternbrooker Weg 20, 24105, Kiel, Germany. Karla Misselbeck and Luca Marchetti contributed equally to this work. Corrado Priami and Patrick J. Stover jointly supervised this work. Correspondence and requests for materials should be addressed to C.P. (email: priami@cosbi.eu) or P.J.S. (email: pjs13@cornell.edu) Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 1 www.nature.com/scientificreports/ Figure 1. The reaction-based specification of the model according to the notation introduced in Gostner et al.61. Rectangles identify model variables, non-boxed substrates are model constants, green circles identify enzymes, dark blue arcs identify matter transformation, and light blue arcs identify regulatory events (dotted lines indicate activations and solid lines indicate inhibitions). The purple boxes indicate reactions and variables associated with the folate cycle and the homocysteine remethylation cycle, respectively. Figure 1. The reaction-based specification of the model according to the notation introduced in Gostner et al.61. Rectangles identify model variables, non-boxed substrates are model constants, green circles identify enzymes, dark blue arcs identify matter transformation, and light blue arcs identify regulatory events (dotted lines indicate activations and solid lines indicate inhibitions). The purple boxes indicate reactions and variables associated with the folate cycle and the homocysteine remethylation cycle, respectively. Figure 1. The reaction-based specification of the model according to the notation introduced in Gostner et al.61. Rectangles identify model variables, non-boxed substrates are model constants, green circles identify enzymes, dark blue arcs identify matter transformation, and light blue arcs identify regulatory events (dotted lines indicate activations and solid lines indicate inhibitions). The purple boxes indicate reactions and variables associated with the folate cycle and the homocysteine remethylation cycle, respectively. complexity and gain an understanding of the cause-and-effect relationships that regulate FOCM functioning in health and disease. www.nature.com/scientificreports/ investigated the consequences of the MTHFR C677T polymorphism, assuming 70% enzyme activity for hete- rozygote and 30% enzyme activity for homozygote, in comparison to CC homozygotes (which was set to 100% activity) using parameters for folate monoglutamates, which are not the physiological form of folate cofactors in cells. Under these conditions, the variant allele decreased concentrations of 5mTHF and SAM and increased the concentrations of homocysteine, SAH, and rates of dTMP and purine biosynthesis21, which is inconsistent with current understanding of biochemical changes associated with NTD risk. The effect on the redistribution of folate cofactors towards 10fTHF that is associated with the 677 T variant, or its impact on other pathways within the network, was not reported21. activity by decreasing its affinity for the FADH cofactor. Such substitution affects enzyme stability and hence the partitioning of folates between dTMP synthesis and homocysteine remethylation25, 26. Decreased MTHFR activity resulting from the polymorphism decreases 5mTHF synthesis, leading to impaired homocysteine remeth- ylation and elevated serum homocysteine24. The 677 T variant is also associated with a redistribution of cellular folate cofactors; 5mTHF is the predominate form of folate in red blood cells in MTHFR 677CC carriers, whereas 10fTHF is the predominate form of folate in MTHFR 677TT carriers27–29. 10fTHF is less chemically stable than 5mTHF, and the MTHFR 677TT variant is associated with lower folate status30 and higher folate requirement31. Recent studies suggest that the contribution of the MTHFR variant to NTD risk is due to its impact on cellular folate status, rather than impaired homocysteine remethylation32. Likewise, mouse models of NTDs indicate that impaired dTMP synthesis, and not homocysteine remethylation, cause folate-responsive NTDs33–35. Reed et al. investigated the consequences of the MTHFR C677T polymorphism, assuming 70% enzyme activity for hete- rozygote and 30% enzyme activity for homozygote, in comparison to CC homozygotes (which was set to 100% activity) using parameters for folate monoglutamates, which are not the physiological form of folate cofactors in cells. Under these conditions, the variant allele decreased concentrations of 5mTHF and SAM and increased the concentrations of homocysteine, SAH, and rates of dTMP and purine biosynthesis21, which is inconsistent with current understanding of biochemical changes associated with NTD risk. The effect on the redistribution of folate cofactors towards 10fTHF that is associated with the 677 T variant, or its impact on other pathways within the network, was not reported21. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 2. Steady states for five different values of glycine (folate cycle in % and the homocysteine remethylation cycle in μM). Where possible, a trend arrow is provided on the right to show the experimental outcome observed in Herbig et al.28. All the trends are consistent with literature (green arrows) except for the total % of 5 mTHF in the glycine range 5–10 mM (red arrow). Figure 2. Steady states for five different values of glycine (folate cycle in % and the homocysteine remethylation cycle in μM). Where possible, a trend arrow is provided on the right to show the experimental outcome observed in Herbig et al.28. All the trends are consistent with literature (green arrows) except for the total % of 5 mTHF in the glycine range 5–10 mM (red arrow). Figure 2. Steady states for five different values of glycine (folate cycle in % and the homocysteine remethylation cycle in μM). Where possible, a trend arrow is provided on the right to show the experimental outcome observed in Herbig et al.28. All the trends are consistent with literature (green arrows) except for the total % of 5 mTHF in the glycine range 5–10 mM (red arrow). igure 2. Steady states for five different values of glycine (folate cycle in % and the homocysteine remethylation ycle in μM). Where possible, a trend arrow is provided on the right to show the experimental outcome activity by decreasing its affinity for the FADH cofactor. Such substitution affects enzyme stability and hence the partitioning of folates between dTMP synthesis and homocysteine remethylation25, 26. Decreased MTHFR activity resulting from the polymorphism decreases 5mTHF synthesis, leading to impaired homocysteine remeth- ylation and elevated serum homocysteine24. The 677 T variant is also associated with a redistribution of cellular folate cofactors; 5mTHF is the predominate form of folate in red blood cells in MTHFR 677CC carriers, whereas 10fTHF is the predominate form of folate in MTHFR 677TT carriers27–29. 10fTHF is less chemically stable than 5mTHF, and the MTHFR 677TT variant is associated with lower folate status30 and higher folate requirement31. Recent studies suggest that the contribution of the MTHFR variant to NTD risk is due to its impact on cellular folate status, rather than impaired homocysteine remethylation32. Likewise, mouse models of NTDs indicate that impaired dTMP synthesis, and not homocysteine remethylation, cause folate-responsive NTDs33–35. Reed et al. www.nature.com/scientificreports/ p Here, we studied the partitioning of CH2F, a cofactor for both homocysteine remethylation and de novo dTMP biosynthesis36–38, and the effects of known genetic and nutritional variables that impact movement of CH2F through the network. Existing models are limited by adopting kinetic parameters determined from the use of folate monoglutamate substrates19, 20. Folate polyglutamates are the functional form of folate cofactors in cells and have much higher affinity for their respective FOCM enzymes than the corresponding monoglu- tamate forms of folate39, 40. Therefore, we updated the parameters in the deterministic model to include the physiologically-relevant polyglutamate forms of folate cofactors and demonstrate that it faithfully recapitulates existing data in the literature (see Methods, Model Validation and Fig. 2). The decisions for selecting individual enzyme kinetic parameters for this model were driven by: 1) available data for physiologically relevant polygluta- mate forms of the folate cofactors derived from characterization of mammalian enzymes, and 2) data from human models, specifically L1210 cells, because of the richness and quality of the data used to derive kinetic parameters. Given the high conservation of folate enzymes among mammals, our model could be applied to mammalian systems in general, even though we are not proposing a completely homogeneous model with respect to species. y g g p p g p y g p p We were able to identify key nodes in the network of Fig. 1 that exhibit high degrees of stochastic behavior, including the influence of nutrient status and genetic variation on stochasticity through simulations. We explored the impact of the MTHFR C677T polymorphism and its interaction with folate status on partitioning of CH2F within the network, including its impact on de novo dTMP biosynthesis to understand the etiology of NTDs. The results of the computational model provide evidence that the rates of de novo dTMP synthesis as currently modeled in the cytosol are insufficient to support DNA synthesis in S-phase in mammals, accounting for uracil misincorporation into DNA that occurs in folate deficiency and in mouse models of NTDs. A hybrid stochastic model of folate-mediated one-carbon metabolism: Effect of the common C677T MTHFR variant on de novo thymidylate biosynthesis In contrast to approaches based solely on deterministic simulation, these studies can be used to assess the contributions of factors such as genetic vari- ation and nutritional status on the stochastic behavior of individual pathways within the network, thereby aiding in establishing which system inputs (i.e. nutrition) and outputs (i.e. biomarkers) are most closely associated with human health outcomes. Dysregulation of the partitioning of one-carbon units in the form of 5,10-methylenetetrahydrofolate (CH2F) cofactors between the de novo dTMP biosynthesis and homocysteine remethylation pathways is believed to underlie FOCM-associated pathologies including NTDs (Fig. 1). A common variant of MTHFR, the C677T poly- morphism, has been associated with numerous pathologies including birth defects, cancer, cardiovascular events, and other pathologies24. MTHFR catalyzes the FADH-dependent, irreversible conversion of CH2F to 5mTHF, which commits folate cofactors away from dTMP synthesis and towards homocysteine remethylation in the cytosol (Fig. 1). The variant results from an alanine to valine substitution in the protein that decreases MTHFR Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 2 Results Th ff Concentrations of model variables (in µM) for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. Table 2. Concentrations of model variables (in µM) for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. yp p y p µM/h FTS MTCH MTD MTHFR MTR Unbinding of SHMT and 5mTHF 10fTHF → CHF CHF → 10fTHF CHF → CH2F CH2F → CHF MTHFR x2 13,101.4 332,703.8 330,313.4 37,733.2 35,342.8 26.8 26.8 17,035.2 MTHFR x1 - CC 23,582.5 756,270.3 744,706.4 89,938.7 78,374.9 21.6 21.6 14,964.5 MTHFR x0.7 - CT 31,430.7 884,173.6 864,936.0 106,222.3 86,984.7 16.0 16.0 12,231.1 MTHFR x0.5 36,054.6 954,589.4 930,806.0 115,378.8 91,595.3 11.7 11.7 9,754.6 MTHFR x0.3 - TT 39,755.7 1,002,680.1 975,243.2 121,653.1 94,216.2 7.1 7.1 6,559.7 Table 3. Fluxes of the reactions catalyzed by the enzymes FTS, MTCH, MTD, MTHFR and MTR and the binding of 5mTHF and SHMT for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. Table 3. Fluxes of the reactions catalyzed by the enzymes FTS, MTCH, MTD, MTHFR and MTR and the binding of 5mTHF and SHMT for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. SAM and SAH levels vary markedly with changes in MTHFR activity (SAM levels decrease and SAH levels increase by more than 50% when comparing the “CC” model to the “TT” model), with a SAM/SAH ratio around one being achieved in CC homozygotes compared to 0.24 in TT homozygotes (Table 2). Although methylation potential, otherwise known as the SAM/SAH ratio, changes markedly with varying MTHFR activity, homocyst- eine concentrations appear relatively insensitive to changes in MTHFR in this model (Table 2), inconsistent with known effects of the MTHFR TT genotype in elevating serum homocysteine levels41. Results Th ff Results The effect of folate status and the MTHFR polymorphism on pathways affecting NTD risk. In the current model, MTHFR activity was decreased to model the effect of the MTHFR C677T polymorphism. In addition, a two-fold increase in MTHFR activity was modeled to examine whether there was a dose-response relationship between MTHFR activity and various readouts of the network. This model shows that 5mTHF levels decrease as MTHFR activity decreases, reflecting the effects of the MTHFR C677T polymorphism (Table 1). The current model also recapitulates the biological observation that decreased MTHFR activity results in accumu- lation of 10fTHF and THF (Table 1). Inclusion of 10fTHF in the folate distribution is a strength of the current model in that it allows for estimation of the effect that perturbations to the system have on accumulation of this unstable form of folate, which likely accounts for the decreased folate status linked to NTDs in carriers of the polymorphism30, 32. Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 3 www.nature.com/scientificreports/ % THF 10fTHF CHF CH2F DHF 5mTHF free bound total MTHFR x2 0.25 14.43 2.86 0.79 0.01 33.45 48.20 81.65 MTHFR x1 – CC 0.70 39.24 7.69 1.93 0.04 8.07 42.34 50.42 MTHFR x0.7 – CT 1.59 48.76 9.45 2.18 0.04 3.37 34.61 37.98 MTHFR x0.5 3.13 54.53 10.51 2.32 0.04 1.86 27.60 29.46 MTHFR x0.3 – TT 8.12 58.72 11.26 2.40 0.04 0.90 18.56 19.46 Table 1. Distribution of folate (in percentage of total folate) for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C677T polymorphism. Table 1. Distribution of folate (in percentage of total folate) for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C677T polymorphism. µM THF 10fTHF CHF CH2F DHF 5 mTHF Unbound SHMT HCY MET SAM SAH free bound total MTHFR x2 0.04 2.58 0.51 0.14 0.00 5.97 8.60 14.57 0.40 3.10 37.99 78.41 31.01 MTHFR x1 – CC 0.12 7.00 1.37 0.34 0.01 1.44 7.56 9.00 1.44 3.47 42.44 51.34 53.26 MTHFR x0.7 – CT 0.28 8.70 1.69 0.39 0.01 0.60 6.18 6.78 2.82 3.67 43.00 34.87 68.98 MTHFR x0.5 0.56 9.73 1.88 0.41 0.01 0.33 4.93 5.26 4.07 3.78 42.59 26.72 77.42 MTHFR x0.3 – TT 1.45 10.48 2.01 0.43 0.01 0.16 3.31 3.47 5.69 3.89 41.87 20.47 84.29 Table 2. Results Th ff µM/h BHMT MAT-I MAT- III GNMT DNMT SAHH SAH → HCY HCY → SAH MTHFR x2 145.9 110.3 62.4 119.6 127.4 264.6 91.9 MTHFR x1 - CC 161.0 123.0 59.6 370.9 87.2 285.2 102.6 MTHFR x0.7 - CT 168.3 127.9 56.3 593.3 59.6 292.4 108.2 MTHFR x0.5 172.1 129.3 54.4 723.1 45.6 295.2 111.5 MTHFR x0.3 - TT 175.5 129.8 52.7 830.4 34.7 297.1 114.5 Table 5. Fluxes of the reactions catalyzed by the enzymes MTR, BHMT, MAT-I, MAT-III, GNMT, DNMT and SAHH for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. Table 5. Fluxes of the reactions catalyzed by the enzymes MTR, BHMT, MAT-I, MAT-III, GNMT, DNMT and SAHH for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. % THF 10fTHF CHF CH2F DHF 5 mTHF free bound total low folate CC 1.27 38.48 7.46 1.65 0.03 4.98 46.13 51.11 TT 4.24 49.75 9.45 1.68 0.03 0.93 33.91 34.84 replete folate CC 0.70 39.24 7.69 1.93 0.04 8.07 42.34 50.42 TT 8.12 58.72 11.26 2.40 0.04 0.90 18.56 19.46 Table 6. Distribution of folate (in percentage of total folate) for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C667T MTHFR polymorphism. Table 6. Distribution of folate (in percentage of total folate) for replete (18 µM) and low (9 µM) levels of total olate and for the CC and TT case of the C667T MTHFR polymorphism. 5mTHF binds to and is a potent inhibitor of SHMT and glycine N-methyltransferase (GNMT). Decreased MTHFR activity, which lowers cellular 5mTHF levels (Table 2), increases the flux through the SHMT1-catalyzed reaction in the direction of serine catabolism to glycine (Table 4). This reflects the depletion of intracellular 5mTHF as well as glycine, resulting from increased GNMT activity that results from 5mTHF depletion (Table 5). Methionine synthase (MTR) flux is highly sensitive to MTHFR genotype reflecting its dependence on availability of its substrate 5mTHF that is generated by MTHFR (Table 3).hf g y The MTHFR C677T variant affects both CH2F partitioning (between homocysteine remethylation and dTMP biosynthesis) and intracellular folate concentrations; the 677 T variant lowers intracellular folate levels. Results Th ff However, in this model the lack of elevation in homocysteine due to the MTHFR C677T polymorphism reflects that the model represents a closed system leading to intracellular conversion of cellular homocysteine to SAH as opposed to export of homo- cysteine into the circulation (Table 2).h y The activity of each enzyme in the FOCM network as predicted by the current model indicates that accumula- tion of 10fTHF resulting from decreased MTHFR activity is due to two factors: (1) an increased flux through both the 10fTHF synthetase activity leading to increased synthesis (FTS), and (2) an increased flux through the cyclo- hydrolase/dehydrogenase activity of MTHFD1 which converts CH2F to CHF to 10fTHF (MTCH, MTD activ- ities, respectively, Table 3). Interestingly, the accumulation of 10fTHF does not affect flux through the enzymes that use 10fTHF as a co-factor for de novo purine synthesis (PGT, AICART; Table 4). This is consistent with empirical experimental findings that increasing cellular levels of 10-formytetrahydrofolate dehydrogenase, which consumes 10fTHF, do not affect de novo purine synthesis42. Flux through the de novo dTMP synthesis pathway increases with decreasing MTHFR activity, consistent with empirical studies indicating these two pathways com- pete for CH2F (Table 4, columns DHFR, TYMS and SHMT)43. These findings are in agreement with empirical data showing that the TT polymorphism results in an increase in CH2F available for dTMP synthesis as indicated by isotope tracer studies in humans26. Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 4 www.nature.com/scientificreports/ µM/h PGT AICARFT DHFR TYMS SHMT Binding of SHMT and 5 mTHF CH2F → THF THF → CH2F MTHFR x2 4,406.3 6,304.6 113.5 113.5 2819.8 569.7 17,035.2 MTHFR x1 - CC 5,268.8 6,749.8 263.4 263.4 15,646.6 4,367.7 14,964.5 MTHFR x0.7 - CT 5,388.7 6,804.3 295.0 295.0 31,989.8 13,063.2 12,231.1 MTHFR x0.5 5,442.7 6,828.4 312.3 312.3 47,127.3 23,667.8 9,754.6 MTHFR x0.3 - TT 5,475.7 6,843.0 322.1 322.1 66,523.2 39,415.5 6,559.7 Table 4. Fluxes of the reactions catalyzed by the enzymes PGT, AICARFT, DHFR, TYMS, SHMT and the unbinding of 5mTHF and SHMT for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. Table 4. Fluxes of the reactions catalyzed by the enzymes PGT, AICARFT, DHFR, TYMS, SHMT and the unbinding of 5mTHF and SHMT for different levels of MTHFR activity (ranging from 2x to 0.3x of wild type). CC, CT and TT refer to the C667T polymorphism. Results Th ff Therefore, the impact of this variant on FOCM was modeled at two different levels of folate (Tables 6–10) (folate replete conditions, 18 μM, and low-folate conditions, 9 μM) to understand how the variants function within the FOCM network as a function of folate cofactor availability. The results demonstrate that the changes in the percentage of 5mTHF (as well as other major one-carbon forms of folate) are more pronounced in the TT genotype than in the CC genotype when cellular folate levels are decreased (Table 6). The percentage of 5 mTHF in CC homozygous does not change in the folate replete and deficiency states, whereas the accumulation of 5 mTHF in TT homozy- gotes differs between the deficient and replete states (Table 6).f gfi p Fluxes through FOCM pathways are affected by both the MTHFR C677T polymorphism and folate levels. The most sensitive pathways to folate deficiency are the MTCH and MTD activities of MTHFD1, MTHFR, MTR, DNMT, DHFR, and TYMS (Tables 8–10, row showing absolute flux differences between folate levels). Flux through the dTMP synthesis pathway (DHFR and TYMS) is highly sensitive to folate status for both the MTHFR CC and TT genotypes (Table 9), with the TT homozygotes being the most sensitive. Flux through GNMT was Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 5 www.nature.com/scientificreports/ μM THF 10fTHF CHF CH2F DHF 5 mTHF Free SHMT HCY MET SAM SAH free bound total low folate CC 0.11 3.43 0.67 0.15 0.00 0.44 4.12 4.56 2.55 3.73 42.84 30.34 73.61 TT 0.38 4.44 0.84 0.15 0.00 0.08 3.03 3.11 10.00 3.95 41.35 17.43 87.79 replete folate CC 0.12 7.00 1.37 0.34 0.01 1.44 7.56 9.00 1.44 3.47 42.44 51.34 53.26 TT 1.45 10.48 2.01 0.43 0.01 0.16 3.31 3.47 5.69 3.89 41.87 20.47 84.29 Table 7. Concentrations of model variables (in µM) for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C667T MTHFR polymorphism. Table 7. Concentrations of model variables (in µM) for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C667T MTHFR polymorphism. Results Th ff µM/h FTS MTCH MTD MTHFR MTR 10fTHF → CHF CHF → 10fTHF CHF → CH2F CH2F → CHF CC replete folate 23,582.5 756,270.3 744,706.4 89,938.7 78,374.9 21.6 21.6 low folate 22,583.6 427,359.0 415,960.1 48,054.6 36,655.7 13.8 13.8 absolute difference (% of replete folate) 4.2 43.5 44.1 46.6 53.2 36.5 36.5 TT replete folate 39,755.7 1,002,680.1 975,243.2 121,653.1 94,216.2 7.1 7.1 low folate 33,618.2 529,771.1 507,690.2 59,371.2 37,290.2 4.2 4.2 absolute difference (% of replete folate) 15.4 47.2 47.9 51.2 60.4 41.2 41.2 Table 8. Fluxes of the reactions catalyzed by the enzymes FTS, MTCH, MTD, MTHFR and MTR for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. Table 8. Fluxes of the reactions catalyzed by the enzymes FTS, MTCH, MTD, MTHFR and MTR for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. µM/h PGT AICARFT DHFR TYMS SHMT Binding of 5 mTHF and SHMT Unbinding of 5 mTHF and SHMT CH2F → THF THF → CH2F CC replete folate 5,268.8 6,749.8 263.4 263.4 15,646.6 4,367.7 14,964.5 1,4964.5 low folate 4,711.3 6,473.4 117.8 117.8 18,535.4 7,268.2 8150.7 8150.7 absolute difference (% of replete folate) 10.6 4.1 55.3 55.3 18.5 66.4 45.5 45.5 TT replete folate 5,475.7 6,843.0 322.1 322.1 66,523.2 39,415.5 6,559.7 6,559.7 low folate 4,943.9 6,593.4 120.0 120.0 73,586.0 51,629.1 5,992.1 5,992.1 absolute difference (% of replete folate) 9.7 3.6 62.8 62.8 10.6 31.0 8.7 8.7 Table 9. Fluxes of the reactions catalyzed by the enzymes PGT, AICARFT, DHFR, TYMS and SHMT and the binding/unbinding of 5 mTHF and SHMT for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. Table 9. Fluxes of the reactions catalyzed by the enzymes PGT, AICARFT, DHFR, TYMS and SHMT and the binding/unbinding of 5 mTHF and SHMT for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. Table 9. Results Th ff Fluxes of the reactions catalyzed by the enzymes PGT, AICARFT, DHFR, TYMS and SHMT and the binding/unbinding of 5 mTHF and SHMT for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. also highly sensitive to folate status in the CC homozygotes, whereas GNMT flux in TT homozygotes was insen- sitive to folate status (Table 10). Similar but less pronounced effects were seen for flux through SHMT (Table 9).f also highly sensitive to folate status in the CC homozygotes, whereas GNMT flux in TT homozygotes was insen- sitive to folate status (Table 10). Similar but less pronounced effects were seen for flux through SHMT (Table 9). To understand if MTHFR genotype affects the stability of the FOCM network at steady state, the determin- istic simulation was coupled with stochastic simulation using a hybrid simulation strategy (see Methods and Supplementary Material). Model steady states were obtained under four different conditions that differed by MTHFR 677 genotype (CC and TT case) and intracellular folate levels (replete and low). Interestingly, the most stable steady state (the one with lowest total sum of reaction propensities a0(x), see Methods for details), was the CC case with folate replete concentrations, consistent with numerous epidemiological studies associating the MTHFR C677T genotype with folate-related pathologies (Table 11)43. The enzyme that exhibited the greatest level of stochasticity in response to folate levels and/or the MTHFR C677T polymorphism was SHMT1 (Table S6). The molecular basis of uracil misincorporation into DNA. Mouse models implicate SHMT1 and impaired de novo dTMP synthesis in NTD risk. Impaired de novo dTMP synthesis causes an increase in dUMP, which when converted to dUTP causes uracil misincorporation into DNA because DNA polymerases do not dis- tinguish between dTTP and dUTP44. The dTMP biosynthesis pathway enzymes (MTHFD1, SHMT, TYMS, and DHFR) are present in both the cytosol and recently have been found to function in the nucleus. In the nucleus, they comprise a multi-enzyme complex at sites of DNA synthesis that may be critical to limit rates of uracil mis- incorporation into DNA, but regulatory mechanisms remain unknown45. These enzymes are modified by the Small Ubiquitin-like MOdifier (SUMO) protein at the G1/S boundary, which permits their nuclear translocation during S-phase of the cell cycle46. Results Th ff Total propensities obtained in four steady state conditions according to folate polymorphism (CC and TT case) and total concentration of available folate (replete, 18 µM; low, 9 µM). To help comparisons, the differences between CC and TT (in % of CC) and between replete and low total folate (in % of replete folate) are indicated. A steady state that is less stable (or more noisy) than another one has higher total propensity. Table 11. Total propensities obtained in four steady state conditions according to folate polymorphism (CC and TT case) and total concentration of available folate (replete, 18 µM; low, 9 µM). To help comparisons, the differences between CC and TT (in % of CC) and between replete and low total folate (in % of replete folate) are indicated. A steady state that is less stable (or more noisy) than another one has higher total propensity. in a mouse model over-expressing SHMT1, rates of uracil misincorporation into DNA increased several fold45. In this model, SHMT1 protein levels were elevated several fold in the liver, yet its localization was restricted to the cytoplasm and nuclear SHMT1 levels were depleted compared to wild-type mice45. Furthermore, nuclei isolated from SHMT1 overexpressing mice exhibited lower rates of de novo dTMP synthesis compared to nuclei isolated from wild-type mice45. This suggests that de novo dTMP synthesis occurs when the enzymes are present in the multi-enzyme complex within the nucleus in mammals. However, no definitive experiment has been performed that identifies the relative contribution of nuclear and cytosolic dTMP synthesis to overall dTMP synthesis. Interestingly, S. cerevesiae do not import the dTMP synthesis pathway into the nucleus47. in a mouse model over-expressing SHMT1, rates of uracil misincorporation into DNA increased several fold45. In this model, SHMT1 protein levels were elevated several fold in the liver, yet its localization was restricted to the cytoplasm and nuclear SHMT1 levels were depleted compared to wild-type mice45. Furthermore, nuclei isolated from SHMT1 overexpressing mice exhibited lower rates of de novo dTMP synthesis compared to nuclei isolated from wild-type mice45. This suggests that de novo dTMP synthesis occurs when the enzymes are present in the multi-enzyme complex within the nucleus in mammals. However, no definitive experiment has been performed that identifies the relative contribution of nuclear and cytosolic dTMP synthesis to overall dTMP synthesis. Interestingly, S. cerevesiae do not import the dTMP synthesis pathway into the nucleus47. Results Th ff g y p y p y To determine if nuclear import of the de novo dTMP pathway was required to meet cellular demands for dTTP during DNA replication, rates of dTMP synthesis were modeled for mammalian cells using standard Michaelis-Menten kinetics (Table 12; Table 4). Based on the number of A-T base pairs in the human genome and an 8-hour S-phase in embryonic stem cells (S-phase in L1210 cells is also 6–10 h48, 49), the rate of dTMP synthesis required for faithful cell replication is 7.8 µM/min (calculations are in Supplementary Material, see also Table 12)50, 51. In the current model, which does not account for SHMT1/TYMS/DHFR/MTHFD1 nuclear localization nor complex formation, cytosolic dTMP synthesis rates are 4.4 µM/min (Table 4, DHFR and TYMS flux, 263.4 µM/h, assuming MTHFR 677 CC genotype). This computational deficit between dTMP requirements and dTMP synthesis rates suggests that dTMP synthesis as currently modeled in the cytosol where the enzymes are not present in a complex cannot meet cellular needs. Nuclear localization and complex formation of the de novo dTMP synthesis complex seem to be unique to mammalian cells. In S. cerevesiae, TYMS is not SUMOylated and localizes to the nuclear periphery47. The measured rate of dTMP synthesis in S. cerevesiae is 1.8 µM/min52 (Table 12). The rate of dTMP synthesis required to replicate the S. Cerevesiae genome over the course of an S-phase (less than one hour52, 53) is 0.5 µM/min, indicating that yeast synthesize dTMP at a rate that is more than 3-fold greater than necessary for adequate dTMP synthesis (calculations are in Supplementary Material). Furthermore, in response to DNA damage, yeast increase dNTP concentrations 6–8 fold54 and E. coli increase dNTP concentra- tions 1.8–3.7 fold55, but dNTP concentrations do not increase after DNA damage in mammals56, 57. Results Th ff One study showed that when nuclear translocation of this complex is impaired Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 6 www.nature.com/scientificreports/ µM/h BHMT MAT-I MAT-III GNMT DNMT SAHH SAH → HCY HCY → SAH CC replete folate 161.0 123.0 59.6 370.9 87.2 285.2 102.6 low folate 170.3 128.8 55.3 664.0 51.8 294.0 110.0 absolute difference (% of replete folate) 5.8 4.7 7.3 79.0 40.5 3.1 7.2 TT replete folate 175.5 129.9 52.7 830.4 34.7 297.1 114.5 low folate 177.4 129.8 51.8 884.4 29.4 297.9 116.3 absolute difference (% of replete folate) 1.1 0.0 1.8 6.5 15.2 0.3 1.6 Table 10. Fluxes of the reactions catalyzed by the enzymes BHMT, MAT-I, MAT-III, GNMT, DNMT and SAHH for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. Table 10. Fluxes of the reactions catalyzed by the enzymes BHMT, MAT-I, MAT-III, GNMT, DNMT and SAHH for replete (18 µM) and low (9 µM) levels of total folate and for the CC and TT case of the C677T MTHFR polymorphism. a0(x) replete folate low folate Difference (% of replete folate) CC 1.6054 · 1015 2.0970 · 1015 30.62 TT 8.5915 · 1015 1.0141 · 1016 18.03 Difference (% of CC) 435.15 383.59 Table 11. Total propensities obtained in four steady state conditions according to folate polymorphism (CC and TT case) and total concentration of available folate (replete, 18 µM; low, 9 µM). To help comparisons, the differences between CC and TT (in % of CC) and between replete and low total folate (in % of replete folate) are indicated. A steady state that is less stable (or more noisy) than another one has higher total propensity. a0(x) replete folate low folate Difference (% of replete folate) CC 1.6054 · 1015 2.0970 · 1015 30.62 TT 8.5915 · 1015 1.0141 · 1016 18.03 Difference (% of CC) 435.15 383.59 Table 11. Total propensities obtained in four steady state conditions according to folate polymorphism (CC and TT case) and total concentration of available folate (replete, 18 µM; low, 9 µM). To help comparisons, the differences between CC and TT (in % of CC) and between replete and low total folate (in % of replete folate) are indicated. A steady state that is less stable (or more noisy) than another one has higher total propensity. Table 11. www.nature.com/scientificreports/ In the cell, newly transported monoglutamate folates are converted to folate polyglutamates, containing 3 to 7 polyglutamate moieties, though the action of folylpolyglutamate synthetase58. The polyglutamate chain (N = 3 glutamate and higher) increases the affinity of folate cofactors for many folate-dependent enzymes by one to two orders of magnitude39, 40. Models that include kinetic parameters derived from the use of folate monoglu- tamates can limit model reliability. Here we established a hierarchy of criteria to select a more homogeneous set of kinetic parameters (i.e. Km and Vmax) by referring, when possible, to L1210 cells because of the richness and quality of the data used to derive kinetic parameters. Furthermore, our preference was to select kinetic coefficients generated using folate polyglutamate cofactors and purified proteins from animal models closest to humans, as the variability in kinetic parameters among mammals is much less than the differences observed between folate monoglutamate and polyglutamate cofactor substrates.h g p yg The current model was validated by demonstrating that it recapitulates empirical observations regarding the impact of intracellular glycine on behavior of the FOCM network (Fig. 2). The validated model was then used to understand how the MTHFR C677T polymorphism, a known genetic risk factors for NTDs in humans, affects FOCM. This model shows that the lower levels of 5mTHF associated with the MTHFR 677 T variant are accom- panied by elevated levels of 10fTHF, which has been observed in animal models and in humans27–29 (Table 1). The model also indicates that 10fTHF accumulates in TT homozygotes as a result of increased flux through both the synthetase activity of MTHFD1 (FTS activity, Table 3), but also due to increased flux through MTHFD1 activity in the direction converting CH2F to 10fTHF (Table 3). Therefore, the model accurately predicts perturbations in FOCM that have been observed in human clinical and epidemiological studies. A recent study suggested that the risk of the MTHFR C677T polymorphism for NTDs was due to its known effect on lowering intracellular folate concentrations, rather than its role in providing 5 mTHF for homocysteine remethylation32. This model demon- strates that the MTHFR C677T polymorphism elevates levels of 10fTHF, which is known to be a chemically unstable form of folate that is susceptible to oxidative degradation, providing a mechanism by which the MTHFR C677T polymorphism depletes intracellular folate levels. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Human S. cerevesiae Genome size 3.0 × 109 bp 1.2 × 107 bp % AT 59% 61.5% T bases needed for replication 1.77 × 109 molecules 7.5 × 106 molecules Length of cell cycle 24 h 2.5 h Length of S-phase 8 h 0.83 h Cell volume 8 × 10−13 L (ES cell) 5 × 10−14 L dTMP synthesis rate required to replicate genome 7.8 µM/min 0.5 µM/min Measured dTMP synthesis rate (model outcomes) 4.4 µM/min 1.8 µM/min Ratio of dTMP production relative to dTMP required for replication 0.6 3.6 Table 12. Cellular capacity for de novo dTMP synthesis in mammals and yeast at S-phase. Human S. cerevesiae Genome size 3.0 × 109 bp 1.2 × 107 bp % AT 59% 61.5% T bases needed for replication 1.77 × 109 molecules 7.5 × 106 molecules Length of cell cycle 24 h 2.5 h Length of S-phase 8 h 0.83 h Cell volume 8 × 10−13 L (ES cell) 5 × 10−14 L dTMP synthesis rate required to replicate genome 7.8 µM/min 0.5 µM/min Measured dTMP synthesis rate (model outcomes) 4.4 µM/min 1.8 µM/min Ratio of dTMP production relative to dTMP required for replication 0.6 3.6 Table 12. Cellular capacity for de novo dTMP synthesis in mammals and yeast at S-phase. Table 12. Cellular capacity for de novo dTMP synthesis in mammals and yeast at S-phase. Specifically, the mouse model exhibits folate-responsive NTDs that occur with minimal perturbation in FOCM, and exhibit low and variable penetrance33, 34. In fact, most if not all, folate-related pathologies whose etiology involves interactions among genetic risk variants and nutrient exposures also exhibit low and/or variable clin- ical presentation. Understanding the stochastic behavior of the various reactions within FOCM that results in increased variability in FOCM network outputs is essential to understand which enzymes in the network contrib- ute to folate-related pathologies. p g Existing FOCM models rely on the limited quantity of kinetic data present in the literature, and the perfor- mance of the model will be dependent upon the kinetic parameters chosen to include in the model. Much of the available kinetic data for FOCM enzymes present in the literature was collected using the commercially available monoglutamate folate substrates, with few studies using the physiologically relevant polyglutamate forms of the cofactor. Discussion Understanding the dynamics of FOCM and its responsiveness to both genetic and environmental perturbations is the key to understanding the etiology of folate-related pathologies. Computational models and related simula- tions permit an identification of the most sensitive reactions within the network that exhibit the greatest degree of stochastic behavior leading to variability in network outputs. Furthermore, computational models allow an understanding of how both genetics and environmental factors can enhance or repress stochastic behavior at defined locations within the network, accelerating the development of diagnostics to identify those at risk for folate-related pathologies as well as lead to the development of targeted nutritional interventions for disease prevention.h p The Shmt1 knockout mouse model (Shmt1+/−, Shmt1−/− embryos) exhibits impaired de novo dTMP synthesis in the absence of perturbations of homocysteine remethylation. It also recapitulates risk for NTDs in humans. Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 7 Methods D i ti Description of the Model and of the Simulation Techniques. The model was constructed as a closed system using the subset of reactions that describe the FOCM pathways and homocysteine remethylation in cyto- plasm19 (Fig. 1). For the simulation, we employed a hybrid stochastic approach using deterministic simulation to compute the initial phase of the dynamics until a model steady state was reached, and then we assessed the stabil- ity of the achieved steady state by relying on exact stochastic simulation. We adopted a hybrid approach, rather than one entirely based on exact stochastic simulation59, because of the intensive computational effort introduced by the stiffness of the system during the simulation. yf y g The deterministic simulation was based on ODEs, where reactions were described in terms of Michaelis-Menten equations consistent with the original model of Reed19 and computed using the MATLAB integrator ode15s, whereas parameter estimates were derived from literature or calculated by nonlinear least squares optimization. The kinetic constants were obtained from folate polyglutamate cofactors and their inter- action with enzymes purified from L1210 cells where possible and otherwise from other mammalian tissue (see Supplementary Material for a detailed discussion of the model and parameter estimates). y The set of ODEs was further translated into a stochastic reaction-based model and a hybrid simulation approach60 was employed to quantify the level of stochasticity in the considered FOCM steady states. We refer to the Supplementary Material for a detailed description of how the ODE model has been translated to a reaction based stochastic one. According to the seminal work of Gillespie59, exact stochastic simulation allows simulation of each reaction event asynchronously when such reaction event is most probable to occur. To allow this, the simulation algorithm computes a propensity function aj(x) at each simulation step for each modeled reaction Rj, where x is the current state of the system providing the abundances of all modeled species at the considered time. The propensity value of a reaction has a direct link with the probability of its execution, that is, reactions with higher propensity are more likely to be fired in the near future. Methods D i ti To evaluate when the next reaction event will occur, also the total sum of propensities = ∑ a x a x ( ) ( ) R j 0 j is computed, because this quantity is linked to the number of reaction events occurring in the next time unit, that is, with increasing total propensity the number of reaction events per unit of time also increases. In Table 11, we provide the total sum of propensities for the con- sidered steady states. On average, we would need to generate up to 1015–1016 reaction events per unit of time during the stochastic simulation, due to stiffness of the system. To circumvent the problem, we relied on the concept of total propensity to evaluate the stability of the steady state, by assuming that a steady state is more stable when it has a lower value of a0(x), that is, a lower averaged number of reaction events that can perturb the equilibrium of the steady state. Model validation. To validate the FOCM model, in silico experiments were performed to determine if the model could recapitulate empirical data generated in MCF-7 cells by Herbig et al.10 focusing on the effect of gly- cine on FOCM (Fig. 2). Glycine is important because, as a second substrate, it has a direct influence on the reac- tion catalyzed by the enzymes GNMT as well as on the reversible reaction transforming CH2F to THF catalyzed by the enzyme SHMT. The purpose is to understand how the steady state of FOCM is affected by altering intracel- lular glycine concentrations. This was achieved by running several model simulations starting from different gly- cine concentrations and comparing the corresponding steady states with empirical data from Herbig et al.10. This study examined the effect of exogenous glycine at concentrations from 0 to 10 mM on the relative distribution of folate one-carbon forms as well as S-adenosylmethionine (SAM) and S-adenosylhomocysteine (SAH) levels. In summary, the empirical data revealed that as glycine concentrations increase intracellular 10fTHF levels increase at the expense of 5mTHF levels that decrease. Furthermore, as glycine concentrations increase SAM levels are depleted and SAH levels rise. www.nature.com/scientificreports/ Importantly, the model reported here demonstrates that the de novo dTMP biosynthesis enzymes are the most sensitive to low intracellular folate concentrations, with both DHFR and TYMS activities being repressed by 63% (Table 9). This finding is consistent with the finding that mouse models with impaired de novo dTMP biosynthesis are susceptible to NTDs in folate deficiency34. The hybrid stochastic simulation also reveals that both folate deficiency and the MTHFR C677T polymorphism create overall instability in the network (Table 11), consistent with a vast body of literature demonstrating an associa- tion of both folate deficiency and the MTHFR C677T polymorphism with various pathologies24, 30. Interestingly, SHMT1 exhibits the greatest increase in stochastic behavior as a result of the MTHFR C677T polymorphism (Table S6); the SHMT1 enzyme is the only FOCM enzyme that when disrupted results in folate-responsive NTDs34.hii The primary findings of this study are that the FOCM network is destabilized as a result of folate deficiency and the MTHFR677T polymorphism, and that SHMT is the most sensitive enzyme within the network to this net- work instability. This finding nicely connects the MTHFR genetic variant, a known risk factor for human NTDs, and SHMT1, the only folate enzyme whose disruption results in folate-responsive NTDs in mice. Furthermore, this model predicts that de novo dTMP synthesis rates in mammals are about half of what is required to meet DNA replication demands for dTMP (Table 12). Although mammals contain two pathways for dTMP synthesis, Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 8 www.nature.com/scientificreports/ the folate-dependent de novo dTMP synthesis pathway described here and a salvage pathway catalyzed by thy- midine kinase 1, the salvage pathway activity is insufficient to meet cellular needs based on observations that folate deficiency results in elevated uracil accumulation in DNA. In mammalian cells, the de novo dTMP syn- thesis enzymes form a multi-enzyme complex that interacts with DNA replication enzymes46. The discrepancy between de novo dTMP synthesis rates required to replicate the genome and the rate of dTMP synthesis currently predicted by the model indicates that the model should be extended to include multi-enzyme complex formation and substrate channeling in the nucleus to model more accurately determinants of FOCM and dTMP synthesis. The inclusion of the dTMP multi-enzyme metabolic complex in the model is expected to limit substrate diffusion and increase the rate of dTMP synthesis. 1. Beaudin, A. E. & Stover, P. J. Insights into metabolic mechanisms underlying folate-responsive neural tube defects: a minireview. Birth Defects Res A Clin Mol Teratol 85, 274–284 (2009). 2. Xu, X. et al. Folic acid therapy delays the progression of chronic kidney disease: the renal sub2 study of the China Stroke Primary Prevention Trial (CSPPT). JAMA Internal Medicine (2016). f 2. Xu, X. et al. Folic acid therapy delays the progression of chronic kidney disease: the renal sub2 study of the China Stroke Primary Prevention Trial (CSPPT). JAMA Internal Medicine (2016). Methods D i ti These changes were interpreted by the effects of glycine concentration driving the reversible SHMT reaction in the direction of serine synthesis10.f We simulated the effect of glycine on folate distribution, SAM, and SAH concentrations using the computa- tional model for values of glycine ranging from 0 to 10 mM (Fig. 2). The trends obtained by the model simula- tions were in agreement with the literature (green arrows in Fig. 2), confirming the coherence between model outcomes and empirical data. We observed only one exception related to the total % of 5 mTHF at 10 mM glycine (red arrow in Fig. 2). This discrepancy could be mainly due to two reasons: (1) 10 mM glycine is an extreme and non-physiological intracellular glycine concentration that could cause pharmacological effects, (2) the large mag- nitude of the experimental error in the Herbig et al. study at this glycine concentration. 1. Beaudin, A. E. & Stover, P. J. Insights into metabolic mechanisms underlying folate-responsive neural tube defects: a minireview Birth Defects Res A Clin Mol Teratol 85, 274–284 (2009). www.nature.com/scientificreports/ Homocysteine synthesis is elevated but total remethylation is unchanged by the methylenetetrahydrofolate d C T l h d b d f l ( ) 28. Davis, S. R. et al. Homocysteine synthesis is elevated but total remethylation is unchanged by the methylenetet reductase 677C->T polymorphism and by dietary folate restriction in young women. J Nutr 135, 1045–1050 (2005).i y y y g y y y reductase 677C->T polymorphism and by dietary folate restriction in young women. J Nutr 135, 1045–1050 (2005).i 9. Ghandour, H., Chen, Z., Selhub, J. & Rozen, R. Mice deficient in methylenetetrahydrofolate reductase exhibit tissue-specific distribution of folates. J Nutr 134, 2975–2978 (2004). 0. Stover, P. J., MacFarlane, A. J. & Field, M. S. Bringing clarity to the role of MTHFR variants in neural tube defect prevention. Am Clin Nutr 101, 1111–1112 (2015). , ( ) 31. Solis, C. et al. Folate intake at RDA levels is inadequate for Mexican American men with the methylenetetrahydrofolate reductase 677TT genotype. J Nutr 138, 67–72 (2008). 32. Tsang, B. L. et al. Assessing the association between the methylenetetrahydrofolate reductase (MTHFR) 677C > T polymorphism and blood folate concentrations: a systematic review and meta-analysis of trials and observational studies. Am J Clin Nutr 101, 1286–1294 (2015). 33. Martiniova, L., Field, M. S., Finkelstein, J. L., Perry, C. A. & Stover, P. J. Maternal dietary uridine causes, and deoxyuridine prev neural tube closure defects in a mouse model of folate-responsive neural tube defects. Am J Clin Nutr 101, 860–869 (2015). neural tube closure defects in a mouse model of folate-responsive neural tube defects. Am J Clin Nutr 101, 860–869 (2015). 34. Beaudin, A. E. et al. Shmt1 and de novo thymidylate biosynthesis underlie folate-responsive neural tube defects in mice. Am J Clin Nutr 93, 789–798 (2011).i 34. Beaudin, A. E. et al. Shmt1 and de novo thymidylate biosynthesis underlie folate-responsive neural tube defects in mice. Am J Nutr 93, 789–798 (2011).i ( ) 35. Beaudin, A. E. et al. Dietary folate, but not choline, modifies neural tube defect risk in Shmt1 knockout mice. Am J Clin Nutr 95, 109–114 (2012).i ( ) 6. Woeller, C. F., Anderson, D. D., Szebenyi, D. M. & Stover, P. J. Evidence for small ubiquitin-like modifier-dependent nuclear impor of the thymidylate biosynthesis pathway. J Biol Chem 282, 17623–17631 (2007). y y y y 37. Field, M. S. et al. www.nature.com/scientificreports/ Folate-pool interconversions and inhibition of biosynthetic processes after exposure of L1210 leukemia cells to antifolates Experimental and network thermodynamic analyses of the role of 12. Seither, R. L., Trent, D. F., Mikulecky, D. C., Rape, T. J. & Goldman, I. D. Folate-pool interconversions and inhibition of biosynthetic processes after exposure of L1210 leukemia cells to antifolates. Experimental and network thermodynamic analyses of the role of dihydrofolate polyglutamylates in antifolate action in cells. J Biol Chem 264, 17016–17023 (1989). y p yg y J ( ) 13. Jackson, R. C. & Harrap, K. R. Studies with a mathematical model of folate metabolism. Arch Biochem Biophys 158, 827–841 (1973). yg y & Harrap, K. R. Studies with a mathematical model of folate metab 13. Jackson, R. C. & Harrap, K. R. Studies with a mathematical mod 13. Jackson, R. C. & Harrap, K. R. Studies with a mathematical model of folate metabolism. Arch Biochem Biophys 158, 827–841 (1973). 14 M i P F & All C J F l t l ki ti i h b t ll J Bi l Ch 264 10552 10566 (1989) 13. Jackson, R. C. & Harrap, K. R. Studies with a mathematical model of folate metabolism. Arch Biochem Biophys 158, 827–8 14. Morrison, P. F. & Allegra, C. J. Folate cycle kinetics in human breast cancer cells. J Biol Chem 264, 10552–10566 (1989).i J p p y 14. Morrison, P. F. & Allegra, C. J. Folate cycle kinetics in human breast cancer cells. J Biol Chem 264, 10552–10566 (1989 g y 15. Strong, W. B., Tendler, S. J., Seither, R. L., Goldman, I. D. & Schirch, V. Purification and properties of serine hydroxymethyltr and C1-tetrahydrofolate synthase from L1210 cells. J Biol Chem 265, 12149–12155 (1990). i ofolate synthase from L1210 cells. J Biol Chem 265, 12149–12155 (1 y y 6. Martinov, M. V., Vitvitsky, V. M., Mosharov, E. V., Banerjee, R. & Ataullakhanov, F. I. A substrate switch: a new mode of regulation in the methionine metabolic pathway. J Theor Biol 204, 521–532 (2000).h p yh 17. Reed, M. C., Nijhout, H. F., Sparks, R. & Ulrich, C. M. A mathematical model of the methionine cycle. J Theor Biol 226, 33–43 (2004). 18. Prudova, A., Martinov, M. V., Vitvitsky, V. M., Ataullakhanov, F. I. & Banerjee, R. Analysis of pathological defects in methionine metabolism using a simple mathematical model. Biochim Biophys Acta 1741, 331–338 (2005).h 19. www.nature.com/scientificreports/ Reed, M. C. et al. A mathematical model of glutathione metabolism. Theor Biol Med Model 5, 8 (2008).fi h 20. Nijhout, H. F. et al. A mathematical model gives insights into the effects of vitamin B-6 deficiency on 1-carbon and glutathione metabolism. J Nutr 139, 784–791 (2009). 1. Reed, M. C. et al. A mathematical model gives insights into nutritional and genetic aspects of folate-mediated one-carbon metabolism. J Nutr 136, 2653–2661 (2006). 2. Goss, P. J. & Peccoud, J. Quantitative modeling of stochastic systems in molecular biology by using stochastic Petri nets. Proc Nat Acad Sci USA 95, 6750–6755 (1998). 23. Gillespie, J. H. Multilocus Behavior in Random Environments. II. Linkage Disequilibrium in an Additive Model. Genetics 87, 569–579 (1977). 4. Schwahn, B. & Rozen, R. Polymorphisms in the methylenetetrahydrofolate reductase gene: clinical consequences. American journa of pharmacogenomics: genomics-related research in drug development and clinical practice 1, 189–201 (2001). f p g g g p p 25. Chiang, E. P., Wang, Y. C. & Tang, F. Y. Folate restriction and methylenetetrahydrofolate reductase 677T polymorphism decreases adoMet synthesis via folate-dependent remethylation in human-transformed lymphoblasts. Leukemia 21, 651–658 (2007). 25. Chiang, E. P., Wang, Y. C. & Tang, F. Y. Folate restriction and methylenetetrahydrofolate reductase 677T polymorphism decreases d M t th i i f l t d d t th l ti i h t f d l h bl t L k i 21 651 658 (2007) 25. Chiang, E. P., Wang, Y. C. & Tang, F. Y. Folate restriction and methylenetetrahydrofolate reductase 677T polymorphism decr adoMet synthesis via folate-dependent remethylation in human-transformed lymphoblasts. Leukemia 21, 651–658 (2007). y p y y p 6. Quinlivan, E. P. et al. Methylenetetrahydrofolate reductase 677C->T polymorphism and folate status affect one-carbon incorporation into human DNA deoxynucleosides. The Journal of nutrition 135, 389–396 (2005). h 7. Bagley, P. J. & Selhub, J. A common mutation in the methylenetetrahydrofolate reductase gene is associated with an accumulation o formylated tetrahydrofolates in red blood cells. Proc Natl Acad Sci USA 95, 13217–13220 (1998). y y 28. Davis, S. R. et al. Homocysteine synthesis is elevated but total remethylation is unchanged by the methylenetetrahydrofolate reductase 677C->T polymorphism and by dietary folate restriction in young women. J Nutr 135, 1045–1050 (2005).ii 8. Davis, S. R. et al. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 3. Zhao, M., Li, X., Xing, C. & Zhou, B. Association of methylenetetrahydrofolate reductase C677T and A1298C polymorphisms with colorectal cancer risk: A meta-analysis. Biomed Rep 1, 781–791 (2013). y p 4. Price, A. J. et al. Circulating Folate and Vitamin B12 and Risk of Prostate Cancer: A Collaborative Analysis of Individual Participan Data from Six Cohorts Including 6875 Cases and 8104 Controls. Eur Urol (2016). y p 4. Price, A. J. et al. Circulating Folate and Vitamin B12 and Risk of Prostate Cancer: A Collaborative Analysis of Ind Data from Six Cohorts Including 6875 Cases and 8104 Controls Eur Urol (2016) g Data from Six Cohorts Including 6875 Cases and 8104 Controls. Eu g ( ) 5. Burr, N. E., Hull, M. A. & Subramanian, V. Folic Acid Supplementation May Reduce Colorectal Cancer Risk in Patients 5. Burr, N. E., Hull, M. A. & Subramanian, V. Folic Acid Supplementation May Reduce Colorectal Cancer Inflammatory Bowel Disease: A Systematic Review and Meta-Analysis. J Clin Gastroenterol (2016).h , , , , pp y k Inflammatory Bowel Disease: A Systematic Review and Meta-Analysis. J Clin Gastroenterol (2016).h pp y Inflammatory Bowel Disease: A Systematic Review and Meta-Analysis. J Clin Gastroenterol (2016).h y y y sentin, M., Zhao, R. & Goldman, I. D. The antifolates. Hematol Onc h ( ) 7. Ducker, G. S. & Rabinowitz, J. D. One-Carbon Metabolism in Health and Disease. Cell metabolism (2016). h Rabinowitz, J. D. One-Carbon Metabolism in Health and Disease. 8. Stover, P. J. Physiology of folate and vitamin B12 in health and disease. Nutr Rev 62, S3-12, discussion S13 (2004). y gy over, P. J. Folate-mediated one-carbon metabolism. Vitam Horm 79 . Fox, J. T. & Stover, P. J. Folate-mediated one-carbon metabolism. 10. Herbig, K. et al. Cytoplasmic serine hydroxymethyltransferase mediates competition between folate-dependent deoxyribonucle and S-adenosylmethionine biosyntheses. J Biol Chem 277, 38381–38389 (2002). y y 1. Scotti, M., Stella, L., Shearer, E. J. & Stover, P. J. Modeling cellular compartmentation in one-carbon metabolism. Wiley interdisciplinary reviews. Systems biology and medicine 5, 343–365 (2013). p y y gy ( ) 12. Seither, R. L., Trent, D. F., Mikulecky, D. C., Rape, T. J. & Goldman, I. D. Folate-pool interconversions and inhibition of biosynthetic t p y y gy 12. Seither, R. L., Trent, D. F., Mikulecky, D. C., Rape, T. J. & Goldman, I. D. References Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 9 Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w www.nature.com/scientificreports/ Mechanisms of dNTP supply that play an essential role in maintai h 57. Niida, H., Shimada, M., Murakami, H. & Nakanishi, M. Mechanisms of dNTP supply that pla genome integrity in eukaryotic cells. Cancer science 101, 2505–2509 (2010).fi pp y p y genome integrity in eukaryotic cells. Cancer science 101, 2505–2509 (2010).fi genome integrity in eukaryotic cells. Cancer science 101, 2505–2509 (2010).fi g g y y 58. Stover, P. J. & Field, M. S. Trafficking of intracellular folates. Adv Nutr 2, 325–331 (2011). r, P. J. & Field, M. S. Trafficking of intracellular folates. Adv Nutr 2, fi g 59. Gillespie, D. T. Stochastic simulation of chemical kinetics. Annu Rev Phys Chem 58, 35–55 (2007).hfi 60. Marchetti, L., Priami, C. & Thanh, V. H. Efficient hybrid stochastic simulation for spatially homogeneous biochemical rea networks. J. Comput. Phys. 317, 301–317 (2016). p y 61. Gostner, R., Baldacci, B., Morine, M. J. & Priami, C. Graphical Modeling Tools for Systems Biology. ACM Comput. Surv. 47, (2015). Author Contributions K.M., L.M., M.S., C.P., M.S.F., P.J.S. contributed to the experimental design. K.M., L.M., M.S. developed the computation model, while M.S.F., P.J.S., and M.S. provided all biological insight and extracted kinetic variables from the literature. M.S.F., P.J.S., and C.P. provided overall guidance of the project. All authors contributed to manuscript preparation and approved the final version. Acknowledgements g We would like to acknowledge Stephen M Stover for technical assistance. We are grateful to Lorenzo Stella g We would like to acknowledge Stephen M Stover for technical assistance. We are grateful to Lorenzo Stella, Tommaso Schiavinotto, Chiara Damiani and Alessandro Romanel for their helpful suggestions and discussions. This work was supported by Public Health Service grant DK58144. Tommaso Schiavinotto, Chiara Damiani and Alessandro Romanel for their helpful suggestions and discussions This work was supported by Public Health Service grant DK58144. www.nature.com/scientificreports/ Nuclear enrichment of folate cofactors and methylenetetrahydrofolate dehydrogenase 1 (MTHFD1) protect de novo thymidylate biosynthesis during folate deficiency. J Biol Chem 289, 29642–29650 (2014). i 38. Field, M. S., Kamynina, E., Watkins, D., Rosenblatt, D. S. & Stover, P. J. Human mutations in methylenetetrahydrofolate dehydrogenase 1 impair nuclear de novo thymidylate biosynthesis. Proceedings of the National Academy of Sciences of the United States of America 112, 400–405 (2015).i f 39. Hilton, J. G., Cooper, B. A. & Rosenblatt, D. S. Folate polyglutamate synthesis and turnover in cultured human fibroblasts. J Biol Chem 254, 8398–8403 (1979). 0. Schirch, V. & Strong, W. B. Interaction of folylpolyglutamates with enzymes in one-carbon metabolism. Arch Biochem Biophys 269 371–380 (1989). 41. Liew, S. C. & Gupta, E. D. Methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism: epidemiology, metabolism and the associated diseases. European journal of medical genetics 58, 1–10 (2015). 42. Anguera, M. C. et al. Regulation of folate-mediated one-carbon metabolism by 10-formyltetrahydrofolate dehydrogenase. J Biol Chem 281, 18335–18342 (2006). l d d ll f Th d l d d d ( l) ( ) 43. Ueland, P. M., Nygard, O., Vollset, S. E. & Refsum, H. The Hordaland Homocysteine Studies. Lipids 36(Suppl), S33–39 (2001). h 44. Blount, B. C. et al. Folate deficiency causes uracil misincorporation into human DNA and chromosome breakage: implications for cancer and neuronal damage. Proc Natl Acad Sci USA 94, 3290–3295 (1997). g 5. Macfarlane, A. J. et al. Nuclear localization of the De Novo thymidylate biosynthesis pathway is required to prevent uraci accumulation in DNA. J Biol Chem (2011). Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 10 www.nature.com/scientificreports/ www.nature.com/scientificreports/ 6. Anderson, D. D., Woeller, C. F., Chiang, E. P., Shane, B. & Stover, P. J. Serine hydroxymethyltransferase anchors de novo thymidylate synthesis pathway to nuclear lamina for DNA synthesis. J Biol Chem 287, 7051–7062 (2012).hh 47. Poon, P. P. & Storms, R. K. Thymidylate synthase is localized to the nuclear periphery in the yeast Saccharomyces cerevisiae Journal of biological chemistry 269, 8341–8347 (1994).hf 47. Poon, P. P. & Storms, R. K. Thymidylate synthase is lo Journal of biological chemistry 269, 8341–8347 (1994).hf f g y ( ) 8. Young, R. C. & DeVita, V. T. The effect of chemotherapy on the growth characteristics and cellular kinetics of leukemia L1210 Cancer Res 30, 1789–1794 (1970).hf ( ) 9. Neil, G. L. & Homan, E. R. The effect of dose interval on the survival of L1210 leukemic mice treated with DNA synthesis inhibitors Cancer Res 33, 895–901 (1973). 50. Becker, K. A. et al. Self-renewal of human embryonic stem cells is supported by a shortened G1 cell cycle phase. Journal of cellular physiology 209, 883–893 (2006).f p y gy ( ) 51. Sagi, I. et al. Derivation and differentiation of haploid human embryonic stem cells. Nature 532, 107–111 (2016). f 52. Greenwood, M. T., Calmels, E. M. & Storms, R. K. Growth-rate-dependent regulation of the expression and inactivation of thymidylate synthase in Saccharomyces cerevisiae. Journal of bacteriology 168, 1336–1342 (1986). y y y y J f gy , ( ) 53. Brewer, B. J., Chlebowicz-Sledziewska, E. & Fangman, W. L. Cell cycle phases in the unequal mother/daughter cell cycles of Saccharomyces cerevisiae. Mol Cell Biol 4, 2529–2531 (1984). Saccharomyces cerevisiae. Mol Cell Biol 4, 2529–2531 (1984) y 4. Chabes, A. et al. Survival of DNA damage in yeast directly depends on increased dNTP levels allowed by relaxed feedback inhibition of ribonucleotide reductase. Cell 112, 391–401 (2003). 55. Gon, S., Napolitano, R., Rocha, W., Coulon, S. & Fuchs, R. P. Increase in dNTP pool size during the DNA damage response plays a key role in spontaneous and induced-mutagenesis in Escherichia coli. Proc Natl Acad Sci USA 108, 19311–19316 (2011).ht y p g 56. Hakansson, P., Hofer, A. & Thelander, L. Regulation of mammalian ribonucleotide reduction and dNTP pools after DNA damage and in resting cells. The Journal of biological chemistry 281, 7834–7841 (2006). gh f g y 57. Niida, H., Shimada, M., Murakami, H. & Nakanishi, M. Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w Additional Information Supplementary information accompanies this paper at doi:10.1038/s41598-017-00854-w Competing Interests: The authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps an institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2017 Scientific Reports \| 7: 797 \| DOI:10.1038/s41598-017-00854-w 11
https://openalex.org/W2010549321	https://ccsenet.org/journal/index.php/ach/article/download/6607/5193	English	null	Advices about How to Make Career Planning for College Students	Asian culture and history	2,010	cc-by	2,242	Advices about How to Make Career Planning for College Students Aiqing Meng Dezhou University, Dezhou 253023, China Tel: 86-534-898-5031 E-mail: mengaq1198@163.com Tel: 86-534-898-5031 E-mail: mengaq1198@163.com Asian Culture and History Asian Culture and History Vol. 2, No. 2; July 2010 www.ccsenet.org/ach Abstract To plan college students’ own careers, they should deeply understand the society, confirm the struggle target, construct reasonable knowledge structure, actively participate in the social practice, and cultivate good psychological quality to establish firm base for their successful employments. Keywords: Career, Planning, Design Keywords: Career, Planning, Design Career planning means one person’s all behaviors and activities about his career in his life, and it is the continual experiences about one person’s life attitude, value view, and individual will in his life, and it is also one person’s change of occupation and post and the realization of working ideal. College expansion and economic crisis have brought large difficulty of employment, and how should college students plan their careers? How to realize their own life value? For these problems, they will feel confused, which could be presented in two aspects. First, they fell confused in their specialty, individual interest, and future post. Second, they generally don’t know to accord with the society. When they select their works or are employed, many of them have not known themselves clearly, the industry and the society. Many students could not realize ideal employment in one time. Therefore, college students should make good preparation before they begin to work, and they should especially plan their career well. How to plan the career? How to make the plans? Following aspect should be emphasized. . Correctly knowing their own specialties and deeply understanding the demand of the society Before planning the career, every college students should make clear that “What person I am? What do I want to do? What work can I do? What work can the exterior environment support me?” When colleges instruct their students to plan career, they should make college students know the cultivated target of their specialties, and the direction of the employment, strengthen their self-consciousness of their specialties, cultivate their learning target, and know and love their future works, and colleges should establish firm thinking base for students’ future occupation target. In the employment market at present, the existence that supply exceeds demand makes the employing units have special requirements for graduates. Their first objectives are graduates’ strong suit in certain specialty, and graduates’ main achievements after they enter into the society depend on the base of their specialty, and their initial successes depend on the specialty knowledge learned in the colleges. If college students leave their specialties, they will increase many “burdens” invisibly when they plan their career, which will increase more difficulties to realize their life values. Therefore, college students should not only more deeply study the specialty knowledge what they learned, but also lean more knowledge. The more important is that they should further widen the knowledge about the specialty, grasp and know more knowledge and technologies which are relative to the specialty. At the same time, high vocational students should comprehensively know their own individual characters, interests, favors, and employment ability. 3. Constructing reasonable knowledge structure College students should not only have certain professional knowledge, but reasonable knowledge structure. Colleges should cultivate students to grasp various qualities, encourage them to participated in various activities, train their own abilities, support them to enhance their comprehensive qualities. Students should actively participate in the part-time jobs and social practices, and engage in the works which are close to their future occupation after school. Students should also actively do more scientific-researches, and enhance their professional ethics, responsibilities, working enthusiasm, and individual frustrated ability in the practice. At the same time, college students should also strengthen their language expression ability, oral foreign language ability, and computer application ability, actively strive for relative certifications about English and computer, and selectively earn credits of other specialties to fulfill themselves. By testing their accumulated abilities and skills, they should modify their individual career planning according to their interests and knowledge quality. There is no a short cut to establish reasonable knowledge structure, and the basic approaches are only learning and accumulation. The reasonable knowledge structure should not be established one for all, and people have to continually pay hardly labors. Only by proper scientific method and continual struggles, college students will certainly establish and perfect their own knowledge structure, and accumulate good base for their future employments. 2. Confirming the target and planning the future If one person wants to succeed, he should make concrete and feasible plan, which is the most important basic factor of the success. With the plan, people will not act blindly, and with the plan, people could concretely carry out and implement the plan, and approach to their own target step by step, even realize and exceed the target. The establishment of the plan should refer to the target which should be scientific and feasible and proper. Too high target is difficult to be realized, even could not be realized, and lose people’s confidence. And too low target is too easy to be realized, and people will feel chesty and even lose the upward mobility. Only proper and scientific target with feasible plan could stimulate people’s fight to realize their own target. College students should realize that the target is the power, and only the struggle could achieve the success. Higher vocational students should learn starting from the most basic and concrete employment post, and only if the tiny post accords with their final employment target, it could be confirmed as their initial employment post. Things should not be distinguished by lowliness and nobleness, and the final results are difficult because peoples’ targets are ISSN 1916-9655 E-ISSN 1916-9663 236 Asian Culture and History Vol. 2, No. 2; July 2010 www.ccsenet.org/ach different, and the accomplishments of numerous tiny things mean the achievement of the final target, and the realization of the life ideal. Colleges students should fully realize people’s career planning should be composed by many tiny things, only they are based on these tiny things, they will accomplish the big thing. In the premise knowing themselves and understanding the society, college students should start from their practice and the demand of the society to confirm the direction of the employment development, make clear the qualities what the employment target needs and the advantages and disadvantages realizing the target. 4. Cultivating the practice ability and certain expression ability needed by the career College students should also strengthen the learning of professional skill, and they should enhance their employment skill and innovational ability. Students could train their ability and creativity to solve problems independently by the quality development activities, and actively participate in some summer practices about the specialty, and increase the association with their graduated school fellows, and communicate experiences applying for jobs, and study to write resumes and application letters, and increase the channels to collect the working information. College students’ comprehensive ability and knowledge are the references for the employing units to select them, and the employing units will check not only their professional knowledge and skills, but their ability to comprehensively utilize knowledge, and their ability to adapt the environment, their ability to integrate the cultures, and their practical ability. The expression ability is the ability to express their own opinions, ideas, and emotions by the language, and it generally includes the oral expression ability, the word expression ability, and the figure expression ability. For college students, the importance of the expression ability is self-evident. When they enter into the society, they will realize that immediately, and in addition, when they apply for a job or select their jobs, they will deeply feel that. For the writing of the recommendation letters, the answer of the application questions, and the interviews, each part needs certain expression ability. Therefore, when cultivating the expression ability, college students should strive for exact, clear, and lively expression as far as possible. 5. Actively participating in beneficial social practice and career training, and strengthening the ability to adapt the career Feng, Lan. (2004). Psychological Analysis and Countermeasures of College Students’ Difficult Employment. Journal of Southwest University for Nationalities (Humanities and Social Science). No. 1, p.408-410. Xiong, Pingzhuo. (2006). Career Planning. Changsha: Zhongnan University Press. April. 6. Strengthening the self-improvement and training, and cultivating good psychological qualit College students should cultivate the optimistic habit of enduring hardships and being capable of hard work, and have the host community awareness. One educator said that the person who did anything well was the person with real quality. In fact, in the world of work, people always love those persons with the host community awareness because they could endure hardships and be capable of hard work. Different enterprises have different cultures, but no one enterprise likes the person who want more demands and fewer contributions, have more complaints and less struggles, require more cares and pay less cares. The host community awareness means the awareness of responsibility, and the person with this awareness could be loyal to his enterprise, respect leaders, love works, care about colleagues, think about the benefits of the enterprise at any time, correctly treat his own gain and loss, failure and success. The host community awareness is the awareness of contribution, without too much individual splitting hairs. To enhance their own status in the world of work, college students should depend on their firm quality and their contributions for the enterprise. Any one enterprise will love the person who could actively contribute the organization and the group. For a long time, the person paying more to the enterprise and others will obtain more admittance. So college students should first confirm their feasible career orientation and target, and decompose the target, and then design reasonable career planning, and implement them. By continual struggles and adjustments, they will certainly realize their own career development targets. At the same time, college students’ career planning depends on the government, the society, the college, and their won enthusiasm and creativity, and the close integration and comprehensive push of above aspects. 7. Enhancing the comprehensive quality and cultivating certain management ability Though not every college students will engage in the management work after they graduate, but every one will need the management skill in their future works. The modern society has shown that the organization and management ability is not the ability for leaders and managers, but for all people. With the development of the times, the college students not matter which specialty they belong should possess not only profound professional knowledge, but certain ability of organization and management, which is not only the demand of employment, but the objective demand of the times. The career planning is a long-term engineering, and the planning in the college stage is only a start. Colleges should take this work as the core content and important project of the education, explore and help college students to establish proper employment target, plan their careers, and enhance the science standard and the actual effect of the college students’ employment, cooperating with various functional departments. 5. Actively participating in beneficial social practice and career training, and strengthening the ability to adapt the career Career training includes the training of career skills, the adaptability checking of the occupation, and the scientific measurement of occupational intention. College students should first check whether their confirmed career targets are definite and whether their preparations are sufficient. Then, they should train themselves aiming at these problems, accept the training of work selection skills, participate in the recruitment, and check their accumulations and preparations in the practice. College students should also fully utilize the conditions of the college, look for the information of the employing units, strengthen their employment application skills, train their interview skills, and complete more exercises with sufficient preparations as far as possible. One person’s ability adapting the society is the comprehensive reflection of his quality and abilities, which is close with his moral quality, knowledge skill, activity ability, innovational ability, the ability dealing with the interpersonal relationship, and healthy status. Generally speaking, the college student with higher quality, strong ability, and healthy body and heart could adapt the environment and work as soon as quickly when he enters into 237 Published by Canadian Center of Science and Education Asian Culture and History Vol. 2, No. 2; July 2010 www.ccsenet.org/ach the society, and even in the bad condition and environment, he could achieve good achievements by his struggles, or turn disadvantageous environment into advantageous environment. Survival of the fittest means that the survival is for the development. Only college students could pay attention to their ability to adapt the society, they could reduce the adapting term and fully exert their intelligences fully after they enter into the society. the society, and even in the bad condition and environment, he could achieve good achievements by his struggles, or turn disadvantageous environment into advantageous environment. Survival of the fittest means that the survival is for the development. Only college students could pay attention to their ability to adapt the society, they could reduce the adapting term and fully exert their intelligences fully after they enter into the society. References Feng, Lan. (2004). Psychological Analysis and Countermeasures of College Students’ Difficult Employment. Journal of Southwest University for Nationalities (Humanities and Social Science). No. 1, p.408-410. Xiong, Pingzhuo. (2006). Career Planning. Changsha: Zhongnan University Press. April. 238 ISSN 1916-9655 E-ISSN 1916-9663 ISSN 1916-9655 E-ISSN 1916-9663
https://openalex.org/W2810266386	https://www.nature.com/articles/s41598-018-28407-9.pdf	English	null	Author Correction: Towards Millimeter-wavelength: Transmission-Mode Fresnel-Zone Plate Lens Antennas using Plastic Material Porosity Control in Homogeneous Medium	Scientific reports	2,018	cc-by	2,319	www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Author Correction: Towards Millimeter-wavelength: Transmission-Mode Fresnel- Zone Plate Lens Antennas using Plastic Material Porosity Control in Homogeneous Medium PEN Published: xx xx xxxx Reference 19 was inadvertently omitted from the final version of our Article. The corrected article provides a comparison between our proposed design and that shown in reference 19 and explains how our Fresnel-Zone Plate design may overcome drawbacks in reference 19. Additionally, there were typographical errors in Equations 8-10. The authors apologize for these errors and any confusion caused. “In millimeter-wave band, the fabrication of illuminators that uniformly illuminates the lens surface to obtain an intended efficiency is difficult. For this reason, amplitude tapering must be considered for illumination of lens feed to obtain the desired efficiency. To accomplish this goal, two crucial factors are studied: the taper efficiency (ηtaper) and the spill-over efficiency (ηSP). In the classic design of Fresnel zone plates, horn antennas are employed as an illuminator to feed the zone plates, the radiation pattern of the lens feeder serves a cosn-like function. Since the fabrication of this kind of feed at millimeter-wave band is complicated. Therefore, a new type of illuminator or cosn-like radiation pattern horn with low SLL must be considered for illumination. To reach this purpose, the- oretical aperture efficiency versus cosn-like radiation pattern amplitude weighting is analyzed to obtain expected efficiency with n variations. The taper efficiency (ηtaper)equation for the loss of non-uniform illumination of the aperture amplitude and the spill-over efficiency (ηSP) are given by (8) and (9), individually15,16: ∫ ∫ ∫ ∫ η π θ φ θ φ θ φ θ θ φ = . .   .   . θ π θ θ π θ . ( ) ( ) G d d G d d 1 tan ( , ) tan ( , ) sin( ) , (8) taper 2 2 0 2 0 0 5 2 2 0 2 0 0 0 0 ∫ ∫ ∫ ∫ η θ φ θ θ φ θ θ π = . . π θ π θπ G G d d ( , ) sin( ) ( , ) sin( ) , (9) SP 0 2 0 0 2 0 0 where G(θ, φ) is the radiation pattern, θ is dedicated to polar angle, and φ is the azimuthal angle15,16. ∫ ∫ ∫ ∫ η π θ φ θ φ θ φ θ θ φ = . .   .   . θ π θ θ π θ . SCIENTIfIC REPOrTs \| (2018) 8:10293 \| DOI:10.1038/s41598-018-28407-9 now reads: “Lens design criteria for the focusing devotions can be classified into two modes of illumination: transmission and reflection modes. The essential devices of each focusing setup are assigned to two central parts that can achieve them in all pieces of literature [1–19] with specific names: a lens device (focusing provider) and an illu- minator/feeder/wave-launcher. Generally, all classes of lens devices for wave-focusing setup are independent of electromagnetic spectrums, and they have had their treatment scenarios and produce predetermined sub-zone permittivity techniques. However, the treatment of the proper illuminator and low-error subzone permittiv- ity implementation is the apparent goal of all focusing mechanisms, which can be considered to guarantee the desired radiation effect in a predetermined setup with a defined focal point (F), lens diameter (D), and antenna gain parameters. As shown in Fig. 1 (a), a Fresnel zone plate is a planar symmetric gradient index structure with a z-axis in the center. Therefore, by providing axially symmetric illumination [9] for feeding lens apertures, the zone plate’s radi- ation patterns should be symmetrical with high gains and high efficiency at the focusing point (P1). Therefore, we can verify that any asymmetrical zone plate’s radiation is due to permittivity estimation errors in the homoge- neous design principles. To achieve this goal, an axially symmetric feed [6] or uniform feed [15, 19] based on the [15, 19] feed analysis is considered. In millimeter-wave band over 30 GHz, the fabrication of illuminators that uniformly [15, 19] or axially symmetric form illuminates [6] the lens surface to obtain an intended efficiency is difficult [19]. For this reason, amplitude tapering must be considered for illumination of lens feed to obtain the desired efficiency for proposed lenses based on [15], and [19] feed analysis. To accomplish this goal, two crucial factors are considered based on Kildal feed analysis, and approximations in [15]: the taper efficiency (ηtaper) and the spill-over efficiency (ηSP) [15, 19]. In the classic design of Fresnel zone plates, horn antennas are employed as an illuminator to feed the zone plates, the radi- ation pattern of the lens feeder serves a cosn-like function [6], [15], [19]. Since the fabrication of this kind of feed at millimeter-wave band is complicated [19]. Therefore, a new type of illuminator or cosn-like radiation pattern horn with low SLL must be considered for illumination [6], [15], [19]. Author Correction: Towards Millimeter-wavelength: Transmission-Mode Fresnel- Zone Plate Lens Antennas using Plastic Material Porosity Control in Homogeneous Medium PEN ( ) ( ) G d d G d d 1 tan ( , ) tan ( , ) sin( ) , (8) taper 2 2 0 2 0 0 5 2 2 0 2 0 0 0 0 (8) ∫ ∫ ∫ ∫ η θ φ θ θ φ θ θ π = . . π θ π θπ G G d d ( , ) sin( ) ( , ) sin( ) , (9) SP 0 2 0 0 2 0 0 (9) where G(θ, φ) is the radiation pattern, θ is dedicated to polar angle, and φ is the azimuthal angle15,16. where G(θ, φ) is the radiation pattern, θ is dedicated to polar angle, and φ is the azimuthal angle15,16. National Institute of Scientific Research (INRS), Centre for Energy, Materials and Telecommunication (EMT), Quebec, Montreal, H5A 1K6, Canada. Javad Pourahmadazar and Tayeb A. Denidni contributed equally to this work. Correspondence and requests for materials should be addressed to J.P. (email: javad.pourahmadazar@emt.inrs.ca) SCIENTIfIC REPOrTs \| (2018) 8:10293 \| DOI:10.1038/s41598-018-28407-9 www.nature.com/scientificreports/ θ θ π = + . ≤ G n n cos for ( , ) {(2 1) ( ), 0 /2; n θ > for pi 0, 2 ; θ > for pi 0, 2 ; (10) (10) Since the axially symmetric radiation pattern for illuminator is given by (10)15, accordingly, with increasing n, the proposed feed will generate high spill-over efficiency (ηSP), and the taper efficiency (ηtaper) will decrease. Considering to the lens diameter and the lens focal length (F), a higher n for cosn-like radiation pattern to obtain a product of two efficiencies as total efficiency (ηT = ηtaper.ηSP) is desirable. Illustration of this efficiency versus n value, as shown in Fig. 6, shows that the optimum point (η) to obtain maximum total efficiency ηT = 0.81 is equal to n = 40. As shown in Fig. 6, to obtain the total efficiency between 0.31 < ηT < 0.71, n must be chosen between 10 <n<20 ranges. Considering to provided information for total efficiency higher than ηT = 0.71, n must be chosen between 22 and 78. For this purpose, two types of feed as illuminators to study of the designed FZPs efficiency are considered: (a) a cos10-like radiation pattern feed, and (b) a cos45-like radiation pattern feed. Author Correction: Towards Millimeter-wavelength: Transmission-Mode Fresnel- Zone Plate Lens Antennas using Plastic Material Porosity Control in Homogeneous Medium PEN To achieve the first feed with cos10-like radiation pattern a microstrip dipole antenna is designed, and to obtain second feed with cos40-like radiation pattern a commercial horn antenna are considered to feed lens.” now reads: To reach this purpose, theoretical aperture efficiency versus cosn-like radiation pattern amplitude weighting is analyzed to obtain expected efficiency with n variations [15, 19]. The taper efficiency (ηtaper)equation for the loss of non-uniform illumination of the aperture amplitude and the spill-over efficiency (ηSP) are given by (Eq. 8) and (Eq. 9), individually [15], [16], [19]: ( ) ( ) G d d G d d 1 tan ( , ) tan ( , ) sin( ) , (8) taper 2 2 0 2 0 0 5 2 2 0 2 0 0 0 0 ∫ ∫ ∫ ∫ η π θ φ θ φ θ φ θ θ φ = . .   .   . θ π θ θ π θ . (8) G d d G d d ( , ) sin( ) ( , ) sin( ) , (9) SP 0 2 0 0 2 0 0 ∫ ∫ ∫ ∫ η θ φ θ θ φ θ φ θ θ φ = . . π θ π π (9) where G(θ, φ) is the radiation pattern, θ is dedicated to polar angle, and φ is the azimuthal angle15,16. where G(θ, φ) is the radiation pattern, θ is dedicated to polar angle, and φ is the azimuthal angle15,16. θ θ θ π θ π = + . < < > G n n cos for for ( , ) {(2 1) ( ), 0 /2; 0 ; n θ θ θ π = + . < < G n n cos for ( , ) {(2 1) ( ), 0 /2; n θ θ θ π θ π = + . < < > G n n cos for for ( , ) {(2 1) ( ), 0 /2; 0, 2 ; (10) n θ π > for 0, 2 ; (10) (10) Since the axially symmetric radiation pattern for illuminator is given by (Eq. 10) [6], accordingly, with increasing n, the proposed feed will generate high spill-over efficiency (ηSP), and the taper efficiency (ηtaper) will decrease [19]. Considering to the lens diameter and the lens focal length (F), a higher n for cosn-like radiation pattern to SCIENTIfIC REPOrTs \| (2018) 8:10293 \| DOI:10.1038/s41598-018-28407-9 2 www.nature.com/scientificreports/ obtain a product of two efficiencies as total efficiency (ηT = ηtaper.ηSP) is desirable [15, 19]. Illustration of this effi- ciency versus n value, as shown in Fig. now reads: 6, shows that the optimum point (η) to obtain maximum total efficiency ηT = 0.81 is equal to n = 40 [15, 19]. As shown in Fig. 6, to obtain the total efficiency between 0.31 < ηT < 0.71, n must be chosen between 10 <n<20 ranges. Considering to provided information for total efficiency higher than ηT = 0.71, n must be chosen between 22 and 78. For this purpose, two types of feed as illuminators to study of the designed FZPs efficiency are considered: (a) a cos 10-like radiation pattern feed, and (b) a cos 45-like radiation pat- tern feed. To achieve the first feed with cos 10-like radiation pattern a microstrip dipole antenna is designed, and to obtain second feed with cos 45-like radiation pattern a commercial horn antenna are considered to feed lens. Concerning structural comparison with [19], the provided lenses do not have any structural similarities or physi- cal relationships and are just determined by the classification of their Fresnel zone plate arrangements with differ- ent applications toward two distinct bodies: planar and corrugated forms. In the case of feed types and illuminator designs, all Fresnel lenses analyses with horn antenna illuminations are similar, which is illustrated in [6] for high-frequency treatments in detail. The presented study in [6] describes the high-efficiency lens treatments while focusing on the type of illuminations such as axially symmetric feeds (Eq. 9) and lens classes. Although based on [6] studies, the results of both hard and soft material lenses are foreseeable, but we and [19] designers tried to solve manufacturing zone plate difficulties with different soft and hard plastic materials. In [19] Fresnel zone prototypes, the designers have decided to use high-efficiency luminosity in the feeding section at the expense of manufacturing problems and maintenance costs in lens platform to achieve high efficiency. However, the general purpose of our structure is concentrated to producing a cubic cell with the ability of intended permittivity control in a homogeneous environment, which is entirely dissimilar and innovative for Fresnel lens treatments. Based on two types of radiation feed applied to both lens surfaces, as described in the next sections, lenses-out radiation has an entirely symmetrical form, a high gain, and high efficiency. now reads: Expected results compared to previ- ously reported devices with similar type feeds, as shown in Table I, indicate the accuracy of the estimated permit- tivity method for phase corrector zones designed with hard plastic cube-shaped cells. Regarding the mentioned results with two distinct materials, the proposed design scheme has already answered the earlier problems with low efficiencies in hard plastic slabs (see Table I). Also, it has responded to the issues of manufacturing and keep- ing in services for the similar Fresnel-type lenses in [19] with soft foam materials.” The original Figure 6 legend, “Illustration of ηtaper, ηSP and ηtotal efficiency over the amplitude weighting generated by cosn-like illumination now reads:- now reads:- “Graphical illustration of ηtaper (Eq. 8), ηSP (Eq. 9), and ηtotal efficiency [19] over the amplitude weighting generated by cosn-like illumination [6, 19].” Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2018 SCIENTIfIC REPOrTs \| (2018) 8:10293 \| DOI:10.1038/s41598-018-28407-9 3
https://openalex.org/W2793094522	https://europepmc.org/articles/pmc5855194?pdf=render	English	null	Crowdsensing in Smart Cities: Overview, Platforms, and Environment Sensing Issues	Sensors	2,018	cc-by	15,500	Crowdsensing in Smart Cities: Overview, Platforms, and Environment Sensing Issues Oscar Alvear 1,2,∗, Carlos T. Calafate 1, Juan-Carlos Cano 1 and Pietro Manzoni 1 Oscar Alvear 1,2,∗, Carlos T. Calafate 1, Juan-Carlos Cano 1 and Pietro Manzoni 1 1 Department of Computer Engineering, Universitat Politecnica de Valencia, 46022 Valencia, Spain; calafate@disca.upv.es (C.T.C.); jucano@disca.upv.es (J.-C.C.); pmanzoni@disca.upv.es (P.M.) 2 Department of Electrical Engineering, Electronics and Telecommunications, Universidad de Cuenca, Cuenca 010150, Ecuador * Correspondence: oscar.alvear@alttics.com Oscar Alvear 1,2,∗, Carlos T. Calafate 1, Juan-Carlos Cano 1 and Pietro Manzoni 1 1 Department of Computer Engineering, Universitat Politecnica de Valencia, 46022 Valencia, Spain; calafate@disca.upv.es (C.T.C.); jucano@disca.upv.es (J.-C.C.); pmanzoni@disca.upv.es (P.M.) 2 Department of Electrical Engineering, Electronics and Telecommunications, Universidad de Cuenca, Cuenca 010150, Ecuador * Correspondence: oscar.alvear@alttics.com Oscar Alvear , Carlos T. Calafate , Juan-Carlos Cano and Pietro Manzoni 1 Department of Computer Engineering, Universitat Politecnica de Valencia, 46022 Valencia, Spain; calafate@disca.upv.es (C.T.C.); jucano@disca.upv.es (J.-C.C.); pmanzoni@disca.upv.es (P.M.) 2 Department of Electrical Engineering, Electronics and Telecommunications, Universidad de Cuenca, Cuenca 010150, Ecuador * Correspondence: oscar.alvear@alttics.com Received: 3 December 2017; Accepted: 31 January 2018; Published: 4 Febuary 2018 Abstract: Evidence shows that Smart Cities are starting to materialise in our lives through the gradual introduction of the Internet of Things (IoT) paradigm. In this scope, crowdsensing emerges as a powerful solution to address environmental monitoring, allowing to control air pollution levels in crowded urban areas in a distributed, collaborative, inexpensive and accurate manner. However, even though technology is already available, such environmental sensing devices have not yet reached consumers. In this paper, we present an analysis of candidate technologies for crowdsensing architectures, along with the requirements for empowering users with air monitoring capabilities. Speciﬁcally, we start by providing an overview of the most relevant IoT architectures and protocols. Then, we present the general design of an off-the-shelf mobile environmental sensor able to cope with air quality monitoring requirements; we explore different hardware options to develop the desired sensing unit using readily available devices, discussing the main technical issues associated with each option, thereby opening new opportunities in terms of environmental monitoring programs. Keywords: Smart city; Internet of Things; crowdsensing; sensor design Sensors 2018, 18, 460; doi:10.3390/s18020460 www.mdpi.com/journal/sensors Keywords: Smart city; Internet of Things; crowdsensing; sensor design 1. Introduction In a smart city, all of these areas must have some level of intelligence to minimise management efforts. Thus, there are various subareas of interest including Smart Governance, Smart Mobility, Smart Utilities, Smart Buildings, and Smart Environment, where the adoption of this paradigm can have a clear impact, being highly beneﬁcial [12,13]. Any city has several areas of concern to the authorities. In a smart city, all of these areas must have some level of intelligence to minimise management efforts. Thus, there are various subareas of interest including Smart Governance, Smart Mobility, Smart Utilities, Smart Buildings, and Smart Environment, where the adoption of this paradigm can have a clear impact, being highly beneﬁcial [12,13]. Sensing processes are one of the most important tasks in a smart city because they allow retrieving the different parameters involved in different control processes. Examples of such processes include transportation, energy management, air conditioning, etc. However, controlling air pollution in smart cities stands out as a key issue, as it has severe consequences on human health, thereby making environment sensing a critical task and a prominent service. Currently, controlling pollution levels is an on-going effort undertaken by most European cities, which invest a considerable amount of money in controlling the different hazards produced by poor air quality. Considering these aspects, an index known as AQI (Air Quality Index [14]) was created to classify the air quality, and it speciﬁes different healthy risks, from Low to Very High. Air quality assessment mainly relies on static monitoring stations, meaning that most cities are endowed with at least one of these stations. Overall, there are about 1500 air quality monitoring stations in Europe. However, the installation and maintenance of these stations is quite expensive. In addition, installing new air monitoring stations is tough since, in crowded places, there is no room for allocating them, meaning that they are typically installed in remote locations, such as parks or sparsely populated areas, which is prone to produce data with little representativeness for the overall population. There are several air pollutant types, and various techniques can be adopted to measure them. 1. Introduction Smart cities are revolutionising our view of the world, and their functioning achieves a very high level of integration, coordination, and cooperation between ordinary objects, providing them with some degree of intelligence. This novel paradigm provides a plethora of systems and technological tools aimed at increasing our life quality, minimising the environmental impact of everyday activities, and optimising resource usage. Such effects are more noticeable in urban areas with millions of citizens, the so called mega cities, which in the near future will be more and more common [1]. The main concept behind a Smart City is the integration of the physical world with the virtual world [2]. This is achieved by providing additional capabilities such as environmental sensing and automatic behaviour to common objects, allowing to capture and to analyse the data from the real world to ensure a better operation of the virtual one. There are several technologies/perspectives that simplify the process of creating a Smart City. Figure 1 provides a global overview of available technologies from diverse perspectives, which cover different aspects that allow creating a Smart City. Thus, in a Smart City, all daily objects, called things, are equipped with extra capabilities, usually sensing and/or acting capabilities, along with communication capabilities, to share information and to optimise their functional operation. This way, and from a communications perspective, the Internet of Things (IoT) [3] focuses on the intercommunication between all things, as well as on the communication between things and data servers (Cloud or Fog/Edge). On the other hand, from an operational perspective, Cyber Physical Systems (CPS) focus on the integration of these physical things with the computational Sensors 2018, 18, 460; doi:10.3390/s18020460 www.mdpi.com/journal/sensors 2 of 28 Sensors 2018, 18, 460 process [4–6] to improve its functionally. Finally, from a service perspective, Cloud Computing [7,8] and Edge/Fog [9–11] Computing focus on the data processing, and on the structure of the Central servers or Local devices. In the remainder of the paper, we will focus on the Internet of Things perspective, analysing available technologies and their adequacy in terms of implementation. Things / Sensors Cloud Computing Interconnection Things / Sensors Smartcities Edge / Fog Computing Cyber-Physical Systems Internet of Things Figure 1. Smart city structure. Internet of Things Interconnection Edge / Fog Computing Figure 1. Smart city structure. Figure 1. Smart city structure. Any city has several areas of concern to the authorities. 1. Introduction In general, air pollutants can be of two types: (i) primary air pollutants, which are gases or particles emitted directly into the atmosphere, including carbon monoxide (CO), carbon dioxide (CO2), particulate matter smaller than 10 microns (PM10), or particulate matter smaller than 2.5 microns (PM2.5); and (ii) secondary air pollutants, which are gases produced by a chemical reaction between 3 of 28 Sensors 2018, 18, 460 primary pollutants and some environment element, including ozone (O3), which is produced by the combination of nitrogen oxides (NOx), Oxygen (O2), Volatile Organic Compounds (VOC), and sunlight [15]. In addition, some pollutants are gases (i.e., CO2, CO, and Ozone), while others are suspended particles in the air (PM10, and PM2.5). Usually, the process of gas monitoring relies on chemical elements that react to the presence of these gases, while the monitoring of suspended particles mostly relies on optical detection procedures. By embracing the Smart City paradigm, crowdsensing becomes a solution able to cope with air pollution monitoring since this novel paradigm assumes that a signiﬁcant number of users perform collaborative sensing tasks, thereby collecting data from different populated locations while doing their daily activities. The collected data is periodically transmitted to a central server (Cloud) for data storage and processing. Overall, this strategy implies that the sensors used must be cheap and tiny enough for comfortable and easy transportation. Otherwise, it becomes hard to achieve a widespread distribution and adoption. Besides, there must be a communications link for transmitting the acquired data to a cloud-based server, where data are constantly being stored and processed. To fulﬁl the ﬁrst two requirements concerning price and size, we get advantage by using emerging prototyping platforms such as Raspberry Pi or Arduino, which allow achieving the aforementioned goals when combined with low-cost sensors available on the market. Such solutions, besides being inexpensive, have the advantage of using compact, powerful, and easy-to-use hardware that is widely adopted nowadays both in research and industry. These compact platforms can also be integrated in vehicles, i.e., cars, bikes, or public transportation, making it easy to monitor different air pollutants while the vehicle travels around the city. On the other hand, smartphones are nowadays widely adopted devices that have become ubiquitous. Thus, it becomes interesting to exploit their communication interfaces, such as Wi-Fi or Cellular, for transmitting the acquired data from mobile sensors. 1. Introduction With these issues in mind, mobiles sensors must be able to create a communication link with smartphones to make such interaction possible. Recent literature suggests that, although there are several hardware options for creating different types of sensors, there are none speciﬁc for air pollution monitoring from a crowdsensing perspective. Hence, it becomes necessary to analyse the different options for creating a small, low-cost, mobile sensor able to communicate with off-the-shelf smartphones in an Internet of Things type of environment. This paper is organised as follows: Section 2 presents a literature overview of crowdsensing projects. Section 3 provides an overview of the current IoT platforms and protocols that are able to support environment-sensing solutions under the crowdsensing paradigm. Section 4 presents an analysis of available architectures and technologies for creating a crowdsensing system while deﬁning the basic requirements for it. Section 5 presents a comparison of possible hardware conﬁgurations for creating a mobile sensor able to perform air monitoring, transmitting the collected data via some mobile sink (e.g., a smartphone) towards a datacenter for storage and processing. Section 6 studies the main requirements of this type of sensors, thoroughly assessing some developed conﬁgurations that fulﬁl the aforementioned requirements. Finally, Section 7 presents the conclusions of the present work. 2. Crowdsensing Architectures in the IoT Context: Literature Overview Following the Smart City paradigm, and focusing on the data collection domain, the concept of crowdsourcing has been introduced to refer to scenarios where a large group of people, through different devices and technologies, actively participate in the data acquisition process [16]. Once data are collected, they are sent to a central server for analysis, and feedback will eventually be returned to citizens through actions and services aimed at improving their life quality. Crowdsensing is a subtype of crowdsourcing where sensors are the actual sources of the data gathered [17]. If air quality sensors are used, crowdsensing becomes a good alternative to traditional stationary air quality stations whereby small sensors are distributed to a large group of people that seamlessly contribute to the system while doing their everyday tasks [18]. Sensors 2018, 18, 460 4 of 28 In the literature, we can ﬁnd several works related to crowdsensing systems applied to air monitoring. Air-cloud [19] is a system to monitor the concentrations of PM2.5 using crowdsensing in China. It focuses on the analysis of the mechanical sensor design to optimise air reception, as well as on data fusion techniques. The sensor calibration process relied on neural network tuning using laboratory measurements as input. U-Air [20] shows how to analyse the data obtained from different sources, such as trafﬁc levels, weather conditions, and pollution using various Big Data techniques, evidencing how these techniques allow inferring environmental pollution levels with better granularity. Researchers from Zurich University [21] have also designed a prototype to monitor ozone levels using a sensor that is connected to the smartphone via an USB cable. The Smart City project in Serbia [22] seeks similar goals, relying on a commercial sensor by Libelium [23] for measuring various air pollutants using the public transport system. Devarakonda et al. [24] propose a system with two types of sensors: (i) a sensor based on Arduino Mega128 to install in vehicles that costs about $700; and (ii) a personal sensor that is smaller than previous ones, and that costs about $400 for end-user applications. In this work, Google technology is used for data processing. Mead et al. [25] analyse the behaviour of electrochemical sensors to monitor air pollution in urban scenarios. They design two types of sensors (static and mobile) based on the PIC18F67J10 microchip, and they show how to deal with electrochemical sensors. Manna et al. 2. Crowdsensing Architectures in the IoT Context: Literature Overview [26] propose a sensor to monitor air pollution on roads, and to track vehicles which cause pollution, by using a sensor based on Arduino, an electrochemical CO gas sensor, and RFID technology. In addition, we can ﬁnd several technological approaches for fabricating small sensors. For instance, Hjiri et al. [27] use Al-doped ZnO (AZO) nanoparticles to create a highly sensitive CO gas sensor. Instead, Borini et al. [28] use graphene oxide to create humidity and temperature sensors with ultrafast response times (30 ms). Chen et al. [29] analyse the use of nanowires to fabricate gas sensors due to their characteristics: ultrasensitive, higher selectivity, low power consumption, and fast response. In addition, Zaatar et al. [30] show how to use the ﬁber optic evanescent wave to monitor air pollution. Even though there are different studies that provide a wide set of approaches for air pollution monitoring through crowdsensing, or that provide an isolate analysis of sensing technologies, there is no speciﬁc hardware solution that is widely available for regular users, and different alternatives have not yet been compared in a detailed manner. Thereby, an exhaustive analysis of the different solutions available to create a small, low-cost device endowed with both air quality sensors and communication capabilities, would represent a step forward in this challenging and fascinating area. 3. Internet of Things Protocols Overview In recent years, the Internet of Things (IoT) has become one of the most challenging research topics, offering a wide range of novel solutions for Smart Cities [3,31]. These proposals mainly analyse the intercommunication between devices, and involve a large variety of domains like home-based solutions [32], intelligent transportation systems [33], healthcare [34], safety and security [35,36], industrial control [37], and environmental monitoring. Thus, the analysis of the sensor design must be able to cope with IoT protocols, which will be described in this section. The main principle underlying the IoT paradigm is that all “things” are, or must be, connected to the Internet, and interact with each other to create developed areas that promote sustainability and a high life quality in multiple key areas [13,38]. The main characteristic of these things is that they are constrained devices such as small sensors, meaning they have restricted processing/storage capacity, restricted battery, restricted communication characteristics, i.e., Low Bandwidth, Low Data Rate, Low Coverage, etc. With this in mind, the Internet of Things has created a subset of protocols divided into various layers, similar to the traditional Internet 5 of 28 Sensors 2018, 18, 460 stack, but taking into account the restrictions of the Internet of Things in terms of processing, battery capacity, and communication capabilities of embedded devices. Figure 2 summarises the different layers deﬁned for IoT, and the differences towards the traditional Internet: (i) the Infrastructure Layer typically relies on wireless technologies, like ZigBee, LoRa, or Bluetooth Low Energy (BLE); (ii) the Addressing Layer focuses on the analysis of the addressing issues to achieve compatibility with Internet protocols; (iii) the Transport Layer is the same than for the TCP/IP (Internet) protocol stack, so either TCP or UDP are available, although UDP is typically used; (iv) the Messaging Layer deﬁnes protocols to transmit data towards the servers; (v) the Message Format Layer deﬁnes encoding types to store and transmit data; and (vi) the Semantic Layer deﬁnes the structure of the data. Below, we will describe the most important protocols involved in the Internet of Things at these different layers. 3. Internet of Things Protocols Overview Zigbee, LoRa, BLE, NFC, Sigfox, WiFi, LTE 6LowPAN UDP XML, JSON, EXI,MessagePack MQTT-SN, CoAP, XMPP-IoT Infrastructure Layer Addresing Layer Transport Layer Messaging Layer Message Format Layer Semantic Layer SensorML, Semantic Sensor Net Ontology WiFi, LTE IPv4/IPv6 TCP, UDP XML, JSON AMQT, MQTT, XMPP, Restful Several Ontologies and Schemas Internet of Things Internet Figure 2. IoT protocols. Figure 2. IoT protocols. 3.1. Infrastructure Layer Currently, there are several communication technologies for the Internet of Things. Notice that, while any communications technology would allow us to create a network for IoT, not all of them take device restrictions into account. Theoretically, 5G allows a lot of possibilities to be offered in the context of the Internet of Things [39]. Similarly, new technologies such as LoRA or SIGFOX allow network sensors to remain connected due to their large coverage. Below, we make a brief analysis of the possible wireless technologies for IoT. 5G Network [40] is the ﬁfth generation cellular network architecture, designed to support great amounts of data, high speed, conﬁgurability, etc. from new emerging technologies such as Internet of Things. Currently, it is in the ﬁrst phase, where new standards and services will be deﬁned, but soon it shall become the default cellular network technology. ZigBee [41] is based on the IEEE 802.15.4 standard, and it was designed for Wireless Sensor Networks. Its main characteristics are its small size and low power consumption. Usually, the transmission range can vary from 10 to 100 m, depending on the output power. The main drawback of this technology, though, is that current smartphones are not equipped with ZigBee interfaces. ZigBee [41] is based on the IEEE 802.15.4 standard, and it was designed for Wireless Sensor Networks. Its main characteristics are its small size and low power consumption. Usually, the transmission range can vary from 10 to 100 m, depending on the output power. The main drawback of this technology, though, is that current smartphones are not equipped with ZigBee interfaces. Wi-Fi [42] is based on the IEEE 802.11 standard, and it was designed for Wireless Local Area Networks. The evolution of this technology provides several variants operating in the 2.4 GHz or in Wi-Fi [42] is based on the IEEE 802.11 standard, and it was designed for Wireless Local Area Networks. The evolution of this technology provides several variants operating in the 2.4 GHz or in Sensors 2018, 18, 460 6 of 28 the 5 GHz band, being currently the 802.11n version the most widespread option. The transmission range for standard interfaces is about 100 m. 3.1. Infrastructure Layer LoRa [43] is a LPWAN (Low-Power Wide-Area Network) technology designed to optimise different aspects such as communication range, battery lifetime, and costs, supporting thousands of devices headed for the Internet of Things in several domains including sensing, metering, and machine-to-machine (M2M) communications. Theoretically, LoRa achieves a transmission range of more than 15 km in rural environments, and of more than 2 km in dense urban areas. Its bandwidth ranges between 250 bps and 50 Kbps in different frequencies: 169 MHz, 433 MHz, and 868 MHz in Europe, and 915 MHz in North America. q p SIGFOX [44] is an emerging technology that offers a proprietary telecommunications network to support Internet of Things solutions. It was designed for LPWA (Low-Power Wide-Area) networks operating in the ISM 868 MHz band, reaching distances greater than 1 km. Since the selected ISM band is restricted, the communications could be of up to 12 bytes per message, and up to 140 messages per day. Near Field Communications (NFC) [45] was designed for communications between two nearby devices (closer than 4 cm). It main target applications are smartphone-based payments and IoT solutions such as access control, or inventory systems. However, its distance requirements and intermittent connectivity features make it a poor option for our purposes. Bluetooth [46] was designed for Personal Area Networks, purposely having a maximum coverage range of 10 meters by default. Currently, it is used for transmitting information between personal devices, such as smartphones, smartwatches, and headsets. Bluetooth Low Energy [47], or Bluetooth Smart, is the name under which Bluetooth version 4 is known. Its main advantage, when compared to previous versions, is using very low power, being nowadays one of the best options for IoT applications. Similarly to previous Bluetooth speciﬁcations, the coverage range is of 10 m. 3.2. Addressing Layer The addressing layer deﬁnes the logic address of the packets by assigning a speciﬁc address to all possible nodes. These protocols deal with the packet forwarding problem, which in the TCP/IP model is handled by IPv4 and/or IPv6. 6lowPAN [48] (IPv6 over Low power Wireless Personal Area Networks) allows using IPv6 over networks based mainly in the IEEE 802.15.4 standard. It is designed for resource-constrained devices by reducing the size of the address to 64 or 16 bits, depending on whether it is for a Local Network or a Personal Area Network, respectively; it uses default values for specifying the network. 3.3. Messaging Layer The messaging layer deﬁnes protocols for data transmission systems considering IoT restrictions. RESTf l [49] (R i l f ) i W b b d hi h The messaging layer deﬁnes protocols for data transmission systems considering IoT restrictions. RESTful [49] (Representational state transfer) is a Web-based architecture to exchange or to The messaging layer deﬁnes protocols for data transmission systems considering IoT restrictions. RESTful [49] (Representational state transfer) is a Web-based architecture to exchange or to manipulate Web resources through a textual representation using preset stateless operations; it means that each message must have all the required information to complete the request. It follows a client/server model based on the HTTP protocol, and relies on its functions: (i) GET, to retrieve a resource; (ii) POST, to create a resource; (iii) PUT, to change the state of a resource; and (iv) DELETE, to delete a resource. The data representation typically adopts the XML or JSON formats. Figure 3 presents a basic overview of a Restful Architecture under the client/server model. 7 of 28 Sensors 2018, 18, 460 Client Server (1) HTTP request (2) HTTP response Figure 3. Basic RESTful architecture. Figure 3. Basic RESTful architecture. There is a speciﬁcation [50] for constrained nodes and networks called Constrained RESTful Environments (CoRE) Link Format. It speciﬁes a set of links to discover resources, and to access these resources in a Machine-to-Machine (M2M) environment. MQTT [51] (Message Queue Telemetry Transport) is a lightweight messaging protocol based on the publisher/subscriber scheme that runs on top of the TCP/IP protocol. It is also designed for constrained networks with limited bandwidth. The MQTT is composed by three elements: publishers, subscribers, and a message broker. A subscriber, which wants to receive a message related to a speciﬁc topic, must subscribe to the message broker; next, when a publisher sends/publishes a message related a certain topic, it is transmitted to all subscribers subscribed to this topic. MQTT has three transmission Quality of Service (QoS) levels: (i) QoS 0: At most once. The message is sent once, but it does not check for ACKs to conﬁrm message reception. (ii) QoS 1: At least once. The message could be sent more than once to each subscriber. (iii) QoS 2: Exactly once. The message is sent exactly once using four-way handshaking. 3.3. Messaging Layer MQTT-SN [52] (MQTT Sensor Network) has been designed to be similar to MQTT, but considering the restrictions of wireless communication environments, such as limited bandwidth, short message length, etc., running over UDP or on Non-IP environments. For interoperating with standard MQTT environments, it needs an MQTT-SN Gateway which connects MQTT-SN nodes, such as constrained sensors, to the MQTT network. Figure 4 shows a basic MQTT architecture. Broker Susbcriber (1) sub(topic) (3) send(data) Publisher (2) pub(topic,data) MQTT-SN Gateway MQTT-SN Client MQTT-SN Environment Figure 4. Basic MQTT (Message Queue Telemetry Transport) architecture. MQTT-SN Gateway MQTT-SN Client Figure 4. Basic MQTT (Message Queue Telemetry Transport) architecture. CoAP [53] (Constrained Application Protocol) is a generic web protocol designed for constrained environments with restricted network capacities and restricted devices, allowing these devices to communicate with the Internet or other constrained devices. It implements a compressed subset of the REST model implementing GET, POST, PUT, and DELETE operations over UDP. CoAP reduces the message header and restricts message exchange, reducing the network overhead. It is very useful for Machine-to-Machine (M2M) communication, and for the Internet of Things (IoT). g CoAP can be easily translated to HTTP for seamless integration with existing Web systems, while reducing network requirements. XMPP [54] (Extensible Messaging and Presence Protocol) is a message-oriented communications protocol based on XML. Initially, it was called Jabber, and it was designed for instant messaging (IM). It allows federations among various XMPP servers, and even communication with different 8 of 28 Sensors 2018, 18, 460 technologies using XMPP gateways. Currently, it is also used for VoIP, video, gaming, or even for IoT applications. Figure 5 shows the basic architecture of an XMPP-based system. technologies using XMPP gateways. Currently, it is also used for VoIP, video, gaming, or even for IoT applications. Figure 5 shows the basic architecture of an XMPP-based system. XMPP Server XMPP Client XMPP Server Other Technology XMPP Client XMPP Client XMPP Gateway Figure 5. Basic XMPP (eXtensible Messaging and Presence Protocol) architecture. XMPP Client Other Technology XMPP Gateway XMPP Client Figure 5. Basic XMPP (eXtensible Messaging and Presence Protocol) architecture. The speciﬁcation for IoT is XEP-0323: Internet of Things—Sensor Data [55], which provides the architecture, basic operations, and data structures for sensor data communication, including a hardware abstraction model for the interconnection of constrained devices. 3.3. Messaging Layer sMAP [56] (Simple Measuring and Actuation Proﬁle) is an example of how RESTful web services can be simpliﬁed, while still allowing instruments and other producers of physical information to directly publish their data in a central server. 3.4. Message Format Layer The Message Format layer presents all data encoding types to store and transmit structured data for IoT applications. XML [57] (eXtensible Markup Language) is a markup language for encoding documents in a text format that is understandable by both human and machines. XML is designed to store data units called entities, where all data structures and document descriptions are achieved through markups. Using these markups, it is possible to create any logical data structure in an easy way. JSON [58] (JavaScript Object Notation) is a format notation to encode structured data (attribute-value pair or array) using human-readable text. It was designed to replace XML by reducing its complexity. It is very common in web systems, especially in AJAX-style ones. It uses pairs (object_id:object_value) and brackets to provide complex object structuring for ﬁtting data in a text document. EXI [59] (Efﬁcient XML Interchange) is a binary and compact representation of XML or JSON documents. It aims at resource-constrained devices and networks, attempting to reduce the size of the data and the computational requirements when compared to other compressors such as gzip. The EXI coder is based on events, and it follows a simpliﬁed Huffman coding to create a binary document. MessagePack [60] is a binary serialization format that encodes messages faster and in a more compact manner than traditional methods such as JSON or XML. This is possible since small integer values are encoded in a single byte, while strings require only an extra byte to identify them. This simpliﬁes the encoding process, but it has some limitations, such as the size of strings or numbers, the number of the key/value association map, etc. Table 1 shows a representation of sensor data using XML and JSON encoding, and Table 2 shows a representation of sensor data using EXI and MessagePack encoding, respectively. We can observe that the EXI encoding and the MessagePack are much smaller (163 and 186 bytes) than the encoding achieved using XML (474 bytes) or JSON (357 bytes). 9 of 28 Sensors 2018, 18, 460 Table 1. Example of data representation using typical encoding types. 3.5. Semantic Layer The Semantic layer presents all approaches that describe a logical representation of things in the IoT context. Therefore, in sensing approaches, we can ﬁnd several representative examples: SensorML [61] is an standard model based on XML encoding for describing sensors and measurement processes. It is developed by the Open Geospatial Consortium, describing a wide range of sensors for different types of architectures, including remote sensors, in-situ sensors and dynamic sensors, among others. Semantic Sensor Net Ontology [62] describes sensors and observations, avoiding to describe domain concepts, location, time, etc. It is developed by the W3C Semantic Sensor Networks Incubator Group (SSN-XG). Web of Things [63] speciﬁes a data model to describe physical devices connected to the Web (Internet) using JSON encoding. It was created for the Mozilla project, and was formally submitted to W3C for discussion. Taking into account that the previously described protocols, a sensor device must be able to cope with a subset of them to allow the exchange of data between the sensors and a central server. 3.4. Message Format Layer EXI Encoding (163 bytes) MessagePack Encoding (186 bytes) 80 40 67 47 26 16 36 5a 80 24 06 d2 c9 50 08 84 3a 39 30 b1 b2 98 80 ee cc 2d 8e ac ae 75 00 48 25 8d 85 c1 d1 d5 c9 95 ce a0 00 00 96 c6 17 46 97 47 56 46 5a 80 44 2c cc e4 b8 d0 dc c0 d4 dc dc 0a 6c 6 f 6e 67 69 74 75 64 65 a8 04 43 0b 4c 0b 8c cc cc cd 8d 8c 0d 00 66 f7 a6 f6 e6 5a 80 44 10 d4 d9 00 0c 02 cc ce 4b 8d 0d cc 0d 8d 4c 80 0c 83 0b 4c 0b 8c cc cd 0c cc cd 8d 40 0d 01 0d 8e 10 00 c0 2c cc e4 b8 d0 dc c0 e0 e4 c8 00 c8 30 b4 c0 b8 cc cc d4 e4 e4 e0 dc 00 d0 10 d4 e5 ea 80 81 a5 74 72 61 63 65 82 a2 69 64 a6 74 72 61 63 65 31 a6 76 61 6c 75 65 73 81 a8 63 61 70 74 75 72 65 73 93 83 a8 6c 61 74 69 74 75 64 65 a9 33 39 2e 34 37 30 35 37 37 a9 6c 6 f 6e 67 69 74 75 64 65 aa 2d 30 2e 33 33 33 36 36 30 34 a5 6 f 7a 6 f 6e 65 a2 35 36 83 a8 6c 61 74 69 74 75 64 65 a9 33 39 2e 34 37 30 36 35 32 a9 6c 6 f 6e 67 69 74 75 64 65 aa 2d 30 2e 33 33 34 33 33 36 35 a5 6 f 7a 6 f 6e 65 a2 36 38 83 a8 6c 61 74 69 74 75 64 65 a9 33 39 2e 34 37 30 38 39 32 a9 6c 6 f 6e 67 69 74 75 64 65 aa 2d 30 2e 33 33 35 39 39 38 37 a5 6 f 7a 6 f 6e 65 a2 35 39 p p g p ( y) g yp EXI Encoding (163 bytes) MessagePack Encoding (186 bytes) 80 40 67 47 26 16 36 5a 80 24 06 d2 c9 50 08 84 3a 39 30 b1 b2 98 80 ee cc 2d 8e ac ae 75 00 48 25 8d 85 c1 d1 d5 c9 95 ce a0 00 00 96 c6 17 46 97 47 56 46 5a 80 44 2c cc e4 b8 d0 dc c0 d4 dc dc 0a 6c 6 f 6e 67 69 74 75 64 65 a8 04 43 0b 4c 0b 8c cc cc cd 8d 8c 0d 00 66 f7 a6 f6 e6 5a 80 44 10 d4 d9 00 0c 02 cc ce 4b 8d 0d cc 0d 8d 4c 80 0c 83 0b 4c 0b 8c cc cd 0c cc cd 8d 40 0d 01 0d 8e 10 00 c0 2c cc e4 b8 d0 dc c0 e0 e4 c8 00 c8 30 b4 c0 b8 cc cc d4 e4 e4 e0 dc 00 d0 10 d4 e5 ea 80 81 a5 74 72 61 63 65 82 a2 69 64 a6 74 72 61 63 65 31 a6 76 61 6c 75 65 73 81 a8 63 61 70 74 75 72 65 73 93 83 a8 6c 61 74 69 74 75 64 65 a9 33 39 2e 34 37 30 35 37 37 a9 6c 6 f 6e 67 69 74 75 64 65 aa 2d 30 2e 33 33 33 36 36 30 34 a5 6 f 7a 6 f 6e 65 a2 35 36 83 a8 6c 61 74 69 74 75 64 65 a9 33 39 2e 34 37 30 36 35 32 a9 6c 6 f 6e 67 69 74 75 64 65 aa 2d 30 2e 33 33 34 33 33 36 35 a5 6 f 7a 6 f 6e 65 a2 36 38 83 a8 6c 61 74 69 74 75 64 65 a9 33 39 2e 34 37 30 38 39 32 a9 6c 6 f 6e 67 69 74 75 64 65 aa 2d 30 2e 33 33 35 39 39 38 37 a5 6 f 7a 6 f 6e 65 a2 35 39 ng (163 bytes) MessagePack Encoding (186 bytes) 3.4. Message Format Layer XML Encoding (474 bytes) JSON Encoding (357 bytes) <?xml version ="1.0" encoding ="UTF−8" ?> <trace > <id>trace1 </id> <values > <captures > <latitude >39.470577 </ latitude > <longitude >−0.3336604</ longitude > <ozone>56</ozone> </captures > <captures > <latitude >39.470652 </ latitude > <longitude >−0.3343365</ longitude > <ozone>68</ozone> </captures > <captures > <latitude >39.470892 </ latitude > <longitude >−0.3359987</ longitude > <ozone>59</ozone> </captures > </values > </trace > { " trace " : { " id " : " trace1 " , " values " : { " captures " : [ { " l a t i t u d e " : "39.470577" , " longitude " : " −0.3336604" , " ozone " : "56" } , { " l a t i t u d e " : "39.470652" , " longitude " : " −0.3343365" , " ozone " : "68" } , { " l a t i t u d e " : "39.470892" , " longitude " : " −0.3359987" , " ozone " : "59" } ] } } } Table 1. Example of data representation using typical encoding types. Table 1. Example of data representation using typical encoding types. XML Encoding (474 bytes) Table 2. Example of data representation using compressed (binary) encoding types. Table 2. Example of data representation using compressed (binary) encoding types. Table 2. Example of data representation using compressed (binary) encoding types. 4. Mobile Sensing Requirements Normally, the sensing process is made through a Wireless Sensor Network (WSN) [64,65], which is composed by a set of nodes or sensors that collect data and send it towards a central sink or gateway. The latter carries out processing tasks, or merely resends collected data to a server for storage and Sensors 2018, 18, 460 10 of 28 processing. Usually, all sensors are resource-constrained devices, and the central sink or gateway has fewer restrictions, often being connected to the power network. Currently, most Wireless Sensor Networks are based on the IoT architecture since it allows us to work with constrained devices and restricted networks. Figure 6 shows the basic structure of a Wireless Sensor Network. We can see that the communications link between the sensors and the sink/gateway is wireless, typically relying on ZigBee, and that the communications link between the sink/gateway and the central server is typically a more robust link, either wired (e.g., Ethernet) or wireless (e.g., Wi-Fi or Cellular). Version January 31, 2018 submitted to Sensors 10 of 27 WSN Sensor Node Sink Central Server Figure 6. Wireless Sensor Network structure. processing. Usually, all sensors are resource-constrained devices, and the central sink or has fewer restrictions, often being connected to the power network. Currently, most of the Sensor Networks are based on the IoT architecture since it allows us to work with constrained and restricted networks. WSN Sensor Node Sink Central Server Central Server rrently, mo ork with co Figure 6. Wireless Sensor Network structure. or Sink Ce S The support for mobility in a Wireless Sensor Network [66,67] can be achieved through different strategies, including sensor mobility, as shown in Figure 7a, or by having mobility on both sensors and gateway, as shown the Figure 7b. Finally, we have crowdsensing architectures where the gateway and the sensor are the same, or are packed together. Commonly, the best way to implement crowdsensing is through smartphones, since nearly all people carry one with them nowadays, and they are endowed with several sensors and communication interfaces. Figure 8 shows an example of a crowdsensing architecture where a smartphone is used as the gateway between the sensor and a Central Server. Figure 6. Wireless Sensor Network structure. Figure 6 shows the basic structure of a Wireless Sensor Network. 4. Mobile Sensing Requirements We can see that the 325 communications link between the sensors and the sink/gateway is wireless, typically relying on 326 Zigbee, and that the communications link between the sink/gateway and the central server is 327 typically a more robust link, either wired (e.g. Ethernet) or wireless (e.g. Wi-Fi or Cellular). 328 y g g Mobile Sensor Node Gateway Central Server Static (a) Mobile Wireless Sensor Network with mobile sensors. Static Mobile Mobile Sensor Node Gateway Central Server (b) Mobile Sensor Network with mobile sensors and mobile gateway. Static Mobile Sensor Node Gateway Central Server (c) Crowdsensing architecture. Figure 7. Different types of Mobile Sensor Networks’ structures. Figure 7. Different types of Mobile Sensor Networks’ structures. Mobile Sensor Node Gateway Central Server Static (a) Mobile Wireless Sensor Network with mobile sensors. Mobile Sensor Node Gateway Central Server Static (a) Mobile Wireless Sensor Network with mobile sensors. (a) Mobile Wireless Sensor Network with mobile sensors. (a) Mobile Wireless Sensor Network with mobile sensors. Static Mobile Mobile Sensor Node Gateway Central Server (b) Mobile Sensor Network with mobile sensors and mobile gateway. Mobile Mobile Central Server Sensor Node Gateway (b) Mobile Sensor Network with mobile sensors and mobile gateway. ( ) g y Static Mobile Sensor Node Gateway Central Server (c) Crowdsensing architecture. Figure 7. Different types of Mobile Sensor Networks’ structures. Figure 7. Different types of Mobile Sensor Networks’ structures. Static Mobile Sensor Node Gateway Central Server (c) Crowdsensing architecture. Figure 7. Different types of Mobile Sensor Networks’ structures. Figure 7. Different types of Mobile Sensor Networks’ structures. Central Server Gateway (c) Crowdsensing architecture. Figure 7. Different types of Mobile Sensor Networks’ structures. Figure 7. Different types of Mobile Sensor Networks’ structures. The support for mobility in a Wireless Sensor Network [66,67] can be achieved through different 329 strategies, including sensor mobility, as shown in Figure 7a, or by having mobility on both sensors 330 and gateway, as shown the Figure 7b. Finally, we have crowdsensing architectures where the 331 Crowdsensing solutions need to be widely disseminated and adopted by users to be successful. In addition, to achieve such widespread acceptance, the impact on the users’ everyday activities 11 of 28 Sensors 2018, 18, 460 must be low. This means that any deployed application must operate in the background and avoid consuming excessive resources, while requiring only a minimal user participation. 4. Mobile Sensing Requirements Sensors 2018, 18, 460 12 of 28 Sampling process refers to the process of capturing pollutant measurements, including the calibration process, where electrical signals are translated to pollution units, ﬁlters, fault detection and diagnosis, etc. Sensor calibration, in Commercial Off-The-Shelf (COTS) sensors such as electrochemical ones, is a process that depends on the physical sensor characteristics, temperature, etc. Basically, electrical outputs must be translated to pollution units, and often there is no lineal relation. The calibration is commonly made in advanced laboratories, taking into account samples taken with different pollution levels and for different temperatures and humidity conditions [69,70]. However, in urban scenarios, the auto calibration procedure is too complicated because all sensors are distributed among different users. Alvear et al. [71] proposed a method to calibrate off-the-shell sensors using mathematical regressions based on high-accuracy samples obtained through the ﬁxed stations deployed in a city. Once a mobile sensor is near to these stations, these samples are used to adjust the translation equation (electrical signal to pollution values). Using COTS, the sampling error and its diagnostics can be a problem [71]. Nevertheless, when focusing on a Crowdsensing solution using a large number of mobile sensors (smart city scenario), this problem could be solved by accounting for redundant data and statistical analysis (i.e., Kriging allows us to deal with sampling error). Filtering process refers to deleting redundant and/or wrong measurements caused by the sensors reading oscillations. By using mobile sensors, the ﬁltering process also has to deal with temporal variations, as describe in [68,71], adjusting samples to a same temporal fragment. 66.0 68.0 70.0 72.0 74.0 76.0 O3 50 70 90 110 16.0 16.5 17.0 17.5 18.0 Hour O3 [ppb] Ozone (a) Sampling process (b) Filtering process (c) Data processing Experimental variogram and fitted variogram model Distance Semi−variance 0.2 0.4 0.6 0.8 1.0 1.2 0.001 0.002 0.003 511747 8051380 15321748 7499 11412 18183 Model: Gau Nugget: 0.000 Sill: 1.001 Range: 0.002 (d) Results presentation Figure 10. Data handling process as described in [68]: (a) sampling process; (b) ﬁltering process after adjusting the temporal variations; (c) data analysis using a semivariogram of the captured data used for interpolating the entire area using the Kriging technique; and (d) pollution distribution map. 4. Mobile Sensing Requirements Concerning the sensor itself, if external to the smartphone, it should be cheap, small, easy to use, and comfortable to carry. Process Storage BLE Man. Rest API Smartphone Sta c Mobile Sensor Analog Sensor Microcontroller BLE Manager BLE WiFi / LTE Server Process Storage Rest API Present Figure 8. Crowdsensing architecture overview. Figure 8. Crowdsensing architecture overview. Crowdsensing approaches have two basic architectural components [68]: a mobile component for the data acquisition process, and a central server for data storage and processing. The mobile component must be able to collect environmental parameters, transmitting them towards the central server. The data acquisition process is based on smartphone sensors, or on small external sensors accessible via smartphone, and the transmission process usually relies on smartphone connectivity towards the Internet. Despite delegating transmission tasks on smartphones, external sensing devices must still be endowed with communication capabilities to transfer the collected data to the smartphone. Thus, the sensing device should be equipped with a wireless communications interface, being technologies such as Wi-Fi, Bluetooth, RFID, NFC, and ZigBee good candidate solutions. The central processing server must be able to receive the transmitted data from the sensors, store and process the data, as well as properly present the obtained results to system managers. In addition, in some cases, they perform remote communication with the mobile devices for conﬁguration tasks, thereby allowing to dynamically change the sensing behaviour. Taking the aforementioned considerations into account, Figure 8 shows a basic hardware architecture applicable to air quality sensing applications that should include: (i) a mobile sensor; (ii) a smartphone; and (iii) a central server. The proposed architecture resembles various approaches from different authors [19–21,68]. Moreover, as shown in Figure 9, the crowdsensing process basically includes ﬁve different tasks: (i) sampling process; (ii) ﬁltering process; (iii) data transfer; (iv) data processing; and (v) results presentation. Notice that all these tasks could be done by different hardware components; for instance, the ﬁltering task could be done by the sensor, the smartphone, or even the central server, depending on the system characteristics. Moreover, characteristics associated to sensing, ﬁltering, and transmission tasks could be deﬁned based on parameters obtained from the processing step. Sampling Process Filtering Process Data Transfer Data Processing Results Presentation Configure Configure Figure 9. Crowdsensing steps. Results Presentation Data Processing Data Transfer Figure 9. Crowdsensing steps. 4. Mobile Sensing Requirements 66.0 68.0 70.0 72.0 74.0 76.0 O3 (a) Sampling process 50 70 90 110 16.0 16.5 17.0 17.5 18.0 Hour O3 [ppb] Ozone O3 [ppb] (b) Filtering process (a) Sampling process g p (d) Results presentation (c) Data processing Experimental variogram and fitted variogram model Distance Semi−variance 0.2 0.4 0.6 0.8 1.0 1.2 0.001 0.002 0.003 511747 8051380 15321748 7499 11412 18183 Model: Gau Nugget: 0.000 Sill: 1.001 Range: 0.002 Experimental variogram and fitted variogram model (d) Results presentation Figure 10. Data handling process as described in [68]: (a) sampling process; (b) ﬁltering process after adjusting the temporal variations; (c) data analysis using a semivariogram of the captured data used for interpolating the entire area using the Kriging technique; and (d) pollution distribution map. Sensors 2018, 18, 460 13 of 28 Data transfer process refers to the upload of data from the sensor to the cloud (Central servers), including sensor-smartphone and smartphone-server communications. In is achieved through the previously described IoT protocols. Data processing refers to the interpolation technique used to recreate a pollution distribution map. It could be made by different methods (Kriging, IDW, and Nearest neighbour Spatial Averaging) as described in [72]. Currently, the most used method is the Kriging interpolation technique, where a semivariogram is calculated for create a complete pollution map. Results presentation refers to the way results are presented to the system administrator. The mos ul representation is a graphical map for the target region. Figure 10 presents the data handling process, as the authors described in [68], showing the four processes for handling pollution information in order to recreate a complete pollution map for a certain target area. 4.1. Basic Smartphone Software Architecture Concerning smartphones, they are devices widely used nowadays for any task, and characterised by powerful computing capabilities, large amounts of memory, and several embedded sensors and communication interfaces [73]. We consider smartphones as the best gateway option for connecting mobile sensors with a central server. In addition, they can perform CPU-intensive tasks such as data ﬁltering or data fusion, simplifying sensor requirements and design to mere data acquisition and data relaying towards the smartphone. Since the smartphone must act as a gateway between sensor and cloud server, it must manage at least two network interfaces: one to collect data from the sensor (Sensing middleware), and another one to upload data to a central server (Cloud middleware). Although both tasks must run independently, the data uploading process is often not made in real time, contrarily to the sensor data collection process, which is a task that should be done very frequently, especially if we aim at a simpliﬁed sensor design, as shown in Figure 11. Moreover, modules to process and store the collected data are also needed. By assuming that the smartphone becomes responsible for all the computationally intensive tasks, it becomes necessary to analyse which are the actual basic requirements when designing a mobile sensor for crowdsensing solutions. Smartphone Internet Connection Bluetooth Process Filtering Adjusting Calibrating Formating Sensing Middleware Cloud Middleware Storage Figure 11. Smartphone software architecture overview. Figure 11. Smartphone software architecture overview. 4.2. Functional Requirements of the Mobile Sensor Despite relying on smartphones for providing system intelligence, basic mobile sensor requirements still involve: Processing: The sensor must be able to process the measured data, perform basic ﬁltering tasks, and transfer data to an external device. Anyway, in terms of processing power, requirements are low. Sensors 2018, 18, 460 14 of 28 14 of 28 Storage: By assuming that a links towards a smartphone or a similar device is available, the sensor does not need to actually store large amounts of collected data. In fact, since data can be seamlessly relayed to the smartphone in real time, the sensor can limit its internal storage to only a few samples. y p g y p Communication: Sensors do not need to have a direct connection to the cloud server via Internet, but they still need to transfer the collected data to the smartphone. Thus, sensors require a communications link compatible with current smartphone technologies like Wi-Fi, Bluetooth, or NFC. Autonomy: Sensors must be able to operate for long periods using a small power supply. Thus, ti i ti b k i t t t k d t f ll b tt i Communication: Sensors do not need to have a direct connection to the cloud server via Internet, but they still need to transfer the collected data to the smartphone. Thus, sensors require a communications link compatible with current smartphone technologies like Wi-Fi, Bluetooth, or NFC. Autonomy: Sensors must be able to operate for long periods using a small power supply. Thus, energy optimisation becomes a key requirement to take advantage of small batteries. Size: Sensors need to be transported by users, or to be quickly installed in vehicles (e.g., bicycle, motorcycles, and cars). Thus, they must be small enough for the sake of aesthetics and comfort. Price: To be attractive to users, sensors must be as cheap as possible. Otherwise, it becomes difﬁcult to meet the broad dissemination requirements of crowdsensing approaches. 4.3. Basic Mobile Sensor Design To fulﬁl the technical requirements, a basic mobile sensor should be composed of a sensor device able to monitor the differences between different pollutant levels, a communications module for transferring the data collected, and a microcontroller/microcomputer acting as a central element for managing all tasks. Figure 12 shows a basic mobile sensor design, and the main characteristics to consider. As shown, the sensor hardware module must be able to connect to a microcontroller/microcomputer, a connection that typically relies on an analog/digital port. Similarly, the communications module must also be connected to the microcontroller via an UART or USB port. Thus, the microcontroller becomes a central element in the sensing module, being responsible for managing the interactions between all the elements. Communica on Module Sensor Module Sensor Node Firmware / SO Applica ons Microcontroller Processor Flash/RAM Analog I/O Digital I/O Ba ery Analog I/O Digital I/O Figure 12. Mobile sensor design. Digital I/O Communica on Module Figure 12. Mobile sensor design. Overall, the sensing module must be equipped with different analog ports, UART/USB ports, a processor, and ﬂash memory (ROM or RAM). 5. Overview of Available Hardware and Software In recent years, the appearance of different embedded prototyping platforms, such as Raspberry Pi or Arduino, which are complemented by a large variety of compatible electronic components, paved the way for the creation of diverse applications related to IoT. When speciﬁcally focusing on environmental monitoring requirements, we ﬁnd that there are different development options, including different types of sensor, and various communication interfaces. Commercially, several companies offer small and yet powerful boards, along with a large variety of electronic components for personalising them according to user needs. In addition, there are various companies offering extra modules or add-ons such as Seeedstudio [74], which offers their own Sensors 2018, 18, 460 15 of 28 15 of 28 sensors for developing personalised frameworks based on Grove technologies. Similarly, Adafruit [75] provides different embedded platforms, as well as all kinds of electronic components, including personalised add-ons for batteries, communication modules, and sensor boards compatible with the most widely extended platforms: Raspberry Pi, Arduino, BeagleBone, Intel Edison, or Intel Galileo. Focusing on ﬁnal solutions, TST [76] offers products in the ﬁeld of Smart Cities (Waste Management, Industrial Control, Light Control, etc.), basing their solutions on their own hardware platform. Likewise, Libelium [23] is a company providing various products in the ﬁeld of monitoring (Environment monitoring, Agriculture, Water monitoring, etc.); most of the components, and the programming tool used, are based on the Arduino platform. However, from a crowdsensing perspective, the solutions offered are inadequate due to the relatively large sizes of the devices, being mostly oriented for public infrastructure deployment. Based on the state of the art, we now provide an in-depth analysis of the different hardware and software components applicable to our mobile air quality sensing context. To enable air quality data acquisition, specialised pollution sensors must be connected to a microcontroller or microcomputer via an analog or digital port. Moreover, for communication tasks, the microcontroller/microcomputer must be connected to the communications module via an USB port or UART interface. In this sense, available options will depend on the microcontroller/ microcomputer characteristics. 5.1. Microcontroller/Microcomputer-Based Embedded Systems Despite the lightweight processing constraints, there are several options available for embedded systems acting as central elements in the sensor design. In fact, it is possible to use microcomputers such as Raspberry Pi, BeagleBone, or Intel Edison, which use a standard operating system, and that allow developing applications for sensing tasks in a straightforward manner. Alternatively, it is possible to use a microcontroller board such as Arduino, and develop application-speciﬁc ﬁrmware instead. Raspberry Pi is one of the most popular microcomputers worldwide. It is a low-cost and small-sized computer that allows connecting standard PC peripherals including a monitor, a keyboard, and a mouse. It was designed to explore computing, and it supports different Operating Systems: Raspbian, which is based on Debian, and also Ubuntu Mate or Windows 10 IoT Core, thereby allowing to use several programming languages. In addition, all Raspberry Pi versions beneﬁt from several input/output ports operating at 5 Volts, thus being ideal for all sorts of IoT projects. There are different versions of the Raspberry Pi, as shown Figure 13, being model A and type 2 the most commonly used. They have different characteristics, e.g., Type B offers better performance in terms of memory and processing, but Model A consumes much less power. Recently, Raspberry Pi 3 has appeared, offering better features than previous versions, being the major difference the integration of a Bluetooth and a Wi-Fi module, thereby facilitating communication tasks in the scope of IoT projects. Figure 13. Overview of different Raspberry Pi models. Figure 13. Overview of different Raspberry Pi models. BeagleBone is a small computer running a Linux Operating System called Angstrom, and supporting various software distributions such as Android or Ubuntu. It has an USB port 16 of 28 16 of 28 Sensors 2018, 18, 460 for connecting distinct peripherals, along with an HDMI port for video connection, allowing to use it as a regular computer. It has two 46 pin headers which operate at 3.3 V, allowing to augment the available functionalities by connecting different digital or analog devices like sensors or actuators. y g g g Commercially, we can ﬁnd several Beaglebone versions, being Beaglebone Black the most commonly used (see Figure 14). Figure 14. Beaglebone Black microcomputer overview. Figure 14. Beaglebone Black microcomputer overview. Intel Edison is a tiny but powerful computer developed by Intel. 5.1. Microcontroller/Microcomputer-Based Embedded Systems It is designed for IoT applications, targeting at both prototypes and commercial solutions with performance constraints. It supports a modiﬁed Linux distribution (Yocto) as its Operating System, and it integrates both Wi-Fi and Bluetooth 4.0 interfaces. In addition, it can beneﬁt from two types of expansion board: an Arduino Expansion board, and a mini breakout board (see Figure 15). Figure 15. Intel Edison with two extension boards. Pycom [77] is a microcontroller based on the ESP32 chip with 24 GPIO pins, 2 UARTs, 1 SPI and 1 I2C port, using a ﬁrmware based on micropython. It can be equipped with several communication interfaces such as Wi-Fi, Bluetooth Low Energy, LoRA, and Sigfox. Moreover, using an expansion board, it can integrate an SD Card, as well as different sensors such as Gyroscope, Accelerometer, or GPS (see Figure 16). Figure 15. Intel Edison with two extension boards. Figure 15. Intel Edison with two extension boards. Pycom [77] is a microcontroller based on the ESP32 chip with 24 GPIO pins, 2 UARTs, 1 SPI and 1 I2C port, using a ﬁrmware based on micropython. It can be equipped with several communication interfaces such as Wi-Fi, Bluetooth Low Energy, LoRA, and Sigfox. Moreover, using an expansion board, it can integrate an SD Card, as well as different sensors such as Gyroscope, Accelerometer, or GPS (see Figure 16). 17 of 28 Sensors 2018, 18, 460 Figure 16. Pycom module. Figure 16. Pycom module. Arduino Uno is an open source prototyping platform characterised by easy-to-use hardware and software. It has several analog and digital input/output pins to connect sensors, actuators or complementary boards, allowing to create a wide variety of IoT solutions. Arduino has its own programming language based on Wiring, and its own Arduino Software based on Processing [78]. It has a central microcontroller, and an USB port for programming and to supply power. Arduino nano is a tiny prototyping platform that maintains the Arduino Uno concept, using the same programming languages and the same libraries. It was also developed for prototyping solutions, but it is smaller than the standard Arduino, and it has less available memory (see Figure 17). Figure 17. Arduino Nano module (right) and Arduino Uno module (left). Figure 17. Arduino Nano module (right) and Arduino Uno module (left). Table 3 allows comparing these ﬁve embedded systems by providing a summary of the most signiﬁcant technical details. 5.1. Microcontroller/Microcomputer-Based Embedded Systems 18 of 28 Sensors 2018, 18, 460 Table 3. Comparison of different processing module components. Board CPU Speed Memory/Storage Power Comp. Ports A/D Size and Weight USB Ports/Wireless Interfaces Price Operating Systems Raspberry Pi Model A 700 MHz 256 MB/SD (4 GB) 0.8 W 0/30 6.5 × 5.5 cm/100 g 1 USB port/- $25 Raspbian Raspberry Pi 2 900 MHz 1 GB/SD (4 GB) 1.5 W 0/30 8.4 × 5.5 cm/136 g 4 USB ports/- $35 Raspbian Ubuntu Mate Windows 10 IoT Raspberry Pi 3 1.2 GHz 1 GB/SD (4 GB) 1.8 W 0/30 8.4 × 5.5 cm/136 g 4 USB ports/Wi-Fi and Bluetooth $40 Raspbian Ubuntu Mate Windows 10 IoT Beagle Bone 720 MHz 512 MB/SD (4 GB) 1.0 W 14/6 8.4 × 5.5 cm/100 g 1 USB port/- $75 BeagleBone Linux Intel Edison 500 MHz 1 GB/SD (4 GB) 0.6 W 14/6 5.9 × 2.8 cm/82 g -/Wi-Fi and Bluetooth $90 Yocto Project Pycom 32 MHz 1 kB/32 kB 0.2 W 14/6 6.8 × 5.4 cm/100 g Wi-Fi, Bluetooth and LoRa $45 Micropython Arduino Uno 32 MHz 1 kB/32 kB 0.2 W 14/6 6.8 × 5.4 cm/100 g -/- $25 Processing-based Arduino Nano 16 MHz 512 B/16 kB 0.2 W 14/6 4.2 × 1.8 cm/20 g -/- $10 Processing-based Table 3. Comparison of different processing module components. Sensors 2018, 18, 460 19 of 28 19 of 28 Operating Systems for IoT Microcomputers Operating Systems for IoT Microcomputers The choice of an adequate Operating System is a very important issue in the design of a sensor since it will shape the corresponding software architecture. Thus, we now proceed by analysing the different operating systems available for the microcomputers referred above. Raspbian is the most common operating system designed for Raspberry Pi. It is supported by all Raspberry Pi versions, and it has two versions: (i) a complete version with a graphical interface and many development tools which facilitate the development of solutions, but that consumes a lot of resources; and (ii) a lite version, without graphical interface, and with just a basic set of preinstalled software, allowing to add only those packages that are actually required. Since it is a Linux-based operating system, it supports several programming languages like C, C++, Java, Scratch, Python, or BASH. Ubuntu MATE is an option for Raspberry Pi microcomputers that is supported by model 2 and model 3. This operating system attempts to be simple from the end user perspective, integrating several entertainment applications, although it is also possible to add development tools to it. Angstrom is a modiﬁed Linux optimised for Beaglebone microcomputers. It has no graphical interface, and it supports several programming languages such as Python, C, Java, or BASH. In addition, it has its own programming language called BoneScript, which is based on the node.js language. g g Yocto Project is a complete embedded Linux development environment with tools and methods to facilitate the creation of embedded systems. It can be conﬁgured to run Arduino-based or Linux-based programs, thereby offering a great ﬂexibility. Micropython [79] is a Python 3.5 implementation optimised for running in microcontrollers. It allows interacting via a prompt, executing commands or running scripts in an autonomous way. It is entirely compatible with python, and it includes modules for accessing low-level hardware. Table 4. Comparison of different operating systems. Operating System Booting Time (s) Min. Operating Systems for IoT Microcomputers Memory (MB) Graphical Interface Type Programming Languages Raspbian 25 150 Yes Multi-thread C Java Python Scratch Raspbian Lite 15 50 No Multi-thread C Java Python Scratch Ubuntu MATE 30 200 Yes Multi-thread C Java Python Yocto Project 20 150 No Multi-thread C Java Python Arduino-based Angstrom Linux 15 100 No Multi-thread C Java Python Bonescript Windows 10 IoT 35 250 No Multi-thread Visual Studio C Micropython <1 <1 No Multi-thread Python - Processing-based <1 <1 No Single-thread Wiring - Windows 10 IoT Core is a Windows 10 based operating system oriented to Internet of Things projects using small devices. It is supported by Raspberry Pi 2 and 3, Arrow DragonBoard 410c, Table 4. Comparison of different operating systems. Table 4. Comparison of different operating systems. Windows 10 IoT Core is a Windows 10 based operating system oriented to Internet of Things projects using small devices. It is supported by Raspberry Pi 2 and 3, Arrow DragonBoard 410c, Windows 10 IoT Core is a Windows 10 based operating system oriented to Internet of Things projects using small devices. It is supported by Raspberry Pi 2 and 3, Arrow DragonBoard 410c, Windows 10 IoT Core is a Windows 10 based operating system oriented to Internet of Things projects using small devices. It is supported by Raspberry Pi 2 and 3, Arrow DragonBoard 410c, 20 of 28 Sensors 2018, 18, 460 and MinnowBoard MAX. Windows 10 IoT Core relies on the rich, extensible Universal Windows Platform (UWP) API for building solutions. It can be used together with the Visual Studio environment for programming. and MinnowBoard MAX. Windows 10 IoT Core relies on the rich, extensible Universal Windows Platform (UWP) API for building solutions. It can be used together with the Visual Studio environment for programming. Table 4 shows a brief comparison of the different Operating Systems currently available. 5.2. Air Pollution Sensors Nowadays, we can ﬁnd a wide variety of sensing technologies (see Figure 18) for gas detection (Metal oxide semiconductor, polymer, carbon nanotubes, moisture absorbing materials, Optics, Acoustics, etc.) as shown in [80]. Each technology has different properties, calibration processes and costs, among other characteristics, and so a comparison between these different technologies is required. Figure 18. Different types of air pollution sensors. Figure 18. Different types of air pollution sensors. To evaluate the different sensing technologies, we must mainly consider some characteristics, especially when focusing on the design of small mobile sensors: (i) sensibility, which refers to the range of values that the sensor can measure; (ii) selectivity, which is the capability of reacting only to the target gas; (iii) linearity, which is the rate of change with respect to gas variations; (iv) response time, which is the time required to start measuring correctly; (v) power consumption; and (vi) price. In the market, we can easily ﬁnd various gas sensors for air pollution monitoring, being the three following types of sensors the most common: electrochemical, semiconductor, and infrared. These have a wide range of prices and characteristics. Below we provide more details about each of these sensor types: Electrochemical gas sensors measure the concentration of some air pollutant by oxidizing or reducing its internal porous membrane, thereby producing current changes. Usually these sensors behave quite linearly, allowing to make accurate measurements. They typically operate at 5 V, having a power consumption of about 600 mW; most pollution sensors in this category cost between $100 and $400. Semiconductor gas sensors are the most common gas sensors because of their low cost and high sensitivity. They have an internal conductive material that increases their conductivity level in the presence of a speciﬁc air pollutant. These sensors are nonlinear and have a low selectivity, adding difﬁculty to the monitoring process. Usually they operate at 5 V, having a power consumption in the 500–900 mW range, and they cost between $10 and $35. Moisture absorbing material sensors are used for measuring temperature and humidity. Their dielectric constant varies according to the water content in the environment. They operate 21 of 28 Sensors 2018, 18, 460 at 5 V, having a power consumption of about 0.5 W. In addition, they are very cheap, with a price of about $5. Infrared sensors measure gas/element variations by detecting interferences in an infrared laser. 5.2. Air Pollution Sensors They are specially adequate for monitoring pollutants such as ﬁne particulate matter sized less than 10 micrometers (PM10), or ﬁne particulate matter sized less than 2.5 micrometers (PM2.5). They usually operate at 5 V, having a power consumption of about 1W, and their cost is about $40. Table 5 provides a brief summary of the most signiﬁcant aspects of these different types of sensors for comparison purposes. Table 5. Comparison of different sensing module components. Sensor Type Sensitivity Selectivity Linearity Response Time Power Comp. Size (cm) Price Electrochemical Medium Medium High Medium 0.6 W 2.0 × 4.0 $200 Semiconductor High Low Low Low 0.5 W 2.0 × 4.0 $10–$35 Moisture Absorbing High Medium High High 0.5 W 2.0 × 4.0 $5 Infrared High Low Medium High 1.0 W 15.0 × 10.0 $40 Table 5. Comparison of different sensing module components. 5.3. Communications Modules 5.3. Communications Modules Although the RFID [81] standard was developed for IoT solutions, there are currently several options available for providing communications between the sensors and the mobile terminal (usually a smartphone), as shown in Section 3. Figure 19 shows some communication modules examples. We now proceed to analyse the technical characteristics associated to the different available options. The main characteristics to consider are: (i) distance, that is, the wireless coverage range; (ii) communication type, which refers to the characteristics of the channel over which messages are transmitted—usually two types are considered, i.e. serial-based, when a communications channel is open to transmit a stream of data, or message-based, when the data are transmitted via a unique message; (iii) message size, which refers to the maximum size of the message when the communication is message-based; (iv) power consumption; and (v) price. Figure 19. Example of some communication modules. Figure 19. Example of some communication modules. Table 6 shows a brief summary of the most signiﬁcant aspects to consider in terms of communication modules. 22 of 28 Sensors 2018, 18, 460 Table 6. Comparison of different network module components. Module Communication Type Max. Message Size Data Rate Distance Power Comp. Price Wi-Fi Serial-based - +54 Mbps 100 m 0.5 W $10 NFC Message-based 32 kB 424 Kbps 0.04 m 0.1 W $35 ZigBee Message-based 128 bits 250 Kbps 100 m 0.1 W $40 SIGFOX Message-based 96 bits 140 msg/day +1 km 0.3 W $60 LoRa Message-based 0.1, 1 or 10% TimeOnAir 250–5400 bps 2–15 km 0.1 W $45 Bluetooth Serial-based - +2.1 Mbps 10 m 0.2 W $10 Bluetooth LE Message-based 160 bits 1 Mbps 10 m 0.05 W $15 We can observe that all options are relatively cheap, with prices in the range from $40–$50, but the more widely used in the market, i.e., Wi-Fi and both Bluetooth and Bluetooth Low Energy, are the cheapest ones, with prices in the range $10–$15. In terms of power consumption, the best option is Bluetooth Low Energy (0.05 W), although ZigBee, LoRA and NFC also exhibit low power consumption levels. In terms of bitrate, the best performing technology is Wi-Fi, being that typical wireless sensing technologies, such as LoRA, ZigBee, or SIGFOX, have a low bitrate. 5.3. Communications Modules Regarding the communications type, there are two options: (i) Wi-Fi and Bluetooth, which open a serial communications channel for transmitting a stream of bytes; and (ii) Bluetooth Low Energy, ZigBee, Sigfox, LoRA, and NFC, which transmit a message per iteration. 6. Mobile Air Pollution Sensor Design The design of a small and cheap mobile sensor is a basic requirement for air pollution monitoring. After analysing the main technical characteristics of hardware components available in the market, it quickly becomes evident that there are several options for creating a mobile air quality sensor for crowdsensing in the Internet of Things context. If focusing on the mobile sensor/smartphone wireless connection, the best option is using Bluetooth Low Energy since it is able to fulﬁl all requirements (low power consumption, low price, and small-sized modules) whereas all other options have different drawbacks. For instance, the ZigBee technology, despite being the most extended technology for wireless sensor networks due to its low power consumption, is not supported by current smartphones. The SIGFOX technology has very strict restrictions regarding the number of messages that it is possible to generate per time slot, thus having little applicability to our aims. The Wi-Fi technology, although being widely used and having a large coverage range, consumes more power and is typically used for Internet connectivity. Finally, the major problem of the NFC technology is its coverage range (only about 4 cm). In terms of sensor device prototyping, the best option for creating a small and cheap mobile sensor for air pollution monitoring is the Semiconductor gas option. Notice that it is cheaper than the electrochemical sensor, and even though the latter is more accurate, the error introduced can be mitigated by combining information from other nearby users. Concerning infrared sensors, they are only applicable to a speciﬁc type of pollutant (ﬁne particulate), while moisture absorbing sensors are mostly used to measure temperature and humidity. Next, we propose and evaluate some possible conﬁgurations for the different studied boards using a semiconductor gas sensor for pollution monitoring, and a Bluetooth Low Energy interface as the transmission technology. For all the proposed hardware conﬁgurations, it is possible to use an external USB charger as a power supply to offer more autonomy, while maintaining the support for mobility. y Figure 20 shows the four solutions we have developed to evaluate the different hardware options, and Table 7 shows a comparison between these solutions. 23 of 28 Sensors 2018, 18, 460 Figure 20. Proposed mobile sensing solutions for air quality monitoring. Figure 20. Proposed mobile sensing solutions for air quality monitoring. Table 7. Comparison of the four different solutions proposed. Table 7. Comparison of the four different solutions proposed. 6. Mobile Air Pollution Sensor Design Module Extras Network Power Comp. Weight Price Flexib. Develop. Comp. Power RPi 3 +converter – 2000 mW 200 g e90 ⋆⋆⋆⋆⋆⋆ ⋆⋆⋆ Beaglebone – +BLE usb 1500 mW 150 g e110 ⋆⋆⋆⋆⋆⋆ ⋆⋆⋆ Intel Edison +breakout +expansion – 1000 mW 100 g 200 g e130 ⋆⋆⋆⋆⋆⋆ ⋆⋆⋆ Arduino +circuit +BLE uart 600 mW 60 g e55 ⋆⋆⋆⋆⋆⋆ ⋆⋆⋆ The ﬁrst hardware option we propose is based on the Raspberry Pi platform (see Figure 20a). Since it has no analog ports, it has to be provided with an analog/digital converter. For this purpose, we propose using GrovePi [82], which is an extension board that allows connecting several analog/digital grove ports to a Raspberry Pi in an easy way. Furthermore, since the Raspberry Pi has several USB ports, we propose using a standard USB Bluetooth module for Raspberry Pi 2, or the built-in Bluetooth module for Raspberry Pi 3. With this solution, it becomes possible to run several programming languages, as it is possible to install a Linux or a Windows 10 IoT operating system, and there are a lot of development efforts around it. This conﬁguration has a power consumption of about 2000 mW, a total weight of 200 g, and it costs about $90. Overall, it is the most power-hungry solution among the four proposed, but it becomes the best option for quick prototyping due to its ﬂexibility and large community of developers. The second hardware approach relies on the BeagleBone board (see Figure 20b). It has several analog ports, allowing us to connect the sensor directly to this board without any intermediate device. In addition, since the BeagleBone has a USB port, it is possible to use a standard USB Bluetooth module. It runs a Linux-based operating system, allowing it to run different programming languages, but there are not too many developments or projects focusing on this solution. This conﬁguration has a power consumption of about 1500 mW, a total weight of 150 g, and it costs about $110, thus being one of the Sensors 2018, 18, 460 24 of 28 24 of 28 most expensive options. It is not a very useful prototyping platform since it has characteristics that are similar to the Raspberry Pi, but it has a smaller developer community and less support. most expensive options. 6. Mobile Air Pollution Sensor Design It is not a very useful prototyping platform since it has characteristics that are similar to the Raspberry Pi, but it has a smaller developer community and less support. The third hardware solution we propose is based on the Intel Edison platform (see Figure 20c). It has an embedded Bluetooth interface, but it does not provide an analog port to directly connect the sensor. For this purpose, it is possible to use: (i) the Arduino expansion board; or (ii) the Breakout board [83]. The ﬁrst sensor connection option is very simple, however the sensor becomes excessively large. Regarding the second option, the overall size remains small, but it is necessary to make an ad hoc circuit to connect the sensor. The Edison board supports a Linux-based operating system (Yocto) including the possibility to run Arduino-based scripts. For this last conﬁguration, the power consumption is of about 1000 mW, the total weight is 200 g, and it costs about $130, making it the most expensive option. It is useful mostly for end solutions due to its price. The last embedded solution we propose is based on the Arduino platform (see Figure 20d). Since it was designed for these types of solutions, it becomes easy to connect a sensor via the existing analog ports; nevertheless, USB ports are not available, and so a Bluetooth module must be connected via an UART port for both Arduino Uno and Arduino Nano boards. This solution only runs Arduino-based scripts, reducing the programming ﬂexibility, but we can ﬁnd a lot of developments using this platform. For the Arduino Uno solution, the power consumption is about 600 mW, the total weight is 150 g, and it costs about $55. This option is improved by the Arduino Nano solution, which has a power consumption of about 600 mW, a total weight of 60 g, and it costs about $50. The latter option is better than all others in terms of consumption, weight, and price, having as its only drawback the limited memory/CPU resources. It is also useful for restricted environments where the power consumption is very limited. 7. Conclusions and Future Work Smart Cities is a trending topic, and many research efforts are being made worldwide to progress towards that new paradigm which encompasses several areas including smart government, smart transport, smart environment, and smart grid. Since air pollution is considered to be one of the most signiﬁcant health risks worldwide, smart environment obviously becomes a very important area in the Smart City context. Although air monitoring stations have been sparsely deployed in most large cities for controlling pollution levels, these are not enough to provide a detailed view of the pollutants’ distribution in a city. In this sense, crowdsensing emerges as a good option to monitor the different pollutants by combining small mobile sensors and smartphones. Since current smartphones have much memory and high computing capabilities, mobile sensors can be kept minimal, focusing on data acquisition tasks alone. In this paper, we have analysed the most relevant Internet of Things architectures and protocols, along with the requirements of an embedded mobile sensor platform from a crowdsensing perspective, identifying the basic tasks the sensor must be able to perform. Besides, an analysis of the hardware architecture requirements has been done, and candidate off-the-shelf hardware components have been analysed. Finally, several complete hardware conﬁgurations meeting all the design requirements have been developed and compared in terms of power consumption, weight, and cost. Overall, we have found that the Arduino Nano platform, despite having very limited resources, is able to fulﬁl the established requirements, thus being the most recommendable alternative in terms of price, weight, and power consumption features. Regarding other hardware alternatives, microcomputers such as Raspberry Pi, BeagleBone, or Intel Edison are more powerful and ﬂexible by supporting a standard Operating System, thereby allowing to quickly deploy any application. We believe that the Raspberry Pi solution can be the best option for quick prototyping. For more professional solutions, requiring higher processing capacity, the Intel Edison becomes a better option, although imposing a higher overhead in terms of development time. Finally, Arduino becomes an option for very restricted environments. Sensors 2018, 18, 460 25 of 28 Overall, we ﬁnd that a hardware solution applicable to all IoT contexts, and meeting low size and low power requirements, along with adequate communication interfaces and battery capacity, is still missing, although in years to come many more products are expected. 7. Conclusions and Future Work The next steps in this research line are to test the adequacy of the proposed solutions not only in the mobile user context, but also in other contexts including coupling these sensors to vehicles such as cars, buses, bikes, or even ﬂying vehicles (multicopters). Acknowledgments: This work was partially supported by the Ministerio de Economía y Competitividad, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Proyectos I+D+I 2014, Spain, under Grant TEC2014-52690-R, the Generalitat Valenciana, Spain, the Secretaría Nacional de Educación Superior, Ciencia, Tecnología e Innovación del Ecuador (SENESCYT), and the Universidad de Cuenca. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. , ; , ; , pp g g , , 19. Cheng, Y.; Li, X.; Li, Z.; Jiang, S.; Li, Y.; Jia, J.; Jiang, X. AirCloud: A cloud-based air-quality monitoring system for everyone. In Proceedings of the 12th ACM Conference on Embedded Network Sensor Systems—SenSys ’14, Memphis, Tennessee, 3–6 November 2014; ACM: New York, NY, USA, 2014; pp. 251–265. References 1. Glasmeier, A.; Christopherson, S. Thinking about smart cities. Camb. J. Reg. Econ. Soc. 2015, 8, 3–12 2. Borgia, E. The Internet of Things vision: Key features, applications and open issues. Comput. Commun. 2014, 54, 1–31. 3. Gubbi, J.; Buyya, R.; Marusic, S.; Palaniswami, M. Internet of Things (IoT): A vision, architectural elements, and future directions. Future Gener. Comput. Syst. 2013, 29, 1645–1660. 4. Lee, E.A. Cyber physical systems: Design challenges. In Proceedings of the 2008 11th IEEE International Symposium on Object Oriented Real-Time Distributed Computing (ISORC), Orlando, FL, USA, 5–7 May 2008; pp. 363–369. 5. Wolf, W. Cyber-physical systems. Computer 2009, 42, 88–89. 5. Wolf, W. Cyber-physical systems. Computer 2009, 42, 88–89. 6 B h ti R Gill H C b h i l t I t C t l T h l 2011 12 161 166 6. Baheti, R.; Gill, H. Cyber-physical systems. Impact Control Technol. 2011, 12, 161–166. ti, R.; Gill, H. Cyber-physical systems. Impact Control 7. Rimal, B.P.; Choi, E.; Lumb, I. A Taxonomy and Survey of Cloud Computing Systems. In Proceedings of the Fifth International Joint Conference on INC, IMS and IDC, NCM ’09, Seoul, Korea, 25–27 August 2009; Volume 9, pp. 44–51. 8. Tsai, W.T.; Sun, X.; Balasooriya, J. Service-oriented cloud computing architecture. In Proceedings of the 2010 Seventh International Conference on Information Technology: New Generations (ITNG), Las Vegas, NV, USA, 12–14 April 2010; pp. 684–689. 9. Dinh, H.T.; Lee, C.; Niyato, D.; Wang, P. A survey of mobile cloud computing: Architecture, applications, and approaches. Wirel. Commun. Mob. Comput. 2013, 13, 1587–1611. 10. Ahmed, A.; Ahmed, E. A survey on mobile edge computing. In Proceedings of the 2016 10th International Conference on Intelligent Systems and Control (ISCO), Coimbatore, India, 7–8 January 2016; pp. 1–8. 11. Zhu, J.; Chan, D.S.; Prabhu, M.S.; Natarajan, P.; Hu, H.; Bonomi, F. Improving web sites performance using edge servers in fog computing architecture. In Proceedings of the 2013 IEEE 7th International Symposium on Service Oriented System Engineering (SOSE), Redwood City, CA, USA, 25–28 March 2013; pp. 320–323. 12 C li I K t l S S t Citi [G t dit ’ i t d ti ] IEEE I t t C t 2013 17 8 11 edge servers in fog computing architecture. In Proceedings of the 2013 IEEE 7th International Symposium on Service Oriented System Engineering (SOSE), Redwood City, CA, USA, 25–28 March 2013; pp. 320–323. 12. References Celino, I.; Kotoulas, S. Smart Cities [Guest editors’ introduction]. IEEE Internet Comput. 2013, 17, 8–11. 12. Celino, I.; Kotoulas, S. Smart Cities [Guest editors’ introduction]. IEEE Internet Comput. 2013, 17, 8–11. 13. Zanella, A.; Bui, N.; Castellani, A.; Vangelista, L.; Zorzi, M. Internet of things for smart cities. IEEE Internet Things J. 2014, 1, 22–32. 4. United States Environmental Protection Agency.Air Quality Index. Available online: http://cfpub.epa.g airnow/index.cfm?action=aqibasics.aqi (accessed on 29 November 2015). 5. Fishman, J.; Ramanathan, V.; Crutzen, P.; Liu, S. Tropospheric ozone and climate. Nature 1979, 282, 818– 15. Fishman, J.; Ramanathan, V.; Crutzen, P.; Liu, S. Tropospheric ozone and climate. Nature 1979, 282, 818–820. 16. Guo, B.; Wang, Z.; Yu, Z.; Wang, Y.; Yen, N.Y.; Huang, R.; Zhou, X. Mobile crowd sensing and computing: The review of an emerging human-powered sensing paradigm. ACM Comput. Surv. (CSUR) 2015, 48, 7. 17. Ganti, R.K.; Ye, F.; Lei, H. Mobile crowdsensing: Current state and future challenges. IEEE Commun. Mag. 2011, 49, 32–39. 16. Guo, B.; Wang, Z.; Yu, Z.; Wang, Y.; Yen, N.Y.; Huang, R.; Zhou, X. Mobile crowd sensing and computing: The review of an emerging human-powered sensing paradigm. ACM Comput. Surv. (CSUR) 2015, 48, 7. 17. Ganti, R.K.; Ye, F.; Lei, H. Mobile crowdsensing: Current state and future challenges. IEEE Commun. Mag. 2011, 49, 32–39. , , Ma, H.; Zhao, D.; Yuan, P. Opportunities in mobile crowd sensing. IEEE Commun. Mag. 2014, 52, 29–35. 19. Cheng, Y.; Li, X.; Li, Z.; Jiang, S.; Li, Y.; Jia, J.; Jiang, X. AirCloud: A cloud-based air-quality monitoring system for everyone. In Proceedings of the 12th ACM Conference on Embedded Network Sensor Systems—SenSys ’14, Memphis, Tennessee, 3–6 November 2014; ACM: New York, NY, USA, 2014; pp. 251–265. Sensors 2018, 18, 460 26 of 28 26 of 28 20. Zheng, Y.; Liu, F.; Hsieh, H.P. U-Air: When urban air quality inference meets big data. In Proceedings of the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining—KDD ’13, Chicago, IL, USA, 11–14 August 2013; pp. 1436–1444. 21. Hasenfratz, D.; Saukh, O.; Sturzenegger, S.; Thiele, L. Participatory air pollution monitoring using smartphones. In Proceedings of the 2nd International Workshop on Mobile Sensing, Beijing, China, 16–20 April 2012; pp. 1–5. 22. Brkovi´c, M.; Sretovi´c, V. Urban Sensing–Smart Solutions for Monitoring Environmental Quality: Case Studies from Serbia. References In Proceedings of the 48th ISOCARP—International Society of City and Regional Planners World Congress: Fast Forward: Planning in a (Hyper) Dynamic Urban Context, Perm, Russia, 10–13 September 2012. p 23. Libelium. Available online: http://www.libelium.com/ (accessed on 27 November 2017). 24. Devarakonda, S.; Sevusu, P.; Liu, H.; Liu, R.; Iftode, L.; Nath, B. Real-time air quality monitoring through mobile sensing in metropolitan areas. In Proceedings of the 2nd ACM SIGKDD International Workshop on Urban Computing, Chicago, IL, USA, 11–14 August 2013; p. 15. 25. Mead, M.; Popoola, O.; Stewart, G.; Landshoff, P.; Calleja, M.; Hayes, M.; Baldovi, J.; McLeod, M.; Hodgson, T.; Dicks, J.; et al. The use of electrochemical sensors for monitoring urban air quality in low-cost, high-density networks. Atmos. Environ. 2013, 70, 186–203. 26. Manna, S.; Bhunia, S.S.; Mukherjee, N. Vehicular pollution monitoring using IoT. In Proceedings of the Recent Advances and Innovations in Engineering (ICRAIE), Jaipur, India, 9–11 May 2014; pp. 1–5. 27. Hjiri, M.; El Mir, L.; Leonardi, S.; Pistone, A.; Mavilia, L.; Neri, G. Al-doped ZnO for highly sensitive CO gas sensors. Sens. Actuators B Chem. 2014, 196, 413–420. 28. Borini, S.; White, R.; Wei, D.; Astley, M.; Haque, S.; Spigone, E.; Harris, N.; Kivioja, J.; Ryhanen, T. Ultrafast graphene oxide humidity sensors. ACS Nano 2013, 7, 11166–11173. 29. Chen, X.; Wong, C.K.; Yuan, C.A.; Zhang, G. Nanowire-based gas sensors. Sens. Actuators B Chem. 2013, 177, 178–195. 30. Zaatar, Y.; Zaouk, D.; Bechara, J.; Khoury, A.; Llinaress, C.; Charles, J.P. Fabrication and characterization of an evanescent wave ﬁber optic sensor for air pollution control. Mater. Sci. Eng. B 2000, 74, 296–298. 31. Khan, W.Z.; Xiang, Y.; Aalsalem, M.Y.; Arshad, Q. Mobile phone sensing systems: A survey. IEEE Commun. Surv. Tutor. 2013, 15, 402–427. 30. Zaatar, Y.; Zaouk, D.; Bechara, J.; Khoury, A.; Llinaress, C.; Charles, J.P. Fabrication and characterization of an evanescent wave ﬁber optic sensor for air pollution control. Mater. Sci. Eng. B 2000, 74, 296–298. 31 Kh W Z Xi Y A l l M Y A h d Q M bil h i t A IEEE C 31. Khan, W.Z.; Xiang, Y.; Aalsalem, M.Y.; Arshad, Q. Mobile phone sensing systems: A survey. IEEE C Surv. Tutor. 2013, 15, 402–427. 32. Ricquebourg, V.; Menga, D.; Durand, D.; Marhic, B.; Delahoche, L.; Loge, C. The smart home concept: Our immediate future. References Constrained RESTful Environments (CoRE) Link Format. Available online: https://tools.ietf.org/ html/rfc6690.html (accessed on 26 November 2017). 51. Locke, D. MQTT V3. 1 Protocol Speciﬁcation; Technical Report; International Business Machines Corporation (IBM) and Eurotech: Udine, Italy, 2010; p. 42. 52. Hunkeler, U.; Truong, H.L.; Stanford-Clark, A. MQTT-S—A publish/subscribe protocol for Wireless Sensor Networks. In Proceedings of the 3rd International conference on Communication Systems Software and Middleware and Workshops, COMSWARE 2008, Bangalore, India, 6–10 January 2008; pp. 791–798. 53. Shelby, Z.; Hartke, K.; Bormann, C. The Constrained Application Protocol (CoAP). Available online: https://tools.ietf.org/html/rfc7252 (accessed on 27 November 2017). 54. Saint-Andre, P. Extensible Messaging and Presence Protocol (XMPP): Core. Available online: https: //tools.ietf.org/html/rfc6120.html (accessed on 27 November 2017). 55. Waher, P. XEP-0323: Internet of Things—Sensor Data. Available online: https://xmpp.org/extensions/xep- 0323.pdf (accessed on November 2017). 56. Dawson-Haggerty, S.; Jiang, X.; Tolle, G.; Ortiz, J.; Culler, D. sMAP: A simple measurement and actuation proﬁle for physical information. In Proceedings of the 8th ACM Conference on Embedded Networked Sensor Systems, Zürich, Switzerland, 3–5 November 2010; pp. 197–210. 7. Bray, T.; Paoli, J.; Sperberg-McQueen, C.M.; Maler, E.; Yergeau, F. Extensible markup language (XM World Wide Web J. 1997, 2, 27–66. 8. Bray, T. The JavaScript Object Notation (JSON) Fata Interchange Format. Available online: https://build tools.ietf.org/html/rfc8259 (accessed on 27 November 2017). 59. Schneider, J.; Kamiya, T.; Peintner, D.; Kyusakov, R. Efﬁcient XML Interchange (EXI) Format 1.0. W3C Proposed Recommendation. Available online: https://www.w3.org/TR/exi/ (accessed on 27 November 2017). 60. Furuhashi, S. MessagePack: It’s Like JSON. But Fast and Small. Available online: http://msgpack.org (accessed on 26 November 2014). 61. Botts, M.; Robin, A.; Greenwood, J.; Wesloh, D. OGC SensorML: Model and XML Encoding Standard; OGC Geospatial Consortium: Wayland, MA, USA, 2014. 62. Compton, M.; Neuhaus, H.; Taylor, K.; Parashar, A. Semantic Sensor Network Ontology. In Web Semantics: Science, Services and Agents on the World Wide Web; CSIRO: Victoria, Australia, 2012; pp. 25–32. 3. Kajimoto, K.; Matsukura, R.; Hund, J.; Kovatsch, M.; Nimura, K. Web of Things (WoT) Architecture (Unofﬁ Draft); WoT W3C Interest Group: Cambridge, MA, USA, 2016. 4. Yick, J.; Mukherjee, B.; Ghosal, D. Wireless sensor network survey. Comput. Netw. 2008, 52, 2292–2330. 65. Akyildiz, I.F.; Su, W.; Sankarasubramaniam, Y.; Cayirci, E. Wireless sensor networks: A survey. Comput. Netw. 2002, 38, 393–422. 66. Munir, S.A.; Ren, B.; Jiao, W.; Wang, B.; Xie, D.; Ma, J. Mobile wireless sensor network: Architecture and enabling technologies for ubiquitous computing. References In Proceedings of the 2006 1st IEEE International Conference on E-Learning in Industrial Electronics, Hammamet, Tunisia, 18–20 December 2006; pp. 23–28. 33. Lenior, D.; Janssen, W.; Neerincx, M.; Schreibers, K. Human-factors engineering for smart transport: Decision support for car drivers and train trafﬁc controllers. Appl. Ergon. 2006, 37, 479–490. 34. Demirkan, H. A smart healthcare systems framework. IT Prof. 2013, 15, 38–45. 35. Gao, W.; Tian, Y.; Huang, T.; Ma, S.; Zhang, X. The IEEE 1857 standard: Empowering smart video surveillance systems. IEEE Intell. Syst. 2014, 29, 30–39. 36. Hewitt, C. Security without IoT Mandatory Backdoors. Available online: https://ssrn.com/abstract=2795682 (accessed on 27 November 2017). 37. McFarlane, D.; Giannikas, V.; Wong, A.C.; Harrison, M. Product intelligence in industrial control: Theory and practice. Annu. Rev. Control 2013, 37, 69–88. p 8. Atzori, L.; Iera, A.; Morabito, G. The internet of things: A survey. Comput. Netw. 2010, 54, 2787–2805. 9. Ejaz, W.; Anpalagan, A.; Imran, M.A.; Jo, M.; Naeem, M.; Qaisar, S.B.; Wang, W. Internet of Things (IoT 5G wireless communications. IEEE Access 2016, 4, 10310–10314. 40. Gupta, A.; Jha, R.K. A survey of 5G network: Architecture and emerging technologies. IEEE Access 2015, 3, 1206–1232. 41. Baronti, P.; Pillai, P.; Chook, V.W.; Chessa, S.; Gotta, A.; Hu, Y.F. Wireless sensor networks: A survey on the state of the art and the 802.15. 4 and ZigBee standards. Comput. Commun. 2007, 30, 1655–1695. 42. Asherson, S.; Kritzinger, P.; Pileggi, P. Wireless Standards and Mesh Networks; Computer Science Department, University of Cape Town: Cape Town, South Africa, 2007. 43. Sornin, N.; Luis, M. LoRa Speciﬁcation. Available online: https://www.rs-online.com/designspark/rel- assets/ds-assets/uploads/knowledge-items/application-notes-for-the-internet-of-things/LoRaWAN% 20Speciﬁcation%201R0.pdf (accessed on 26 November 2017). 27 of 28 Sensors 2018, 18, 460 44. Sigfox. Available online: http://www.sigfox.com (accessed on 25 November 2017). 45. Haselsteiner, E.; Breitfuß, K. Security in near ﬁeld communication (NFC). In Proceedings of the Workshop on RFID Security, Graz, Austria, 1 August 2006; pp. 12–14. 6. Bluetooth Core Speciﬁcation Version 4.0. Available online: https://www.bluetooth.org/docman/handl downloaddoc.ashx?doc_id=229737 (accessed on 26 November 2017). 47. Gomez, C.; Oller, J.; Paradells, J. Overview and evaluation of bluetooth low energy: An emerging low-power wireless technology. Sensors 2012, 12, 11734–11753. 48. Mulligan, G. The 6LoWPAN architecture. In Proceedings of the 4th workshop on Embedded Networked Sensors, Cork, Ireland, 25–26 June 2007; pp. 78–82. 49. Fielding, R.T. REST: Architectural Styles and the Design of Network-Based Software Architectures. Ph.D. Thesis, University of California, Oakland, CA, USA, 2000. 50. Shelby, Z. References In Proceedings of the 21st International Conference on Advanced Information Networking and Applications Workshops, AINAW’07, Niagara Falls, ON, Canada, 21–23 May 2007; Volume 2, pp. 113–120. 67. Tilak, S.; Abu-Ghazaleh, N.B.; Heinzelman, W. A taxonomy of wireless micro-sensor network models. ACM SIGMOBILE Mob. Comput. Commun. Rev. 2002, 6, 28–36. 68. Alvear, O.; Zamora, W.; Calafate, C.; Cano, J.C.; Manzoni, P. An Architecture Offering Mobile Pollution Sensing with High Spatial Resolution. J. Sens. 2016, 2016, 1458147. Sensors 2018, 18, 460 28 of 28 28 of 28 69. Tsujita, W.; Yoshino, A.; Ishida, H.; Moriizumi, T. Gas sensor network for air-pollution monitoring. Sens. Actuators B Chem. 2005, 110, 304–311. 70. Kamionka, M.; Breuil, P.; Pijolat, C. Calibration of a multivariate gas sensing device for atmospheric pollution measurement. Sens. Actuators B Chem. 2006, 118, 323–327. 71. Alvear, Ó.; Calafate, C.T.; Cano, J.C.; Manzoni, P. Calibrating low-end sensors for ozone monitoring. In International Internet of Things Summit; Springer: Cham, Switzerland, 2015; pp. 251–256. 72. Wong, D.W.; Yuan, L.; Perlin, S.A. Comparison of spatial interpolation methods for the estimation of air quality data. J. Expo. Sci. Environ. Epidemiol. 2004, 14, 404–415. 73. Lane, N.D.; Miluzzo, E.; Lu, H.; Peebles, D.; Choudhury, T.; Campbell, A.T. A survey of mobile phone sensing. IEEE Commun. Mag. 2010, 48, 140–150. g g 74. Seeedstudio. Available online: http://www.seeedstudio.com/ (accessed on 26 November 2017). Seeedstudio. Available online: http://www.seeedst 75. Adafruit. Available online: https://www.adafruit.com/ (accessed on 26 November 2017). 76. TSN Systems . Available online: http://www.tst-sistemas.es/en/ (accessed on 26 November 2017 77. Pycom. Available online: https://pycom.io/ (accessed on 26 November 2017). 78. Ben Fry, C.R. Processing. Available online: https://www.processing.org/ (accessed on 26 N 79. Micropython. Available online: https://micropython.org/ (accessed on 26 November 2017). 79. Micropython. Available online: https://micropython.org/ (accessed on 26 November 2017). 80 Liu X ; Cheng S ; Liu H ; Hu S ; Zhang D ; Ning H A survey on gas sensing technology Sensors 2012 0. Liu, X.; Cheng, S.; Liu, H.; Hu, S.; Zhang, D.; Ning, H. A survey on gas sensing technology. Sensors 2 12, 9635–9665. 81. Landt, J. The history of RFID. IEEE Potentials 2005, 24, 8–11. 82. Seeed-Studio. Grovepi Extension Board. Available online: http://www.seeedstudio.com/depot/GrovePi- Starter-Kit-for-Raspberry-Pi-p-2240.html (accessed on 26 November 2017). 83. Intel-Corp. Intel Edison Compute Module. Available online: http://www.intel.eu/content/www/eu/en/ do-it-yourself/edison.html (accessed on 26 November 2017). c⃝2018 by the authors. Licensee MDPI, Basel, Switzerland. References This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2942116119	https://europepmc.org/articles/pmc6497753?pdf=render	English	null	Myeloid-Derived Suppressor Cells in Lung Transplantation	Frontiers in immunology	2,019	cc-by	4,970	Myeloid-Derived Suppressor Cells in Lung Transplantation Tobias Heigl 1†, Anurag Singh 2†, Berta Saez-Gimenez 1, Janne Kaes 1, Anke Van Herck 1, Annelore Sacreas 1, Hanne Beeckmans 1, Arno Vanstapel 1, Stijn E. Verleden 1, Dirk E. Van Raemdonck 1, Geert Verleden 1, Bart M. Vanaudenaerde 1, Dominik Hartl 2 and Robin Vos 1 1 Lung Transplant Unit, Lab of Respiratory Diseases, Department of Chronic Diseases, Metabolism and Ageing (CHROMETA), KU Leuven, Leuven, Belgium, 2 Universitätsklinik für Kinder-und Jugendmedizin, Tübingen, Germany Myeloid-derived suppressor cells (MDSC) are a heterogeneous group of immune cells from the myeloid lineage. MDSCs expand in pathological situations, such as chronic infection, cancer, autoimmunity, and allograft rejection. As chronic lung allograft dysfunction (CLAD) limits long-term survival after lung transplantation (LTx), MDSCs may play a role in its pathophysiology. We assessed phenotype and frequency of MDSCs in peripheral blood from lung transplant recipients and its relationship to post-transplant complications and immunosuppression. Granulocytic (G)-MDSC were identiﬁed and quantiﬁed by ﬂow cytometry of blood from 4 control subjects and 20 lung transplant patients (stable n = 6, infection n = 5; CLAD n = 9). G-MDSC functionality was assessed in vitro by their capability to block CD4 and CD8 T cell proliferation. More G-MDSC could be assessed using EDTA tubes compared to heparin tubes (p = 0.004). G-MDSC were increased in stable lung transplant recipients vs. non-transplant controls (52.1% vs. 9.4%; p = 0.0095). The infection or CLAD groups had lower G-MDSCs vs. stable recipients (28.2%p = 0.041 and 33.0%; p = 0.088, respectively), but were not different among CLAD phenotypes. G-MDSC tended to correlate with cyclosporine A and tacrolimus levels (r² = 0.18; r² = 0.17). CD4 and CD8 cells proliferation decreased by 50 and 80% if co-cultured with MDSCs (1:6 and 1:2 MDSC:T-cell ratio, respectively). In conclusion, circulating MDSCs are measurable, functional and have a G-MDSC phenotype in lung transplant patients. Their frequency is increased in stable patients, decreased during post-transplant complications, and related to level of immunosuppression. This study may pave the way for further investigations of MDSC in the context of lung transplantation. Citation: Heigl T, Singh A, Saez-Gimenez B, Kaes J, Van Herck A, Sacreas A, Beeckmans H, Vanstapel A, Verleden SE, Van Raemdonck DE, Verleden G, Vanaudenaerde BM, Hartl D and Vos R (2019) Myeloid-Derived Suppressor Cells in Lung Transplantation. Front. Immunol. 10:900. doi: 10.3389/ﬁmmu.2019.00900 Heigl T, Singh A, Saez-Gimenez B, Kaes J, Van Herck A, Sacreas A, Beeckmans H, Vanstapel A, Verleden SE, Van Raemdonck DE, Verleden G, Vanaudenaerde BM, Hartl D and Vos R (2019) Keywords: myeloid-derived suppressor cells, blood, lung transplantation, allograft, chronic rejection, immunosuppression, infection, phenotypes Specialty section: Specialty section: This article was submitted to Alloimmunity and Transplantation, a section of the journal Frontiers in Immunology Received: 21 December 2018 Accepted: 08 April 2019 Published: 26 April 2019 ORIGINAL RESEARCH published: 26 April 2019 doi: 10.3389/ﬁmmu.2019.00900 ORIGINAL RESEARCH Edited by: Olivera J. Finn, University of Pittsburgh, United States Reviewed by: Régis Josien, University of Nantes, France Stanislaw Stepkowski, University of Toledo, United States Correspondence: Robin Vos robin.vos@uzleuven.be Anurag Singh anurag.singh@med.uni-tuebingen.de Edited by: Olivera J. Finn, University of Pittsburgh, United States Reviewed by: Régis Josien, University of Nantes, France Stanislaw Stepkowski, University of Toledo, United States Correspondence: Robin Vos robin.vos@uzleuven.be Anurag Singh anurag.singh@med.uni-tuebingen.de *Correspondence: Robin Vos robin.vos@uzleuven.be Anurag Singh anurag.singh@med.uni-tuebingen.de †These authors have contributed equally to this work †These authors have contributed equally to this work †These authors have contributed equally to this work INTRODUCTION From a transplant immunological point of view, graft acceptance is the fundamental element in allograft survival. Graft acceptance is realized by blocking the immune system with immunosuppression preventing host immune cells to recognized and attack the “non-self” donor (lung) tissue. Immune regulatory cells are thought to play a major role in the balance between graft acceptance and chronic rejection. Most attention has gone to natural and inducible FoxP3 positive From a transplant immunological point of view, graft acceptance is the fundamental element in allograft survival. Graft acceptance is realized by blocking the immune system with immunosuppression preventing host immune cells to recognized and attack the “non-self” donor (lung) tissue. Immune regulatory cells are thought to play a major role in the balance between graft acceptance and chronic rejection. Most attention has gone to natural and inducible FoxP3 positive April 2019 \| Volume 10 \| Article 900 Frontiers in Immunology \| www.frontiersin.org Myeloid-Derived Suppressor Cells in Lung Transplantation Heigl et al. regulatory T cells (Treg) (1). Immune regulation and graft acceptance, however, encompasses many more cells including regulatory B cells, regulatory dendritic cells and innate regulatory cells like the myeloid-derived suppressor cells (MDSCs), which were introduced 10 years ago by Gabrilovich et al., MDSCs were initially described as a heterogeneous group of immune cells from the myeloid lineage with a potent immune-regulatory activity (2). In the last few years, more insights into the nature and biological role of MDSCs have been reported and consequently MDSCs have emerged as a universal regulator of immune function in many pathologic conditions. MDSCs are known to expand in pathological situations such as chronic infection, cancer, transplant rejection and autoimmunity (3–5). Within the MDSC population, two main subgroups of cells were identiﬁed: granulocytic MDSCs (G-MDSCs) also nominated as polymorphomononuclear (PMN-MDSCs) and monocytic (M)- MDSCs. G-MDSCs are phenotypically and morphologically similar to neutrophils, whereas M-MDSCs resemble monocytes (6). Looking at the functionality of both M- and G-MDCS, the suppressive activity has been mainly attributed to arginine 1 (ARG1) and nitric oxide (NO) for M-MDSC and upregulation of reactive oxygen species (ROS) for G-MDSC (7, 8). Upregulation of ARG1, NO, and ROS are key mechanism to suppress T cell proliferation (9) and the production of IFNγ (10). Another hallmark is the upregulation of the transcription factor signal transducer and activator of transcription 3 (STAT3). INTRODUCTION STAT3, which functions as a signaling hub, integrating the diﬀerent cues of the immunologic micro-environment (11, 12) regulates the expansion of MDSCs by stimulating myelopoiesis and inhibiting myeloid-cell diﬀerentiation. Further, it promotes MDSC survival by inducing the expression of cyclin D1, B-cell lymphoma XL (BCL-XL) and MYC (4). Within transplantation, MDSCs are involved in maintaining allogeneic acceptance in bone marrow, kidney and liver transplantation (13–16). Moreover, it has also been shown that commonly used immunosuppressive drugs can aﬀect MDSC diﬀerentiation and functionality (17, 18). Our goal was to characterize phenotype (M-MDSC or G-MDSC) and frequency of MDSCs in lung transplant recipients. And consequently, to assess if MDSCs can serve as a potential new research target in the ﬁeld of lung transplantation since chronic lung allograft dysfunction (CLAD), considered to be driven by an overactive T cell response, FIGURE 1 \| Gating Strategy to determine MDSC phenotype. (A) The low-density fraction of PBMC was stained with speciﬁc markers to differentiate between G-MDSCs (CD66b/CD33) and M-MDSCs (HLA-DR/CD14). (B) Different coatings of blood tubes (EDTA vs. Heparin) affect the MDSC cell numbers. (C) Exemplary FACS plots of the healthy controls and different LTx patient groups. FIGURE 1 \| Gating Strategy to determine MDSC phenotype. (A) The low-density fraction of PBMC was stained with speciﬁc markers to differentiate between G-MDSCs (CD66b/CD33) and M-MDSCs (HLA-DR/CD14). (B) Different coatings of blood tubes (EDTA vs. Heparin) affect the MDSC cell numbers. (C) Exemplary FACS plots of the healthy controls and different LTx patient groups. April 2019 \| Volume 10 \| Article 900 2 2 Frontiers in Immunology \| www.frontiersin.org Myeloid-Derived Suppressor Cells in Lung Transplantation Heigl et al. remains the most important factor limiting long-term survival after transplantation. for CLAD and its phenotypes. Lung transplant recipients were selected according to their clinical status upon recruitment: 6 were considered stable, 5 recipients had an acute infection (2 CMV; 1 Pseudomonas aeruginosa; 1 Inﬂuenza + E. coli; 1 Inﬂuenza + Aspergillus fumigatus) and 9 were aﬀected by diﬀerent phenotypes of CLAD (5 BOS and 4 RAS cases). Blood of 15 individuals was used to compare Heparin vs. EDTA coated blood tubes (2 control, 3 Infection, 5 Stable, and 5 CLAD). The clinical status was assessed by an expert clinician (RV) according to current guidelines (19, 20). Patient Characteristics This study included 20 lung transplant recipients and 4 healthy controls recruited at the University Hospitals Leuven (Belgium). All lung transplant recipients gave informed consent at time of listing for transplantation and routine blood sampling was approved by the University hospital (S51577). Relevant patient information retrieved from the clinical database included age, gender, type of transplantation, underlying disease, allograft ischemic time during transplantation, immunosuppressive dose, and trough levels, time post-transplant of blood sampling, time of death, infection information, and diagnostic criteria RESULTS Clinical characteristics of study participants are included in Table 1. G-MDSC were present in the low-density fraction of PBMCs, based on physical (FSC/SSC) and ﬂow cytometric characteristics (CD33+CD66b+ cells) (Figure 1A). M-MDSC, on the other hand, were not observed in the low-density fraction of PBMCs, based on physical (FSC/SSC) and cell surface marker characteristics (CD14−HLA-DR−) (Figure 1A). T-Cell Suppression Assays pp y The MDSC functional assay assessed T-cell suppression (both CD4 and CD8) by isolated MDSC (Figure 2) (23). MDSCs were isolated from blood of 2 lung transplant recipients, 1 stable and 1 with CLAD (BOS), using anti-CD66b and anti- FITC magnetic microbeads with the autoMACS R⃝Pro Separator (Miltenyi Biotec) according to manufacturer’s instructions. CD4+ and CD8+ T cells were isolated using CD4 and CD8 antibody (BD Pharmingen) combined with anti-FITC magnetic microbeads and autoMACS R⃝Pro Separator (Miltenyi Biotec). Isolated CD4 or CD8 cells were labeled with CFSE dissolved in RPMI-1640, supplemented with 10% heat-inactivated human serum, 2 mM glutamine, 100 IU/ml penicillin, and 100 mg/ml streptomycin and 60,000 cells were plated per well in a 96-well microtiter plate. Cells were further stimulated with 100 U/ml IL- 2 (R&D Systems) and 1 µg/ml OKT3 (Janssen Cilag). Diﬀerent numbers of G-MDSCs were added to obtain an MDSC:T-cell ratio 1:6 and 1:2 and incubated for 3 days in a humidiﬁed chamber at 37◦C and 5% CO2. After incubation, cells were harvested and CFSE-ﬂuorescence intensity analyzed by ﬂow cytometry to determine T-cell proliferation. Proliferation was calculated as the ratio of the divided cells (P1 to P5) over all cells (P0 to P5) with control T cells as reference value. MDSC Characterization All Stable Infection CLAD p n = 20 n = 6 n = 5 n = 9 Age, median (IQR) 55 (32–60) 51.5 (20–60) 58 (36–62) 54 (39–57) 0.65 Gender: Male, n (%) 9 (42.9) 2 (33.3) 2 (40.0) 4 (44.4) 0.17 Diagnosis, n (%) COPD 10 (50.0) 3 (50.0) 3 (60.0) 4 (44.4) 0.86 ILD 2 (10.0) 0 (0.0) 0 (0.0) 2 (22.2) 0.47 CF 5 (25.0) 2 (33.3) 1 (20.0) 2 (22.2) 1.00 Other 3 (15.0) 1 (16.7) 1 (20.0) 1 (11.1) 1.00 Immunosupressive treatment, n (%) CsA+AZA+P 2 (10.0) 1 (16.7) 0 (0.0) 1 (11.1) 1.00 CsA+MMF+P 2 (10.0) 0 (0.0) 2 (40.0) 0(0.0) 0.05 CsA+P 1 (5.0) 0 (0.0) 1 (20.0) 0 (0.0) 0.25 FK+AZA+P 5 (25.0) 1 (16.7) 1 (20.0) 3 (33.3) 0.82 FK+MMF+P 5 (25.0) 3 (50.0) 1 (20.0) 1 (11.1) 0.35 FK+P 4 (20.0) 1 (16.7) 0 (0.0) 3 (33.3) 0.41 FK 1 (5.0) 0 (0.0) 0 (0.0) 1 (11.1) 1.00 Type of LTx SSLT 18 (90.0) 6 (100.0) 5 (100.0) 7 (77.8) 0.48 SLT 2 (10.0) 0 (0.0) 0 (0.0) 2 (22.2) Survival post LTx (years), median (IQR) 7.0 (4.1–9.7) 5.6 (3.8–8.7) 4.3 (3.0–7.4) 7.5 (5.9–11.9) 0.18 Sampling time post LTx (months), median (IQR) 3.9 (0.9–6.6) 1.9 (0.7–5.1) 0.9 (0.5–4.7) 6.6 (4.8–9.3) 0.02 TABLE 1 \| Characteristics of lung transplant patients. characterized as previously described (21, 22). In brief, peripheral blood mononuclear cells (PBMCs) were isolated from whole blood by Ficoll density gradient centrifugation (Lymphocyte Separation Medium; Biochrom), washed with RPMI-1640 and cell viability was conﬁrmed by trypan blue staining. The isolated PBMC, containing only low density granulocytes, were stained with speciﬁc antibodies for G-MDSC (CD66b-FITC, CD33-PE) and M-MDSC (CD14-FITC and HLADR-PerCP) (Miltenyi Biotec) and quantiﬁed by ﬂow cytometry using a FACSCalibur (BD). G-MDSCs were phenotypically characterized as low- density fraction granulocytes CD33+CD66b+ cells (Figure 1). The percentage of G-MDSC was determined as ratio of CD33+CD66b+ cells (P2 in Figure 1) over total PBMCs containing the low density granulocyte fraction (P1 in Figure 1). Calculations were performed with BD CellQuest Pro analysis software and FlowJo V7. MDSC Characterization Peripheral blood was collected using EDTA and Heparin-coated tubes and samples were shipped to the Universitätsklinik für Kinder-und Jugendmedizin, Tübingen (Germany) at room temperature and analyzed within 24 h. MDSCs were FIGURE 2 \| G-MDSC percentages measured in blood of lung transplant recipients and healthy controls. (A) the effect of LiHe vs. EDTA tubes on G-MDSC percentages; (B) G-MDSC in healthy controls and lung transplant recipients who were stable, had an infection or were diagnosed with CLAD; (C) CLAD sub-phenotypes in BOS and RAS. (D,E) G-MDSC correlated with CNI level of the patients. FIGURE 2 \| G-MDSC percentages measured in blood of lung transplant recipients and healthy controls. (A) the effect of LiHe vs. EDTA tubes on G-MDSC percentages; (B) G-MDSC in healthy controls and lung transplant recipients who were stable, had an infection or were diagnosed with CLAD; (C) CLAD sub-phenotypes in BOS and RAS. (D,E) G-MDSC correlated with CNI level of the patients. April 2019 \| Volume 10 \| Article 900 Frontiers in Immunology \| www.frontiersin.org 3 Myeloid-Derived Suppressor Cells in Lung Transplantation Heigl et al. TABLE 1 \| Characteristics of lung transplant patients. Statistical Analysis Qualitative variables are expressed as absolute numbers and percentages. Normally distributed quantitative variables are expressed as mean and standard deviation; non-normally distributed variables are expressed as median and interquartile range (25–75 percentile). Demographic and clinical variables of patients were compared using the chi-square test for qualitative variables or Fisher’s exact test when one of the expected eﬀects was <5. Normally distributed quantitative variables were compared using one-way ANOVA test; non-normally distributed quantitative variables were compared using the Kruskal-Wallis test. One-way ANOVA Test was used to compare MDSCs counts between groups. Linear regression was used for investigating the interaction of MDSC% and immunosuppressive trough levels. Data were analyzed using Graph Pad prism 7.0 software (San Diego, CA, USA). Percentages of G-MDSC were increased when using EDTA tubes compared to using LiHe tubes (mean: 33.38% [range: 18.32–50.36] vs. 6.24% [4.02–20.53], p = 0.004) (Figures 1B, 2A). EDTA and LiHe tubes were equally (statistically not signiﬁcantly diﬀerent) distributed across the control and patient groups. G- MDSC were increased in stable lung transplant recipients vs. healthy control subjects (52.1% [33.3–61.9] vs. 9.4% [7.6–16.4], p = 0.0095) (Figures 1C, 2B). Lung transplant recipients with an infection or CLAD tended to have lower percentage of G-MDSC compared to stable recipients (28.2% [17.2–36.6], p = 0.041 and 33.0% [25.6-38.1], p = 0.088, respectively) (Figure 2B). Within CLAD patients, the proportion of G-MDSC were comparable in BOS (5 cases) and RAS (4 cases) (p = 0.99) (Figure 2C). G-MDSC percentages seemed to increase with increasing blood levels April 2019 \| Volume 10 \| Article 900 Frontiers in Immunology \| www.frontiersin.org 4 Heigl et al. Myeloid-Derived Suppressor Cells in Lung Transplantation FIGURE 3 \| G-MDSCs isolated from lung transplant patients functionally supress T cell proliferation. The suppressive effect of CD66b+-MACS-isolated MDSCs (isolated from lung transplant recipients; 1 with CLAD and 1 stable) on CFSE labeled T cell CD4+ (green) and CD8+ (purple) proliferation. (A) Different ratios of MDSC vs. T cells (1:6 and 1:2) were assessed and compared with T cell proliferation without MDSCs. P0 represents undivided cells, P1 cells divided 1 time; P2 cells divided twice and so on. T cell proliferation ratio is portion of divided cells over all cells. The bar graphs represent the proliferation index compared to control conditions (n = 2). FIGURE 3 \| G-MDSCs isolated from lung transplant patients functionally supress T cell proliferation. Statistical Analysis The suppressive effect of CD66b+-MACS-isolated MDSCs (isolated from lung transplant recipients; 1 with CLAD and 1 stable) on CFSE labeled T cell CD4+ (green) and CD8+ (purple) proliferation. (A) Different ratios of MDSC vs. T cells (1:6 and 1:2) were assessed and compared with T cell proliferation without MDSCs. P0 represents undivided cells, P1 cells divided 1 time; P2 cells divided twice and so on. T cell proliferation ratio is portion of divided cells over all cells. The bar graphs represent the proliferation index compared to control conditions (n = 2). cells also expressed CD11b, CXCR4 and HLA-DRlow. The absence of CD14 expression conﬁrmed their G-MDSC phenotype (Figure S1). of the calcineurin inhibitors (Tacrolimus r² = 0.17, p = 0.12; Cyclosporine r² = 0.18, p = 0.39) used as immunosuppressive therapy, which however was not signiﬁcant most probably due to the small sample size (Figures 2D,E). MDSCs are known for their role in immune regulation and allograft acceptance, and are involved in delayed graft rejection (17, 24, 25). Our data showed an expansion of G- MDSCs (not M-MDSCs) in stable lung transplant recipients and a decrease of G-MDSCs in patients with CLAD. Lung transplant recipients suﬀering from an infection also demonstrated a reduction in G-MDSCs, pointing to the fact that infection interferes with immune regulation and allograft acceptance. For example, it has been shown in mice that CMV infection impairs MDSC diﬀerentiation (26). CMV is a clinically relevant post- transplant pathogen, which is considered as a risk factor for later development of CLAD (27) Also in our study population, we found that recipients with diagnosed CMV within the infection group showed a lower G-MDSC percentage compared to the other patient groups (data not shown). G-MDSCs isolated from lung transplant patients eﬀectively suppressed T-cell proliferation in a CFSE based polyclonal proliferation assay. The T-cell suppression assay was used as a proof-of-concept assay to demonstrate that G-MDSCs expanded in transplant recipient patients indeed represent a suppressive G-MDSC cell type and do not reﬂect myeloid cell populations with G-MDSC-like markers, but without T cell suppressive activities. Isolated patient G-MDSCs exhibited a strong suppressive function on T cell proliferation of about 50 and 80% with a 1:6 and 1:2 ratio of MDSC, vs. CD4+ or CD8+ T cells, respectively (Figure 3). FUNDING BV and GV are supported by a research grant of KU Leuven, Belgium (C2/15/030) and by the Research Foundation Flanders (FWO) (G083818N); GV is supported by a research grant from the Broere Charitable Foundation; RV is supported by a research grant of UZ Leuven, Belgium (STG15/023); RV and SV are senior research fellows of the Research Foundation Flanders (FWO), Belgium. There are several limitations to our study. As a pilot study, the number of studied patients is limited. Furthermore, there are several confounding factors such as the heterogeneity of patient characteristics, diﬀerences in immunosuppressive therapy, use of azithromycin, diﬀerent blood sampling tubes and diﬀerent timings of sampling after lung transplantation. ETHICS STATEMENT (17, 28). Calcineurin inhibitors are indispensable in lung transplantation as eﬃcient immunosuppressive drugs to block the immune response toward the allograft; hence, induction of MDSCs and their immunosuppressive function might be a part of their mechanism of action. It has been shown in a mouse skin transplant model that mechanistically, CsA treatment enhances the expression of indoleamine 2,3-dioxygenase (IDO) and thereby induces the suppressive activities of MDSCs in allograft recipients (29). Since the myeloid compartment consists of many diﬀerent cell types with often overlapping phenotypic markers, we wanted to assess if the G-MDSCs, isolated from our lung transplant population, demonstrated suppressive eﬀector properties. We conﬁrmed that G-MDSCs did exert CD4+ and CD8+ T cell suppression in two independent patient samples. Due to the low number of replications, we can only speculate that in the setting of transplant immunology, G-MDSCs would act upstream of T cells to induce a cascade of peripheral tolerance toward the graft tissue. Challenging from a technical standpoint was the diﬀerence observed between the Lithium-Heparin and EDTA coated blood-drawing tubes used for PBMC isolations and the resulting diﬀerences in G-MDSC. At this point, we speculate that EDTA, as an iron chelator, inhibits cell degranulation, and may be the reason why more G-MDSC can be measured when using EDTA coating compared to Lithium-Heparin, at least in our experimental settings. However, it is important to mention that in a study by Pallet et al., the opposite eﬀect, increased G- MDSC counts in Heparin vs. EDTA tubes, has been observed (30), which thus needs further investigation. This study included 20 lung transplant recipients and 4 healthy controls recruited at the University Hospitals Leuven (Belgium). All lung transplant recipients gave informed consent at time of listing for transplantation and routine blood sampling was approved by the University hospital (S51577). DISCUSSION MDSCs were evaluated in lung transplant recipients and G-MDSC (CD33+/CD66b+) could be identiﬁed in the low- density fraction of PBMCs. G-MDSC (CD33+/CD66b+) Furthermore, we evaluated the eﬀect of immunosuppression on G-MDSCs: G-MDSCs showed a modest correlation with increasing CNI trough levels, a previously reported phenomenon April 2019 \| Volume 10 \| Article 900 Frontiers in Immunology \| www.frontiersin.org 5 Heigl et al. Myeloid-Derived Suppressor Cells in Lung Transplantation SUPPLEMENTARY MATERIAL However, our ﬁndings remain interesting, and may warrant more in-depth research on the role of G-MDSCs in lung transplantation. In our opinion, elucidating the functional hierarchy of immune regulatory cells in the context of transplant tolerance/rejection is of importance to understand graft acceptance. We believe that the up-stream suppressive activity of G-MDSC may be an intriguing starting point to dissect this highly complex interconnected immune regulatory system consisting of Treg, Bregs, Mregs, and other cell types. The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ﬁmmu. 2019.00900/full#supplementary-material Figure S1 \| Representative dot plots and histograms of surface marker proﬁling of G-MDSCs isolated from lung transplant patients. MDSCs were analyzed in peripheral blood mononuclear cells (PBMCs) isolated from low density fraction of whole blood after Ficoll density centrifugation. Lung transplant MDSCs exhibited characteristic G-MDSC phenotype of CD33b+CD14−cells (left panel, also see Figure 1) and also expressed CXCR4 (right panel). Histograms show individual surface marker staining (red) in comparison to unstained control (black) for CXCR4. AUTHOR CONTRIBUTIONS TH and ASi performed this study from sampling, analyzing, writing, and submitting the paper. ASi performed the MDSC FACS analysis, the in vitro MDSC proliferation test and the writing of the paper. JK, ASa, and SV helped in organizing the sampling and writing of the paper. DH helped in writing of the paper. BS-G, HB, AV, and AVH helped by searching for the clinical patient information and helped preparing the data. DV was the surgeon performing the lung translations and critical evaluated the manuscript. GV and RV were the clinicians performing the daily care taking of the patients and critical evaluated the manuscript. BV and RV designed and funded the study. 4. Gabrilovich DI, Nagaraj S. Myeloid-derived suppressor cells as regulators of the immune system. Nat Rev Immunol. (2009) 9:162–74. doi: 10.1038/nri2506 REFERENCES 4. Gabrilovich DI, Nagaraj S. Myeloid-derived suppressor cells as regulators of the immune system. Nat Rev Immunol. (2009) 9:162–74. doi: 10.1038/nri2506 4. Gabrilovich DI, Nagaraj S. Myeloid-derived suppressor cells as regulators of the immune system. Nat Rev Immunol. (2009) 9:162–74. doi: 10.1038/nri2506 1. Wood KJ, Bushell A, Hester J. Regulatory immune cells in transplantation. Nat Rev Immunol. (2012) 12:417–30. doi: 10.1038/nri3227 5. Bunt SK, Sinha P, Clements VK, Leips J, Ostrand-Rosenberg S. Inﬂammation induces myeloid-derived suppressor cells that facilitate tumor progression. J Immunol. (2006) 176:284–90. doi: 10.4049/jimmunol.176.1.284 5. Bunt SK, Sinha P, Clements VK, Leips J, Ostrand-Rosenberg S. Inﬂammation induces myeloid-derived suppressor cells that facilitate tumor progression. J Immunol. (2006) 176:284–90. doi: 10.4049/jimmunol.176.1.284 2. Gabrilovich DI, Bronte V, Chen S-H, Colombo MP, Ochoa A, Ostrand- Rosenberg S, et al. The terminology issue for myeloid-derived suppressor cells. Cancer Res. (2007) 67:425. doi: 10.1158/0008-5472.CAN-06-3037 6. Gabrilovich DI, Ostrand-Rosenberg S, Bronte V. Coordinated regulation of myeloid cells by tumours. Nat Rev Immunol. (2012) 12:253–68. doi: 10.1038/nri3175 6. Gabrilovich DI, Ostrand-Rosenberg S, Bronte V. Coordinated regulation of myeloid cells by tumours. Nat Rev Immunol. (2012) 12:253–68. doi: 10.1038/nri3175 3. Ostrand-Rosenberg S, Sinha P. Myeloid-derived suppressor cells: linking inﬂammation and cancer. J Immunol. (2009) 182:4499– 506. doi: 10.4049/jimmunol.0802740 3. Ostrand-Rosenberg S, Sinha P. Myeloid-derived suppressor cells: linking inﬂammation and cancer. J Immunol. (2009) 182:4499– 506. doi: 10.4049/jimmunol.0802740 7. Youn J-I, Collazo M, Shalova IN, Biswas SK, Gabrilovich DI. Characterization of the nature of granulocytic myeloid-derived suppressor cells in 7. Youn J-I, Collazo M, Shalova IN, Biswas SK, Gabrilovich DI. Characterization of the nature of granulocytic myeloid-derived suppressor cells in April 2019 \| Volume 10 \| Article 900 Frontiers in Immunology \| www.frontiersin.org 6 Heigl et al. Myeloid-Derived Suppressor Cells in Lung Transplantation tumor-bearing mice. J Leukoc Biol. (2012) 91:167–81. doi: 10.1189/ jlb.0311177 tumor-bearing mice. J Leukoc Biol. (2012) 91:167–81. doi: 10.1189/ jlb.0311177 21. Brandau S, Trellakis S, Bruderek K, Schmaltz D, Steller G, Elian M, et al. Myeloid-derived suppressor cells in the peripheral blood of cancer patients contain a subset of immature neutrophils with impaired migratory properties. J Leukoc Biol. (2011) 89:311–7. doi: 10.1189/jlb.0310162 8. Umansky V, Blattner C, Gebhardt C, Utikal J. The Role of Myeloid- Derived Suppressor Cells (MDSC) in cancer progression. Vaccines. (2016) 4:E36. doi: 10.3390/vaccines4040036 22. Rieber N, Brand A, Hector A, Graepler-Mainka U, Ost M, Schäfer I, et al. REFERENCES Flagellin induces myeloid-derived suppressor cells: implications for pseudomonas aeruginosa infection in cystic ﬁbrosis lung disease. J Immunol. (2013) 190:1276–84. doi: 10.4049/jimmunol.1202144 9. Rodriguez PC, Hernandez CP, Quiceno D, Dubinett SM, Zabaleta J, Ochoa JB, et al. Arginase I in myeloid suppressor cells is induced by COX- 2 in lung carcinoma. J Exp Med. (2005) 202:931–9. doi: 10.1084/jem. 20050715 23. Singh A, Lelis F, Braig S, Schäfer I, Hartl D, Rieber N. Diﬀerential regulation of myeloid-derived suppressor cells by Candida species. Front Microbiol. (2016) 7:1624. doi: 10.3389/fmicb.2016.01624 10. Nathan C, Cunningham-Bussel A. Beyond oxidative stress: an immunologist’s guide to reactive oxygen species. Nat Rev Immunol. (2013) 13:349– 61. doi: 10.1038/nri3423 24. Wu T, Zhao Y, Zhao Y. The roles of myeloid-derived suppressor cells in transplantation. Expert Rev Clin Immunol. (2014) 10:1385– 94. doi: 10.1586/1744666X.2014.948424 11. Nefedova Y, Huang M, Kusmartsev S, Bhattacharya R, Cheng P, Salup R, et al. Hyperactivation of STAT3 is involved in abnormal diﬀerentiation of dendritic cells in cancer. J Immunol. (2004) 172:464–74. doi: 10.4049/jimmunol.172.1.464 25. Escors D, Liechtenstein T, Perez-Janices N, Schwarze J, Dufait I, Goyvaerts C, et al. Assessing T-cell responses in anticancer immunotherapy: dendritic cells or myeloid-derived suppressor cells? Oncoimmunology. (2013) 2:e26148. doi: 10.4161/onci.26148 j 12. Rébé C, Végran F, Berger H, Ghiringhelli F. STAT3 activation: a key factor in tumor immunoescape. JAKSTAT. (2013) 2:e23010. doi: 10.4161/ jkst.23010 26. Dangi A, Zhang L, Zhang X, Luo X. Murine CMV induces type 1 IFN that impairs diﬀerentiation of MDSCs critical for transplantation tolerance. Blood Adv. (2018) 2:669–80. doi: 10.1182/bloodadvances.2017 012187 13. Highﬁll SL, Rodriguez PC, Zhou Q, Goetz CA, Koehn BH, Veenstra R, et al. Bone marrow myeloid-derived suppressor cells (MDSCs) inhibit graft-versus-host disease (GVHD) via an arginase-1-dependent mechanism that is up-regulated by interleukin-13. Blood. (2010) 116:5738–47. doi: 10.1182/blood-2010-06-287839 27. Paraskeva M, Bailey M, Levvey BJ, Griﬃths AP, Kotsimbos TC, Williams TP, et al. Cytomegalovirus replication within the lung allograft is associated with bronchiolitis obliterans syndrome. Am J Transplant. (2011) 11:2190– 6. doi: 10.1111/j.1600-6143.2011.03663.x 14. Meng F, Chen S, Guo X, Chen Z, Huang X, Lai Y, et al. Clinical signiﬁcance of myeloid-derived suppressor cells in human renal transplantation with acute T cell-mediated rejection. Inﬂammation. (2014) 37:1799– 805. doi: 10.1007/s10753-014-9910-5 28. Yang F, Li Y, Zhang Q, Tan L, Peng L, Zhao Y. The eﬀect of immunosuppressive drugs on MDSCs in transplantation. J Immunol Res. (2018) 2018:5414808. doi: 10.1155/2018/5414808 15. Frontiers in Immunology \| www.frontiersin.org REFERENCES Luan Y, Mosheir E, Menon MC, Wilson D, Woytovich C, Ochando J, et al. Monocytic myeloid-derived suppressor cells accumulate in renal transplant patients and mediate CD4(+) Foxp3(+) Treg expansion. Am J Transplant. (2013) 13:3123–31. doi: 10.1111/ajt.12461 29. Wang X, Bi Y, Xue L, Liao J, Chen X, Lu Y, et al. The calcineurin- NFAT axis controls allograft immunity in myeloid-derived suppressor cells through reprogramming T cell diﬀerentiation. Mol Cell Biol. (2015) 35:598– 609. doi: 10.1128/MCB.01251-14 16. Natarajan S, Thomson AW. Tolerogenic dendritic cells and myeloid-derived suppressor cells: potential for regulation and therapy of liver auto- and alloimmunity. Immunobiology. (2010) 215:698–703. doi: 10.1016/j.imbio.2010.05.024 30. Pallett LJ, Gill US, Quaglia A, Sinclair LV, Jover-Cobos M, Schurich A, et al. Metabolic regulation of hepatitis B immunopathology by myeloid- derived suppressor cells. Nat Med. (2015) 21:591–600. doi: 10.1038/ nm.3856 17. Han C, Wu T, Na N, Zhao Y, Li W, Zhao Y. The eﬀect of immunosuppressive drug cyclosporine A on myeloid-derived suppressor cells in transplanted mice. Inﬂamm Res. (2016) 65:679–88. doi: 10.1007/s00011-016-0949-7 Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 18. Okano S, Abu-Elmagd K, Kish DD, Keslar K, Baldwin WM, Fairchild RL, et al. Myeloid-derived suppressor cells increase and inhibit donor-reactive T cell responses to graft intestinal epithelium in intestinal transplant patients. Am J Transplant. (2018) 18:2544–58. doi: 10.1111/ajt.14718 Copyright © 2019 Heigl, Singh, Saez-Gimenez, Kaes, Van Herck, Sacreas, Beeckmans, Vanstapel, Verleden, Van Raemdonck, Verleden, Vanaudenaerde, Hartl and Vos. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 19. Meyer KC, Raghu G, Verleden GM, Corris PA, Aurora P, Wilson KC, et al. An international ISHLT/ATS/ERS clinical practice guideline: diagnosis and management of bronchiolitis obliterans syndrome. Eur Respir J. (2014) 44:1479–503. doi: 10.1183/09031936.00107514 20. Verleden GM, Raghu G, Meyer KC, Glanville AR, Corris P. A new classiﬁcation system for chronic lung allograft dysfunction. J Heart Lung Transplant. (2014) 33:127–33. REFERENCES doi: 10.1016/j.healun.2013.10.022 April 2019 \| Volume 10 \| Article 900 Frontiers in Immunology \| www.frontiersin.org 7
https://openalex.org/W3173945610	https://www.preprints.org/manuscript/202106.0650/v1/download	English	null	Cellular and Humoral Immunogenicity of a Candidate DNA Vaccine Expressing SARS-CoV-2 Spike Subunit 1	Vaccines	2,021	cc-by	6,151	Cellular and humoral immunogenicity of a candidate DNA vac- cine expressing SARS-CoV-2 spike subunit 1 Khalid A Alluhaybi1,2, Rahaf H Alharbi1, Rowa Y Alhabbab1,3, Najwa D Aljehani1,4, Sawsan S Alamri1,4, Moham- mad Basabrain1, Rehaf Alharbi1, Wesam Abdulaal4, Mohamed A Alfaleh1,2, Levi Tamming5,6, Wanyue Zhang5,6, Ma- zen Hassanain7, Abdullah Algaissi8,9, Adel M. Abuzenadah3, Xuguang Li5,6, Anwar M Hashem1,10* 1 Vaccines and Immunotherapy Unit, King Fahd Medical Research Center, King Abdulaziz University, Jeddah, Saudi Arabia 1 Vaccines and Immunotherapy Unit, King Fahd Medical Research Center, King Abdulaziz University, Jeddah, Saudi Arabia 1 Vaccines and Immunotherapy Unit, King Fahd Medical Research Center, King Abdulaziz University, Jeddah, Saudi Arabia 2 Faculty of Pharmacy, King Abdulaziz University, Jeddah, Saudi Arabia 2 Faculty of Pharmacy, King Abdulaziz University, Jeddah, Saudi Arabia 3 Department of Medical Laboratory Technology, Faculty of Applied Medical Sciences, King Abdulaziz Univer- sity, Jeddah, Saudi Arabia 5 Centre for Biologics Evaluation, Biologics and Radiopharmaceutical Drug Directorate, Health Products and Food Branch (HPFB), Health Canada and WHO Collaborating Center for Standardization and Evaluation of Biologicals, Ottawa, ON, Canada 6 Department of Biochemistry, Microbiology and Immunology, Faculty of Medicine, University of Ottawa, Ot- tawa, ON, Canada 7 Department of Surgery, Faculty of Medicine, King Saud University, Riyadh, Saudi Arabia 8 Department of Medical Laboratories Technology, College of Applied Medical Sciences, Jazan University, Jazan, Saudi Arabia 9 Medical Research Center, Jazan University, Jazan, Saudi Arabia dical Research Center, Jazan University, Jazan, Saudi Arabia 10 Department of Medical Microbiology and Parasitology, Faculty of Medicine, King Abdulaziz University, Jed- dah, Saudi Arabia Correspondence: Anwar M Hashem; Faculty of Medicine and King Fahd Medical Research Center, King Ab- dulaziz University, P.O.Box 80205, Jeddah 21859, Saudi Arabia. Tel. +966 (12) 6400000 ext. 21033, E-mail: am- hashem@kau.edu.sa Abstract: The urgent need for effective, safe and equitably accessible vaccines to tackle the ongoing spread of COVID-19 led researchers to generate vaccine candidates targeting varieties of immuno- gens of SARS-CoV-2. Because of its crucial role in mediating binding and entry to host cell and its proven safety profile, the subunit 1 (S1) of the spike protein represents an attractive immunogen for vaccine development. Here, we developed and assessed the immunogenicity of a DNA vaccine en- coding the SARS-CoV-2 S1. Following in vitro confirmation and characterization, the humoral and cellular immune responses of our vaccine candidate (pVAX-S1) was evaluated in BALB/c mice using two different doses, 25 µg and 50 µg. Cellular and humoral immunogenicity of a candidate DNA vac- cine expressing SARS-CoV-2 spike subunit 1 Our data showed high levels of SARS-CoV-2 specific IgG and neutralizing antibodies in mice immunized with three doses of pVAX-S1. Analysis of the induced IgG subclasses showed a Th1-polarized immune response as demonstrated by the significant eleva- tion of spike-specific IgG2a and IgG2b compared to IgG1. Furthermore, we found that immuniza- tion of mice with three doses of 50 µg of pVAX-S1 could elicit significant memory CD4+ and CD8+ T cell responses. Taken together, our data indicates that pVAX-S1 is immunogenic and safe in mice and is worthy of further preclinical and clinical evaluation. Keywords: SARS-CoV-2 vaccine; cellular and humoral immunogenicity; DNA vaccine Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0 Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0 Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 Article 2.1 Generation of DNA construct SARS-CoV-2 S1 coding sequence was PCR amplified from codon-optimized full- length S gene (GenBank accession number: MN908947) synthesized by GenScript USA Inc. (Piscataway, NJ). The coding sequence of S1 (1-681 aa) was subcloned into the mam- malian expression vector pVAX1 under the control of the cytomegalovirus immediate- early promoter, denoted as pVAX-S1. The construct was cloned between NheI and KpnI restriction sites. The construct was then confirmed by restriction digestion and sequenc- ing. Bulk endotoxin-free preparations of pVAX-S1 and empty vector (pVAX) were pre- pared using the GenElute™ HP Select Plasmid Gigaprep Kit (Sigma, Germany) for animal studies. Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 acute respiratory infections and deaths. As of June 2021, around 178 million confirmed cases have been reported with at least 3.8 million deaths [5]. Furthermore, the high trans- missibility rate of SARS-CoV-2 among humans as well as the emergence of new variants of concern (VOC) of the virus pose significant obstacles toward controlling its spread [6- 8], highlighting the urgent need for the development of safe, effective and equitably ac- cessible vaccines. Hundreds of SARS-CoV-2 vaccines have been or being developed using traditional and innovative technology platforms [9]. Several of these vaccines were approved for emergency use by multiple regulatory agencies across the globe. Examples of traditional vaccines include killed/inactivated vaccines which although demonstrated safety and ef- ficacy in humans, several potential risks may still exist [10]. Other developers have adopted innovative technologies and/or novel approaches for antigen design, gene ex- pression and vector optimization including adenovirus-based vaccines [11-14] as well as mRNA vaccines [15]. Additional platforms being used include novel viral vectors, recom- binant subunit proteins, nanoparticles and plasmid DNA [16-18]. The spike (S) protein of SARS-CoV-2 is composed of a globular head S1 subunit con- taining the receptor-binding domain (RBD), and a membrane-proximal S2 subunit con- taining the fusion machinery of the virus [19]. Most of these aforementioned vaccines rely on using either the full-length S protein or the RBD as the immunogen because of their critical roles in viral entry and host tropism [19,20], and ability to elicit protective immun- ity in animals and humans after vaccination or infection [9, 10, 21]. Use of full-length S protein as immunogen, however, could be associated with undesired responses by induc- ing non-neutralizing antibodies which may contribute to disease enhancement, immuno- pathological inflammation and fatality [22-28]. As such, targeting the S1 subunit could help minimize potentially undesirable effect. Here, we evaluated the humoral and cellular immunogenicity of a plasmid DNA vaccine candidate expressing the S1 subunit of the S protein in an attempt to focus the immune response towards the neutralizing-epitope rich domains. 1. Introduction The emergence and rapid spread of the severe acute respiratory syndrome-corona- virus-2 (SARS-CoV-2), the causative agent of the coronavirus disease 2019 (COVID-19) pandemic, represent a serious public health and economic burden to humanity [1-4]. While the majority of COVID-19 patients are either asymptomatic carriers or have mild symptoms such as fever, myalgia and cough, millions have suffered from life-threatening 2.4 Immunization and samples collection 6- to 8-week-old female BALB/c mice were provided from King Fahd Medical Re- search Center (KFMRC) core animal facility, King Abdulaziz University (KAU). All ani- mal experiments were done according to guidelines and the approval of the Animal Care and Use Committee (ACUC) at KFMRC and ethical approval from the bioethical commit- tee at KAU (approval number 04-CEGMR-Bioeth-2020). Mice were randomly divided into three experimental groups (8 mice/group) and immunized with three doses of 25 μg or 50 μg of pVAX-S1, or 50 μg of pVAX on days 0, 21 and 42. Mice were immunized intramus- cularly using customized needle-free Tropis system (PharmaJet, Golden, CO). Serum sam- ples were collected every three weeks and mice were euthanized on day 63 to collect final bleed and spleens for immune response analysis. 2.5 Binding antibodies measurement by indirect ELISA. End-point titers of anti-S1 total IgG or its isotypes (IgG1, IgG2a and IgG2b) from immunized mice were determined by ELISA as previously described [29]. Briefly, 96-well plates were coated with SARS-CoV-2 S1 protein (Sino Biological, China) at 1 μg/ml in at 4°C overnight. Plates were washed three times with PBS containing 0.1% Tween-20 before blocking with 5% skim milk in PBS-T for 1 h at room temperature. After washing, 2-fold serial dilution of mouse sera starting from 1:100 were added to wells and incubated for 1 h at 37°C. Then, peroxidase-conjugated rabbit anti-mouse IgG secondary antibodies (Sigma, Germany) were added at the recommended concentrations and incubated for 1 h at 37°C. After extensive washing, 3,3’,5,5’-tetramethylbenzidine (TMB) substrate (KPL, Gaithersburg, MD) was added for 30 min to develop a colorimetric reaction. Finally, reac- tion was stopped with 0.16 M sulfuric acid, and absorbance was read spectrophotometri- cally at 450 nm on ELx808™ Microplate Reader (BioTek, Winooski, VT). End-point titers were determined as the reciprocals of the highest dilution with an OD above the cut-off value which was defined as the OD mean from the control group plus three standard deviations (SDs). 2.3 Detection of SARS-CoV-2 S1 protein expression by immunofluorescence 2.3 Detection of SARS-CoV-2 S1 protein expression by immunofluorescence HEK-293A cells were seeded on cell culture slide and incubate at 37°C in a 7% CO2 incubator to be 70% confluent by the next day. Cells were then transfected with 1 μg of either pVAX-S1 or pVAX control plasmid using Lipofectamine 2000 Transfection Reagent (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions and incubated at 37°C in a 7% CO2 incubator for 36 h. The media was removed, cells were washed with PBS, fixed with 4% formaldehyde at 4°C for 10 min and permeabilized with 0.2% PBS-Triton 100 (PBS-T) at 4°C for 20 min. Cells were washed twice with PBS-T and blocked with 2% goat serum/PBS-T at room temperature for 30 min. Cells were then stained with in-house mouse anti-S1 primary polyclonal antibodies at a 1:1000 dilution in 2% goat serum/PBS-T at 4°C for 1 h. This was followed by three washes and staining with Alexa Fluor-488 la- beled goat anti-mouse IgG H&L secondary antibody (Abcam, UK) at 1:500 dilution in blocking buffer in the dark at room temperature for 1 h. Cells were finally washed three times with PBS-T and mounted with VECTASHIELD with DAPI counterstain antifade mounting medium (Invitrogen, Carlsbad, CA). Images were captured using Olympus BX51 Fluorescence Microscope and analyzed using ImageJ 1.53e Software. 2.4 Immunization and samples collection 2.2 Detection of SARS-CoV-2 S1 protein expression by Western blot HEK-293A cells (70-90% confluent) in 6-well plates were transfected with 2 μg of pVAX-S1, pVAX1, or pcDNA3.1 expressing full-length S protein (pcDNA-S) using Lipofectamine 2000 Transfection Reagent (Invitrogen, Carlsbad, CA) according to manu- facturer’s instructions, followed by incubation at 37°C in a 7% CO2 incubator for 48 h. After that, media were removed, and transfected cells were then washed with phosphate-buff- ered saline (PBS) and lysed with radioimmunoprecipitation assay buffer (RIPA buffer) (Sigma, Germany). The lysates were subjected to western blot analysis to test protein ex- pression using in-house anti-S1 (SARS-CoV-2) mouse polyclonal antibodies at a 1:1000 Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 dilution. Also, β-actin was detected with anti β-actin antibodies at a 1:3000 dilution (Ori- Gene Technologies, Inc., Rockville, MD) as a loading control. dilution. Also, β-actin was detected with anti β-actin antibodies at a 1:3000 dilution (Ori- Gene Technologies, Inc., Rockville, MD) as a loading control. 2.8 Statistical analysis Statistical analysis and graphical presentations were generated using GraphPad Prism version 9.0.2 software (Graph-Pad Software, Inc., CA, USA). Statistical analysis was conducted using the Mann–Whitney test or one-way analysis of variance with Bonferroni post-hoc test to adjust for multiple comparisons between groups. All values are repre- sented as mean ± SD and statistical significance is reported as , P≤0.05, , P≤0.01, , p ≤ 0.001, and **, p ≤ 0.0001. Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 vesicular stomatitis virus expressing codon-optimized full-length SARS-CoV-2 S protein (GenBank accession number: MN908947) (rVSV-ΔG/SARS-2-S-luciferase pseudovirus) was generated in BHK21/WI-2 cells. Pseudovirus was collected and titrated on Vero E6 cells as previously described [30]. Then, neutralization assay was conducted by co-incu- bating two-fold serial dilutions of heat-inactivated mouse sera from vaccinated and con- trol groups started from 1:20 dilution (in duplicate) with media containing 5x104 relative luciferase unit (RLU) of rVSV-ΔG/SARS- 2-S-luciferase pseudovirus for 1 h at 37°C in a 7% CO2 incubator. The mixtures were then transferred onto confluent Vero E6 cell mono- layers in white 96-well plates and incubated for 24 h at 37°C in a 7% CO2 incubator. Fol- lowing that, cells were lysed, luciferase activity was measured using Luciferase Assay System (Promega) according to the manufacturer’s instructions, and luminescence activ- ity was measured using BioTek Synergy 2 microplate reader (BioTek, Winooski, VT). Cell only control (CC) and virus control (VC) were included with each assay run. The median inhibitory concentration (IC50) of neutralizing antibodies (nAbs) was determined using GraphPad Prism version 9.0.2 software. 2.7 Determination of T cell response by flow cytometry Single cell suspensions of splenocytes were prepared from each mouse in immun- ized and control groups. One million splenocytes/well were re-stimulated with 5 μg/ml of a pool of 15-mer peptides overlapping with 11 amino acid and covering the SARS-CoV- 2 S1 protein (GenScript USA Inc, Piscataway, NJ) for 6 h at 37°C in a 7% CO2 incubator in the presence of brefeldin A (BD Biosciences, San Jose, CA) at a final concentration of 1:1000. Phorbol myristate acetate/ionomycin was used as a positive control and RPMI 1640 medium as a negative unstimulated control. Cells were then washed in FACS buffer (PBS with 2% heat inactivated FBS) and stained with LIVE/DEAD Fixable Violet Dead Cell Stain Kit (Invitrogen, Carlsbad, CA) for 30 min at room temperature. After washing with FACS buffer, cells were stained for surface markers with PB-conjugated anti-mouse CD8, PB-conjugated anti-mouse CD4, APC-conjugated anti-mouse CD44 antibody and Pe-Cy7- conjugated anti-mouse CD62L antibodies (BioLegend, UK). The cells were then washed with FACS buffer and fixed and permeabilized using Cytofix/Cytoperm Solution (BD Bi- osciences, San Jose, CA) according to the manufacturer’s protocol. For intracellular stain- ing, cells were labeled with FITC-conjugated anti–mouse IFN-γ (clone XMG1.2), PE-con- jugated anti-mouse TNF-α (clone MP6-XT22) and PerCP/Cy5.5–conjugated anti-mouse IL-2 (clone JES6-5H4) antibodies (BioLegend, UK) for 20 min at 4°C. Cells were then washed twice with permeabilization buffer and once with FACS buffer. All data were col- lected using BD FACSAria™ III flow cytometer (BD Biosciences, San Jose, CA) and ana- lyzed using FlowJo v10 software (Tree Star). 2.6 Neutralizing antibodies measurement by pseudovirus neutralization assay To assess the ability of induced antibodies to inhibit virus entry, pseudovirus neu- tralization assay was performed as previously described [30]. Briefly, recombinant Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 The generated DNA vaccine candidate (Figure 1a) was evaluated for protein expression in vitro in HEK293A cells prior to animal experiments. As shown in Figure 1b, western blot analysis confirmed that the recombinant construct was able to express S1 subunit protein at the expected molecular weight. A plasmid expressing full-length S protein (pcDNA3.1-Full S) was used as a positive control. Similarly, immunofluorescence analysis showed expression of SARS-CoV-2 S1 protein in transfected cells (Figure 1c), suggesting that the expressed protein maintained structural confirmation to be detected by polyclo- nal anti-S antibodies. As expected, no protein was detected from cells transfected with the empty control plasmid pVAX (Figures 1b and 1c). Figure 1. Vaccine design and characterization. a) DNA vaccine (pVAX-S1) map. The inserted gene (SARS-CoV-2 S subunit 1) is indicated by orange color in the pVAX1 plasmid. b) Western blot anal- ysis. Figure shows bands of expressed full-length S from cells transfected with pcDNA3.1-Full S (positive control) and S1 subunit protein expressed from pVAX-S1. Empty pVAX was used as neg- ative control. c) Immunofluorescence analysis. Cells transfected with pVAX-S1 or empty control pVAX were stained with anti-SARS-CoV-2 S1 mouse polyclonal antibodies. Scale bars are 50 µm. Red square is magnified to scale bar of 10 µm. Merging and magnification were processed by ImageJ 1.53e. Figure 1. Vaccine design and characterization. a) DNA vaccine (pVAX-S1) map. The inserted gene (SARS-CoV-2 S subunit 1) is indicated by orange color in the pVAX1 plasmid. b) Western blot anal- ysis. Figure shows bands of expressed full-length S from cells transfected with pcDNA3.1-Full S (positive control) and S1 subunit protein expressed from pVAX-S1. Empty pVAX was used as neg- ative control. c) Immunofluorescence analysis. Cells transfected with pVAX-S1 or empty control pVAX were stained with anti-SARS-CoV-2 S1 mouse polyclonal antibodies. Scale bars are 50 µm. Red square is magnified to scale bar of 10 µm. Merging and magnification were processed by ImageJ 1.53e. 3. Results 3.1 In vitro confirmation of protein expression from the candidate vaccine 3.1 In vitro confirmation of protein expression from the candidate vaccine 3.3 Evaluation of cellular immune response in immunized mice Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106 Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 3.2 Evaluation of binding and neutralizing antibodies in immunized mice Mice were intramuscularly immunized with 3 doses of 25 μg or 50 μg of pVAX-S1 in a three-week interval regimen (Figure 2a). As a control, a group of mice was immunized with 50 μg of empty control vector (pVAX). Vaccine-induced binding antibodies were assessed by indirect ELISA from serum samples collected on weeks 3, 6 and 9. As shown in Figure 2b, analysis of S1-specific total IgG showed significant levels only after 3 doses with both doses of pVAX-S1 vaccine compared to control group (pVAX). Although no significant difference was found between 25 μg and 50 μg of pVAX-S1, the 50 μg dose of pVAX-S1 showed significantly higher level of antibodies than the 25 μg dose when both were compared to the control group. Testing levels of IgG subclasses (IgG1, IgG2a and IgG2b) in samples collected on week 9 from mice immunized with 50 µg suggested a Th1- skewed immune response as shown by the high IgG2a:IgG1 and IgG2b:IgG1 ratios (Figure Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 2c). Determination of end-point titers of S1-specific total IgG also confirmed the induction of significant levels of binding antibodies in samples collected on week 9 from each im- munized mouse (Figure 2d). To investigate whether the vaccine-induced antibodies were able to inhibit viral entry in cells, levels of nAbs were determined using pseudovirus neu- tralization assay from samples collected on week 9 in the 50 µg group. As shown in Figure 2e, immunized mice were only able to induce low levels of nAbs against SARS-CoV-2 pseudovirus in Vero cells. Collectively, these results confirm the ability of the vaccine to elicit significant Th1-skwed humoral immunity against SARS-CoV-2. Figure 2. Humoral immunity in pVAX-S1 immunized mice. a) Animal experimental plan. Mice were divided into three experimental groups (n=8) and immunized intramuscularly with three doses on days 0, 21 and 42 using 25 μg or 50 μg of pVAX-S1, or 50 μg of pVAX1 using a customized needle-free Tropis system (PharmaJet, Golden, CO). b) Values of optic density (OD) of S1-specific binding total IgG at 1:100 dilution from every single mouse were determined by ELISA at 3, 6 and 9 weeks after immunizations. c) IgG2a:IgG1 and IgG2b:IgG1 ratios were calculated in samples col- lected on week 9 from immunized mice in the group that received the 50 µg dose of pVAX-S1. d) End-point titers of S1-specific total IgG were determined by ELISA in samples collected on week 9 from each mouse in all groups. e) The median inhibitory concentration (IC50) of neutralizing anti- bodies (nAbs) was determined against rVSV-ΔG/SARS-2-S-luciferase pseudovirus as described in the materials and methods. binding total IgG at 1:100 dilution from every single mouse were determined by ELISA at 3, 6 and 9 weeks after immunizations. c) IgG2a:IgG1 and IgG2b:IgG1 ratios were calculated in samples col- lected on week 9 from immunized mice in the group that received the 50 µg dose of pVAX-S1. d) End-point titers of S1-specific total IgG were determined by ELISA in samples collected on week 9 from each mouse in all groups. e) The median inhibitory concentration (IC50) of neutralizing anti- bodies (nAbs) was determined against rVSV-ΔG/SARS-2-S*-luciferase pseudovirus as described in the materials and methods. 3.3 Evaluation of cellular immune response in immunized mice 4. Discussion There is still an urgent need for multiple safe and protective vaccines against SARS- CoV-2 to combat the ongoing COVID-19 pandemic [9]. DNA-based vaccines represent a fast and safe approach to develop vaccines for such unprecedented situations [31]. Nu- merous studies on SARS-CoV-2 and other pathogenic human CoVs such as MERS-CoV and SARS-CoV have dementated that most of the neutralizing antibodies that are gener- ated to either natural infection or full-length S based vaccines candidates target the S1 subunit, making S1 an attractive and probably safer immunogen for vaccine development [32-36]. This is due to the fact S1 contains the RBD and the N-terminal motif (NTD) which are critical for mediating binding to the host receptor. In this work, we successfully devel- oped and evaluated the immunogenicity of a new DNA vaccine candidate against SARS- CoV-2 encoding the S1 subunit of the S protein. After in vitro confirmation and character- ization of the S1 expression, we evaluated the immunogenicity and safety of the vaccine in mice. Overall, our data showed that pVAX-S1 was able to induce strong antibody re- sponses in mice after three doses regimen of intramuscular immunization in a dose de- pendent manner. Furthermore, we showed that pVAX-S1 induced a Th1-biased protec- tive immune response, characterized by antibody production predominantly of IgG2a and IgG2b subclasses and secretion of significantly elevated levels of Th-1 cytokines (IFN-γ and TNF-α) by memory CD4+ and CD8+ T cells. Interestingly, while the vaccine candidate induced high levels of S1 specific Abs titers, we noticed that the level of the produced nAbs was relatively low which should be investigated for further improvement may be through novel antigen design or use of molecular adjuvants. Nonetheless, similar immune response induced by a similar DNA vaccine provided protection in non-human primates [37], suggesting protection through multiple mechanisms. It is of note that several vaccine candidates have been developed using S1 subunit as immunogen such as DNA vaccine expressing S1 domain with a foldon trimerization motif [37], live-attenuated YF17D expressing S1 (YF-S1) [38], S1-Fc fusion subunit protein [39], and S1 subunit protein alone (S1) or fused to the norovirus shell domain (S1-S) [33]. Those candidates used different technologies to test S1 immunogenicity in numbers of animal models. Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 mouse. Representative FACS plots are shown. Data are shown as mean ± SD for each group from one experiment. mouse. Representative FACS plots are shown. Data are shown as mean ± SD for each group from one experiment. Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 Next, we investigated the memory CD8+ and CD4+ T cell response from 9-week samples (3 weeks post last immunization) using peptide pool covering the entire S1 protein. As shown in Figure 3a, memory CD4+ T cells (CD4+CD62L-CD44+ T cells) from mice immun- ized with 50 μg dose of pVAX-S1 produced significant levels of IFN-γ and TNF-α but not IL-2 compered to control group. Similarly, antigen-specific CD8+CD44+CD62L+ central memory T cells showed significant levels of IFN-γ and TNF-α but not IL-2 compared to pVAX control group (Figure 3b). On the other hand, effector CD8+CD62L-CD44+ memory T cells only produced IFN-γ at significant level compared to control group (Figure 3c). These data show that pVAX-S1 could elicit significant memory CD4+ and CD8+ T cell re- sponses in mice. Figure 3. Memory T cells response in immunized mice. Histograms and FACS plots display IFN- γ, TNF-α and IL-2 expression from ex vivo re-stimulated a) memory CD4 (CD4+CD62L-CD44+ T cells), b) central memory CD8 (CD8+CD62L+CD44+ T cells) and c) effector memory CD8 (CD8+CD62L-CD44+ T cells). Data in histograms are shown as percentages of induced cytokines from peptides-stimulated cells after subtracting levels produced by unstimulated splenocytes from each Figure 3. Memory T cells response in immunized mice. Histograms and FACS plots display IFN- γ, TNF-α and IL-2 expression from ex vivo re-stimulated a) memory CD4 (CD4+CD62L-CD44+ T cells), b) central memory CD8 (CD8+CD62L+CD44+ T cells) and c) effector memory CD8 (CD8+CD62L-CD44+ T cells). Data in histograms are shown as percentages of induced cytokines from peptides-stimulated cells after subtracting levels produced by unstimulated splenocytes from each Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 has not been fully elucidated, it has been postulated that the non-neutralizing epitopes within the S protein may be responsible for the harmful immune response in vaccinated hosts [40, 41]. These data suggested that using the neutralizing-epitope rich S1 subunit of the S protein could be a better approach to avoid any potential safety concerns. Within the past year and half, several COVID-19 vaccines have been approved for emergency use with hundreds of others being currently in different stages of clinical de- velopment [9, 42, 43]. Among the ones that have been approved and in late stages of clin- ical development are the nucleic acid-based vaccines. DNA and mRNA vaccines have sev- eral advantages over the platforms. For example, these vaccines can be developed rapidly without the need for the cultivation of the target pathogen and can be easily produced in a large industrial scale. Furthermore, nucleic acid-based vaccines can be rapidly and easily adapted to respond to potential mutations/variants. Compared to mRNA vaccines which requires a cold-chain system, DNA vaccine are more thermo-stable with less stringent storage condition and easier formulation, enabling it as a promising platform for wide distribution across the globe. Our approach using plasmid DNA (pVAX) to encode the SARS-CoV-2 S1 as a proof of principle have demonstrated that S1 can lead to high hu- moral and cellular immunity in mice with a predominant Th1-biased response. Such ap- proach can be further enhanced by the use of efficient adjuvant and improved method of delivery. Author Contributions: K.A.A., R.Y.A., M.H., A.A., X.L. and A.M.H. designed and conceptualized the work. K.A.A., R.H.A., R.Y.A., N.D.A., S.S.A., M.B., R.H., M.A.A., L.T., and W.Z. performed and optimized experiments and analyzed data. W.A., M.H., A.A., X.L., and A.M.A. provided resources and materials. A.M.A. and A.M.H. acquired the funding. K.A.A., A.A., X.L. and A.M.H. drafted the manuscript. All authors have reviewed and edited the manuscript and agreed to the published ver- sion of the manuscript. Author Contributions: K.A.A., R.Y.A., M.H., A.A., X.L. and A.M.H. designed and conceptualized the work. K.A.A., R.H.A., R.Y.A., N.D.A., S.S.A., M.B., R.H., M.A.A., L.T., and W.Z. performed and optimized experiments and analyzed data. W.A., M.H., A.A., X.L., and A.M.A. provided resources and materials. A.M.A. and A.M.H. acquired the funding. K.A.A., A.A., X.L. and A.M.H. drafted the manuscript. All authors have reviewed and edited the manuscript and agreed to the published ver- sion of the manuscript. 1. Hui, D.S. et al. The continuing 2019-nCoV epidemic threat of novel coronaviruses to global health - The latest 2019 novel coronavirus outbreak in Wuhan, China. Int J Infect Dis. 2020,91,264-266. doi:10.1016/j.ijid.2020.01.009. j j 2. Zheng, B.J.; et al. SARS-related virus predating SARS outbreak, Hong Kong. Emerg Infect Dis. 2004,10,176-178. doi:10.3201/eid1002.030533. 4. Discussion Similar to our work, single-dose of YF-S1 in hamsters or two doses of DNA vac- cine expressing S1-foldon in rhesus macaques induced significant levels of binding anti- bodies and low-to-medium levels of nAbs compared to other tested vaccines expressing full-length cleavable S protein, prefusion-stabilized S or other truncated versions such as those lacking the transmembrane or the cytoplasmic domains [37, 38]. Also consistent with our data, both of these S1-based vaccines induced highly elevated levels of Th1- skewed T cell responses compared to other vaccines [37, 38]. Interestingly, while YF-S1 failed to protect hamsters from viral replication [38], S1-foldon DNA vaccine led to reduc- tion of viral RNA after SARS-CoV-2 challenge in rhesus macaques [37]. On the other hand, other developed subunit vaccines such as S1-Fc and S1-S fusion proteins elicited signifi- cantly high levels of nAbs in multiple animal models that exceeded the levels observed in acutely infected individuals [33, 39]. These previous reports as well as our current data clearly show the potential of SARS-CoV-2 S1 as a promising immunogen [33, 37-39]. The use of S1 as an immunogen has been proposed for other highly pathogenic coro- naviruses such as MERS-CoV and SARS-CoV because its potential high safety profile com- pared to the use of full-length S. Although full-length S protein can induce the highest immune response, some reports suggest its association with possible side effects in the currently used COVID-19 vaccines [22, 23]. Additionally, previous reports on MERS-CoV, SARS-CoV and other coronaviruses have suggested that the use of full-length S based vaccine could lead to undesired immune response upon infection [24-28]. Although, exact mechanism of this vaccination-induced immunopathology and/or disease enhancement Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 Funding: This project was funded by the Deanship of Scientific Research (DSR) at King Abdulaziz University, Jeddah under Grant Number GCV19-43-1441. The authors therefore acknowledge with thanks DSR for financial and technical support. Institutional Review Board Statement: The study was conducted according to the guidelines of the Declaration of Helsinki, and all animal experiments were done according to guidelines and the ap- proval of the Animal Care and Use Committee (ACUC) at KFMRC and ethical approval from the bioethical committee at KAU (approval number 04-CEGMR-Bioeth-2020). Data Availability Statement: The raw data collected in this study are available on request from the corresponding author. Acknowledgments: We would like to thank King Fahd Medical Research Center (KFMRC) and King Abdulaziz University (KAU) for their continuous support. Acknowledgments: We would like to thank King Fahd Medical Research Center (KFMRC) and King Abdulaziz University (KAU) for their continuous support. Conflicts of Interest: A.M.H., M.A.A., K.A.A., S.S.A. and A.A. are named as inventors on a US pa- tent application related to this work. A.M.H. and A.A. work as scientific consultants in SaudiVax ltd and receive fees for consulting. M.H. is an employee of SaudiVax ltd and receives salary and benefits. All other authors report no competing interests. Conflicts of Interest: A.M.H., M.A.A., K.A.A., S.S.A. and A.A. are named as inventors on a US pa- tent application related to this work. A.M.H. and A.A. work as scientific consultants in SaudiVax ltd and receive fees for consulting. M.H. is an employee of SaudiVax ltd and receives salary and benefits. All other authors report no competing interests. References Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 A.; et al. SARS-CoV-2 immunity: review and applications to phase 3 vaccine candidates. L 1606. doi:10.1016/S0140-6736(20)32137-1. 17. Poland, G. A.; et al. SARS-CoV-2 immunity: review and applications to phase 3 vaccine candidates. Lancet. 2020,396,1595- 1606. doi:10.1016/S0140-6736(20)32137-1. 18 Tebas P ; et al Safety and immunogenicity of INO 4800 DNA vaccine against SARS CoV 2: A preliminary report of an 1606. doi:10.1016/S0140-6736(20)32137-1. 18. Tebas, P.; et al. Safety and immunogenicity of INO-4800 DNA vaccine against SARS-CoV-2: A preliminary report of an 18. Tebas, P.; et al. Safety and immunogenicity of INO-4800 DNA vaccine against SARS-CoV-2: A pre open-label, Phase 1 clinical trial. EClinicalMedicine. 2021,31. doi:10.1016/j.eclinm.2020.100689. p j 19. Walls, A.C.; et al. Structure, Function, and Antigenicity of the SARS-CoV-2 Spike Glycoprotein. Cell. 2020,181,281-292.e6. doi:10.1016/j.cell.2020.02.058. j 20. Hoffmann, M.; et al. SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Pro- tease Inhibitor. Cell. 2020;181,271-280.e8. doi:10.1016/j.cell.2020.02.052. j 21. Krammer, F. SARS-CoV-2 vaccines in development. Nature. 2020,586, 516-527. doi:10.1038/s41586-02 22. McNaughton, A. L.; et al. Fatal COVID-19 outcomes are associated with an antibody response targeting epitopes shared with endemic coronaviruses. medRxiv. 2021. doi: https://doi.org/10.1101/2021.05.04.21256571. p g 23. Fang, Y.T.; et al. Annexin A2 on lung epithelial cell surface is recognized by severe acute respiratory syndrome-associated coronavirus spike domain 2 antibodies. Mol Immunol. 2010;47,1000-1009. doi:10.1016/j.molimm.2009.11.019. 24. Weingartl, H.; et al. Immunization with modified vaccinia virus Ankara-based recombinant vaccine against severe acute respiratory syndrome is associated with enhanced hepatitis in ferrets. J Virol. 2004,78,12672–6. doi:10.1128/JVI.78.22.12672- 12676.2004. 25. Jaume, M.; et al. SARS CoV subunit vaccine: antibody-mediated neutralisation and enhancement. Hong Kong Med J. 2012,18 Suppl 2:31-36. pp 26. Jaume, M.; et al. Anti-severe acute respiratory syndrome coronavirus spike antibodies trigger infection of human immune cells via a pH- and cysteine protease-independent FcγR pathway. J Virol. 2011,85,10582-10597. doi:10.1128/JVI.00671-11. 27. Weiss, R.C.; Scott, F.W. Antibody-mediated enhancement of disease in feline infectious peritonitis: comparisons with den- gue hemorrhagic fever. Comp Immunol Microbiol Infect Dis. 1981,4,175-89. doi: 10.1016/0147-9571(81)90003-5. g g p 28. Olsen, C.W. A review of feline infectious peritonitis virus: molecular biology, immunopathogenesis, clinical aspects, and vaccination. Vet Microbiol. 1993,36,1-37. doi:10.1016/0378-1135(93)90126-r. 29. Al-Amri, S.S.; et al. Immunogenicity of Candidate MERS-CoV DNA Vaccines Based on the Spike Pr doi:10.1038/srep44875. 30. Almahboub, S.A.; et al. Evaluation of Neutralizing Antibodies against Highly Pathogenic Coronaviruses: A Detailed Pro- tocol for a Rapid Evaluation of Neutralizing Antibodies Using Vesicular Stomatitis Virus (Vsv) Pseudovirus-Based Assay. Front Microbiol. 2020,4. Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 3. Oboho, I.K.; et al. 2014 MERS-CoV outbreak in Jeddah--a link to health care facilities. N Engl J Med. 2015,372,846-854. doi:10.1056/NEJMoa1408636. J 4. Lu, R.; et al. Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding. Lancet. 2020,395,565-574. doi:10.1016/S0140-6736(20)30251-8. 5. Coronavirus disease (COVID-19) Pandemic. Available online: https://www.who.int/emergencies/diseases/novel-corona- virus-2019. (Accessed on 12-June-2021). 6. Adhikari, S.P.; et al. Epidemiology, causes, clinical manifestation and diagnosis, prevention and control of coronavirus disease (COVID-19) during the early outbreak period: a scoping review. Infect Dis Poverty. 2020,9 doi:10.1186/s40249-020- 00646-x. 7. Bai, Y.; et al. Presumed Asymptomatic Carrier Transmission of COVID-19. JAMA. 2020,323,1406-1407. doi:10.1001/jama.2020.2565. j 8. Xu, Z.; et al. Pathological findings of COVID-19 associated with acute respiratory distress syndrome. Lancet Respir Med. 2020,8,420-422. doi:10.1016/S2213-2600(20)30076-X. 9. Zhang, J.; et al. Progress and Prospects on Vaccine Development against SARS-CoV-2. Vaccines. 2020; 8. doi:10.3390/vac- cines8020153. al. A systematic review of SARS-CoV-2 vaccine candidates. Signal Transduct Target Ther. 2020,5 392-020-00352-y. 10. Dong, Y.; et al. A systematic review of SARS-CoV-2 vaccine candidates. Signal Transduct doi:10.1038/s41392-020-00352-y. 10. Dong, Y.; et al. A systematic review of SARS-CoV-2 vaccine candidates. Signal Transduct Target Ther. 2020,5 doi:10.1038/s41392-020-00352-y. l f d ff f h hAdO A A l y 11. Voysey, M.; et al. Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK. Lancet. 2021,397,99-111. doi:10.1016/S0140- 6736(20)32661-1. ( ) 12. Zhu, F.C.; et al. Immunogenicity and safety of a recombinant adenovirus type-5-vectored COVID-19 vaccine in healthy adults aged 18 years or older: a randomised, double-blind, placebo-controlled, phase 2 trial. Lancet. 2020,396,479-488. doi:10.1016/S0140-6736(20)31605-6. ( ) 13. Sadoff, J.; et al. Interim Results of a Phase 1-2a Trial of Ad26.COV2.S Covid-19 Vaccine. N Engl J Med. 2021,384,1824-1835. doi:10.1056/NEJMoa2034201. J 14. Logunov, D.Y.; et al. Safety and efficacy of an rAd26 and rAd5 vector-based heterologous prime-boost COVID-19 vaccine: an interim analysis of a randomised controlled phase 3 trial in Russia. Lancet. 2021,397,671-681. doi:10.1016/S0140- 6736(21)00234-8. ( ) 15. Vasireddy, D.; et al. Review of COVID-19 Vaccines Approved in the United States of America for Emergency Use. J Clin Med Res. 2021,13, 204-213. doi:10.14740/jocmr4490. j 16. Keech, C.; et al. Phase 1-2 Trial of a SARS-CoV-2 Recombinant Spike Protein Nanoparticle Vaccine. N Engl J Med. 2020,383,2320-2332. doi: 10.1056/NEJMoa2026920. t al. SARS-CoV-2 immunity: review and applications to phase 3 vaccine candidates. Lancet. 2020,396,1595- 6/S0140-6736(20)32137-1. 17. Poland, G. Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi: 10.3389/fmicb.2020.02020. 30. Almahboub, S.A.; et al. Evaluation of Neutralizing Antibodies against Highly Pathogenic Coronaviruses: A Detailed Pro- tocol for a Rapid Evaluation of Neutralizing Antibodies Using Vesicular Stomatitis Virus (Vsv) Pseudovirus-Based Assay. Front Microbiol. 2020,4. doi: 10.3389/fmicb.2020.02020. Preprints (www.preprints.org) \| NOT PEER-REVIEWED \| Posted: 28 June 2021 doi:10.20944/preprints202106.0650.v1 31. Williams, J.A. Vector Design for Improved DNA Vaccine Efficacy, Safety and Production. Vaccines. 2013,1,225-249. https://doi.org/10.3390/vaccines1030225. p g 32. Fernández, A. SARS-CoV-2 Glycosylation Suggests That Vaccines Should Have Adopted the S1 Subunit as Antigen. ACS Pharmacol Transl Sci. 2021,4,1016-1017. doi: 10.1021/acsptsci.1c00036. 33. Wang, Y.; et al. SARS-CoV-2 S1 is superior to the RBD as a COVID-19 subunit vaccine antigen. J Med Virol. 2021,93,892- 898. doi: 10.1002/jmv.26320. 34. Ababneh, M.; et al. Recombinant adenoviral vaccine encoding the spike 1 subunit of the Middle East Respiratory Syndrome Coronavirus elicits strong humoral and cellular immune responses in mice. Vet World. 2019,12,1554-1562. doi: 10.14202/vet- world.2019.1554-1562. 35. Hashem, A.M.; et al. A Highly Immunogenic, Protective, and Safe Adenovirus-Based Vaccine Expressing Middle East Res- piratory Syndrome Coronavirus S1-CD40L Fusion Protein in a Transgenic Human Dipeptidyl Peptidase 4 Mouse Model. J Infect Dis. 2019,8,1558-1567. doi: 10.1093/infdis/jiz137. j 36. Yang, Y.; Du, L. SARS-CoV-2 spike protein: a key target for eliciting persistent neutralizing antibodies. Signal Transduct Target Ther. 2021,6. doi: 10.1038/s41392-021-00523-5. g 37. Yu, J.; et al. DNA vaccine protection against SARS-CoV-2 in rhesus macaques. Science. 2020,369,806-811. doi:10.1126/sci- ence.abc6284. 38. Sanchez-Felipe, L.; et al. A single-dose live-attenuated YF17D vectored SARS-CoV-2 vaccine candidate. Nature. 2021,590,320-325. doi:10.1038/s41586-020-3035-9. 39. Ren, W.; et al. Recombinant SARS-CoV-2 spike S1-Fc fusion protein induced high levels of neutralizing responses in non- human primates. Vaccine. 2020,38,5653-5658. doi:10.1016/j.vaccine.2020.06.066. 40. He, Y.; et al. Identification of immunodominant sites on the spike protein of severe acute respiratory syndrome (SARS) coronavirus: implication for developing SARS diagnostics and vaccines. J Immunol. 2004,173,4050-4057. doi:10.4049/jim- munol.173.6.4050. 41. Hotez, P. J.; et al. Calling for rapid development of a safe and effective MERS vaccine. Microbes Infect. 2014,16, 529-531. doi: 10.1016/j.micinf.2014.05.002. 42. Pushparajah, D.; et al. Advances in gene-based vaccine platforms to address the COVID-19 pandemic. Adv Drug Deliv Rev. 2021,170,113-141. doi:10.1016/j.addr.2021.01.003. j 43. COVID-19 vaccine tracker and landscape. Available online: https://www.who.int/publications/m/item/draft-landscape-of- covid-19-candidate-vaccines. (Accessed on 24-June-2021).
https://openalex.org/W1535100260	https://figshare.com/articles/journal_contribution/Contextual_Influences_in_Visual_Processing/6604454/1/files/12094859.pdf	English	null	Contextual Influences in Visual Processing	Springer eBooks	2,008	cc-by	4,527	Definition Vision is the analysis of patterns in visual images with the view to understanding the objects and the physical processes in the world that generate them. Locally, visual patterns are highly ambiguous and subject to multiple interpretations. Image structures surrounding the pattern being analyzed can provide additional constraints or context to disambiguate the interpreta- tion. The resulting ▶contextual influences are ubiqui- tous in visual perception and manifest at the neuronal level as the modulation of the activity of neurons by image structures outside their ▶classical receptive fields. A variety of extra-classical receptive field effects have been identified. A commonly reported phenom- enon is called ▶surround suppression: the response of a neuron to an oriented bar or grating within its receptive field is suppressed when stimuli are simultaneously introduced to the surrounding area outside its receptive field. There are several types of surround suppression effects, mediated by a number of ▶local circuits as well as ▶recurrent feedback circuits [2]. The early phase of surround suppression is fast and is not sensitive to the exact parameters of the surround stimuli. However, the later phase of surround suppression is stimulus-specific. Simply put, while the neuron can detect the presence of stimuli in the surround immediately, its sensitivity to the precise nature of the surround stimulus or global context takes time to develop. The onset delay of this sensitivity varies considerably depending on the types of the stimuli and the spatial extent of the contextual stimuli. Synonyms Surroundinfluence;Contextualmodulation;Localglobal interaction; Extra-classical receptive field modulation Contextual Influences in the Primary Visual Cortex Neurons in the primary visual cortex receive converg- ing input from the ▶lateral geniculate nucleus (LGN). A neuron’s classical receptive field, also known as the minimum responsive field, is the part of visual space in which the presence of appropriate features can excite the neuron. By definition, stimulating the visual space outside a neuron’s classical receptive field cannot evoke a response. Modulation of neuronal activity by surround stimulation can be observed, however, only when the neuron is responding to a stimulus presented to its receptive field. This modulation is called the non- classical or ▶extra-classical receptive field effect. Such effects have been considered neural manifestations of contextual influences in visual perception. Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:24 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:24 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:24 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: C Contextual Influences in Visual Processing Contextual Influences in Visual Processing Recent approaches seek to understand the neural basis of the perceptual interpretation of the local recep- tive field stimulus by changing the global image context (Fig. 1b). With this approach, a number of neural correlates of perception have been revealed, providing insights into the representation of subjective perceptual experience in the brain. Contextual Influences in Visual P i TAI SING LEE Computer Science Department and Center for Neural Basis of Cognition, Carnegie Mellon University, Pittsburgh, PA, USA Interestingly, the enhancement is uniform across the surface of a compact region, with a sudden drop off at the region’s border. Hence, it has been proposed to be a signal that could highlight a figure against its background and is called the ▶figure enhancement effect [4]. According to most studies, the onset delay of this figure enhancement effect is proportional to the size of the region. When the receptive field is at the center of a region that is six times larger than its size, the onset delay is typically 40 ms relative to stimulus is different from that of the center receptive field stimulus than when the orientations are the same. When the receptive field stimulus is a bar, iso-orientation suppression emerges at about 10 ms after the onset of the response to the receptive field stimulus [3]. When the receptive field stimulus is a part of an oriented texture region significantly larger than that of the receptive field, the later part of the neuron’s response is inversely proportional to the size of the region – the larger the region, the smaller the response. This results in a relative enhancement of response when the neuron’s receptive field is inside a smaller region than when it is in the larger background region. Interestingly, the enhancement is uniform across the surface of a compact region, with a sudden drop off at the region’s border. Hence, it has been proposed to be a signal that could highlight a figure against its background and is called the ▶figure enhancement effect [4]. According to most studies, the onset delay of this figure enhancement effect is proportional to the size of the region. When the receptive field is at the center of a region that is six times larger than its size, the onset delay is typically 40 ms relative to Perceptual computations supported by the complex machinery in V1 likely go beyond feature detection and feature contrast enhancement. From a computational perspective, contextual effects reflect the influence of computational constraints, realized by neuronal con- nectivity and interaction, necessary for solving visual inference problems. Surround interaction can bring in contextual information to improve local estimates of visual cues, as evident in the observations that ▶orientation tuning curves and ▶disparity tuning curves tend to sharpen over time during the analysis of each visual image. Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:24 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: Contextual Influences in Visual Processing. Figure 1 Stimuli used in contextual modulation studies. (a) Classic center-surround stimuli that have been typically used in neurophysiological studies on iso-orientation surround suppression [3]. Neurons tend to respond better when the orientations of the center and surround gratings are different (left image) than when they are the same (right image). The red ellipse outlines the spatial extent of the receptive field of the neuron. A similar effect observed in a larger center patch with a significantly longer delay is called figure enhancement [4]. (b) Surround context can change the perceptual saliency of the receptive field stimulus. The receptive field stimulus is said to pop out from the background on the left image, but not on the right image. This pop-out phenomenon depends on 3D interpretation of the stimulus elements. Early visual neurons’ activity is correlated with the perceptual saliency of this pop-out phenomenon [5]. Surround interaction can be quite complex and can vary according to the luminance contrast or the spatial scale of the stimuli. While surround modulation tends to be suppressive when the luminance contrast of the stimulus is strong, it can become facilitatory when the luminance contrast is weak. Neuronal ▶adaptation, well known in the ▶retina and LGN, is sensitive to the absolute luminance and luminance contrast levels in the entire scene. In a dark and low-contrast environ- ment, retinal and LGN neurons are known to expand their receptive fields temporally and spatially with a simultaneous increase in their sensitivity gains. Such a strategy serves to optimize feature detection in the presence of noise. The contrast dependence in surround influence likely results from V1 neurons inheriting and extending these adaptation or optimization strategies. stimulus is different from that of the center receptive field stimulus than when the orientations are the same. When the receptive field stimulus is a bar, iso-orientation suppression emerges at about 10 ms after the onset of the response to the receptive field stimulus [3]. When the receptive field stimulus is a part of an oriented texture region significantly larger than that of the receptive field, the later part of the neuron’s response is inversely proportional to the size of the region – the larger the region, the smaller the response. This results in a relative enhancement of response when the neuron’s receptive field is inside a smaller region than when it is in the larger background region. Characteristics The study of contextual influences in visual processing has a long history in psychology and neuroscience [1]. Investigations of these effects in the visual system have focused on the ▶modulatory effect on the activity of a neuron by image structures outside its localized ▶receptive field. The classical approach employs the simplest stimuli such as bars and sinusoidal gratings to probe the interaction between the stimuli presented inside and outside a neuron’s classical receptive field. A prevalent finding is that neurons in both the ▶primary visual cortex (striate cortex, V1) and the ▶extrastriate cortex exhibit ▶feature contrast enhancement, i.e., the cells respond better when the stimulus attributes in the area surrounding their receptive fields, such as bar orientation, are different from those inside their receptive fields (Fig. 1a). One well-known stimulus-specific surround suppres- sion, observed with an onset delay, is called ▶iso- orientation suppression. In this phenomenon, a neuron’s response is stronger when the orientation of the surround Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:24 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: 2 Contextual Influences in Visual Processing response onset on the average. The figure enhancement effect is more general than iso-orientation suppression as it has been observed in studies with motion or shape from shading stimuli without any orientation contrast between the receptive field stimulus and the surround [4,5]. p [ ] Functionally, both iso-orientation suppression and figure enhancement can serve to enhance stimulus feature contrast, resulting in an increase in ▶perceptual saliency of the representation of less expected or surprising visual events to facilitate further processing. Indeed, it has been demonstrated that this response enhancement is directly proportional to perceptual saliency of the visual pattern, as measured in terms of the reaction time for target detection, and it is disso- ciable from luminance contrast or orientation contrast in the stimulus (Fig. 1b) [5]. The broader spatial extent and the longer onset latency of the figure enhancement effect suggest that, while iso-orientation suppression might be mediated primarily by inhibitory ▶local circuits, the figure enhancement or perceptual saliency effect likely involves additional long range facilitation circuits including recurrent ▶feedback from the extra- striate cortex, as suggested by both anatomical and deactivation studies. The ▶retinotopic organization, the connection infrastructure, and the tuning properties of neurons in V1 make it ideally suitable for supporting a variety of visual computations. One such computation is the grouping of edges into contours and features into Contextual Influences in Visual Processing 3 ▶disparity selective neurons when the 3D depth plane of the random dot stereogram stimulus intersects with the cells’ optimal disparity tunings. This process appears to contribute to the gradual sharpening of the neurons’ disparity tunings over time, providing a plausible mechanism for improving local estimates of visual cues based on global context. Such cooperative or mutual facilitatory mechanisms might also contribute to surface association by increasing the firing rates of the neurons analyzing different parts of the same visual surface simultaneously. The resulting enhanced and correlated activities, partly represented in the figure enhancement effect, can highlight the relevant coincident features in visual input as a group to provide a stronger drive for downstream neurons in the extrastriate cortex to learn explicit representations for higher order features and structures. coherent regions. There is some evidence that V1 plays an important role in this computation to be discussed below. First, the activity of some V1 neurons is enhanced if the surrounding bars outside their receptive fields line up with the bar presented within their receptive fields to form a longer contour (Fig. 2a). g ( g ) Moreover, some V1 neurons respond to the ▶sub- jective contour of a ▶Kanizsa figure, even when no feature is presented to their classical receptive fields (Fig. 2b). There is also evidence that neurons can interpolate contours across the blind spot or behind an occlusion. Furthermore, collinear contours have been found to induce neuronal synchrony in V1 neurons of the same ▶orientation selectivity. Recently, it was also found that neurons with different orientation tunings, when stimulated simultaneously by curved contours, also exhibit an increase in synchrony or ▶effective connectivity, as revealed by multi-electrode recordings [6]. This dynamic change in effective connectivity between neurons as a function of stimulus is suggestive of a mechanism for ▶contour completion. Contextual Influences in the Extrastriate Visual Cortex The extrastriate cortex, downstream from the striate or primary visual cortex, is partitioned into many different visual areas. The feature contrast enhancement effect observed in V1 is also prevalent in extrastriate visual areas, expressed in the respective feature dimensions that neurons in those areas are tuned to. Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:24 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: In area ▶MT (medial temporal), for example, the motion of surround stimuli has been shown to significantly modulate the response of a neuron to moving stimuli presented to its receptive field. The response of the neuron is suppressed when the direction of surround motion is the same as the motion detected in the neuron’s receptive field. This is analogous to the iso-orientation suppression in V1 but in the motion domain. In addition, the disparity-tuned MT neurons also experience iso-disparity suppression. In addition, similar changes in effective connectivity have also been observed among spatially disjoint Contextual Influences in Visual Processing. Figure 2 Neurophysiological evidence of contour completion in V1. (a) Oriented bars in the surround (left image), when aligned with the receptive field stimulus to form a contour, can increase a cell’s response to its receptive field stimulus (right image) (Kapadia, Westheimer and Gilbert 2000). The red ellipse outlines the spatial extent of the receptive field of the neuron. (b) The subjective contour of a Kanizsa’s illusory square can evoke response in a V1 neuron even when no stimulus feature is present in its receptive field (red ellipse) (Lee and Nguyen 2001). The subtle addition of thin circles on the right image changes the perceptual interpretation of the image from a white square occluding four black circular disks, with a vivid subjective contour over the receptive field (left image), to that of a white square in a background visible through four circular windows on a white wall in front (right image). The extrastriate cortical areas, however, exhibit some additional contextual effects that are rarely observed in the striate cortex. Many of these new contextual effects are concerned with the inference of 3D surfaces, their occlusion and depth ordering relationships, also known as ▶figure-ground organization. In MT, it has been shown that the responses of direction-selective neurons to a motion stimulus are sensitive to the figure-ground context defined by the surrounding surface depth structures in a way that is consistent with ▶Barber Pole illusion [7]. Contextual Influences in Visual Processing. Contextual Influences in Visual Processing. Figure 2 Neurophysiological evidence of contour completion in V1. (a) Oriented bars in the surround (left image), when aligned with the receptive field stimulus to form a contour, can increase a cell’s response to its receptive field stimulus (right image) (Kapadia, Westheimer and Gilbert 2000). (b) In the Cornsweet-O’Brien illusion, the presence of a contrast edge can change the perception of the brightness of a region. A V2 neuron that prefers darkness over brightness would respond better to the perceptually darker region (left image) than to the perceptually brighter region (right image) even though the physical luminance of the receptive field stimulus in the two cases is exactly the same [10]. Figure 3 Neurophysiological evidence of surface inference in V2. (a) A left-border cell will respond more strongly when its receptive field (red ellipse) is analyzing the left border of a figure (left image) than when it is analyzing the right border of the figure (right image), even when the visual pattern on the receptive field and in its immediate surround is identical [8]. This class of cells, observed primarily in V2, is said to convey information about border-ownership or surface occlusion. (b) In the Cornsweet-O’Brien illusion, the presence of a contrast edge can change the perception of the brightness of a region. A V2 neuron that prefers darkness over brightness would respond better to the perceptually darker region (left image) than to the perceptually brighter region (right image) even though the physical luminance of the receptive field stimulus in the two cases is exactly the same [10]. The perception of surface attributes such as bright- ness, shading and color depends very strongly on the interpretation of the underlying 3D surface geometry and the illumination direction in the visual scene. Two observations suggest that these surface attributes might also be inferred and represented in V2 because of the dependence of such inference on 3D surface interpretation. First, the neural correlate of ▶shape- from-shading pop-out, a perceptual phenomenon that crucially depends on 3D surface interpretation, is observed in V2 but not in V1 pre-attentively [4]. Second, the neural correlate of the ▶Cornsweet- O’Brien illusion, an illusion in perceived brightness induced by edge contrast, which ultimately can be traced back to surface geometry and lighting direction interpretations in natural scenes, is observed in V2 but not V1 [10] (Fig. 3b). There has been, however, some evidence for brightness representation in V1 [1]. It is possible that the construction of brightness represen- tation is a gradual and distributed process, computed first at V1 based on surround luminance contrast, but achieving a more abstract and invariant representation in V2 as the 3D surface representation is made explicit. In general, neuronal activities tend to become progres- sively more abstract and more correlated with our subjective perceptual experience as one moves up the visual hierarchy. such as the task of tracing a curve. Beyond V2, extrastriate neurons tend to have large receptive fields. Attentional modulation in neurons of these higher areas typically manifests as the selection of one relevant feature over the others present within their individual recep- tive fields. Attention can be voluntary, as in selecting a particular spatial location (spatial attention) or a partic- ular feature (feature attention) in the receptive field for further analysis. But it can also be reflexive, driven or captured by the saliency of the stimuli computed automatically in early visual areas. The variety of ▶feature contrast and perceptual saliency effects observed in V1 and in the extrastriate cortex likely serves as a part of this reflexive attention mechanism. Recently, higher-order non-spatial contextual effects, such as context familiarity and associative memory, have also been shown to modify the activities of neurons in ▶inferotemporal cortex (IT) and medial temporal (MT) respectively. such as the task of tracing a curve. Beyond V2, extrastriate neurons tend to have large receptive fields. Attentional modulation in neurons of these higher areas typically manifests as the selection of one relevant feature over the others present within their individual recep- tive fields. Attention can be voluntary, as in selecting a particular spatial location (spatial attention) or a partic- ular feature (feature attention) in the receptive field for further analysis. But it can also be reflexive, driven or captured by the saliency of the stimuli computed automatically in early visual areas. The variety of ▶feature contrast and perceptual saliency effects observed in V1 and in the extrastriate cortex likely serves as a part of this reflexive attention mechanism. Recently, higher-order non-spatial contextual effects, such as context familiarity and associative memory, have also been shown to modify the activities of neurons in ▶inferotemporal cortex (IT) and medial temporal (MT) respectively. In addition to global image structures, behavior, task demands and memory are also known to provide strong contextual information to influence visual perception and object recognition. ▶Attentional modulation of neuronalresponses has beenwidely observedand studied in the extrastriate cortex (see ▶Visual Attention). The red ellipse outlines the spatial extent of the receptive field of the neuron. (b) The subjective contour of a Kanizsa’s illusory square can evoke response in a V1 neuron even when no stimulus feature is present in its receptive field (red ellipse) (Lee and Nguyen 2001). The subtle addition of thin circles on the right image changes the perceptual interpretation of the image from a white square occluding four black circular disks, with a vivid subjective contour over the receptive field (left image), to that of a white square in a background visible through four circular windows on a white wall in front (right image). Contextual Influences in Visual Processing. Figure 2 Neurophysiological evidence of contour completion in V1. (a) Oriented bars in the surround (left image), when aligned with the receptive field stimulus to form a contour, can increase a cell’s response to its receptive field stimulus (right image) (Kapadia, Westheimer and Gilbert 2000). The red ellipse outlines the spatial extent of the receptive field of the neuron. (b) The subjective contour of a Kanizsa’s illusory square can evoke response in a V1 neuron even when no stimulus feature is present in its receptive field (red ellipse) (Lee and Nguyen 2001). The subtle addition of thin circles on the right image changes the perceptual interpretation of the image from a white square occluding four black circular disks, with a vivid subjective contour over the receptive field (left image), to that of a white square in a background visible through four circular windows on a white wall in front (right image). Several lines of evidence suggest that the computa- tions underlying figure-ground segregation and 3D surface inference might start in visual area V2. First, a significant fraction of V2 neurons (and a small number of V1 neurons) have been shown to signal whether their receptive fields are at the left border or the right border of a figure in an image regardless of the polarity of contrast at the border (Fig. 3a). A left-border-preferring neuron carries the informa- tion that the border within its receptive field belongs to (or is owned by) the surface or region to its right [8]. 4 Contextual Influences in Visual Processing Contextual Influences in Visual Processing. Figure 3 Neurophysiological evidence of surface inference in V2. Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:24 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: (a) A left-border cell will respond more strongly when its receptive field (red ellipse) is analyzing the left border of a figure (left image) than when it is analyzing the right border of the figure (right image), even when the visual pattern on the receptive field and in its immediate surround is identical [8]. This class of cells, observed primarily in V2, is said to convey information about border-ownership or surface occlusion. (b) In the Cornsweet-O’Brien illusion, the presence of a contrast edge can change the perception of the brightness of a region. A V2 neuron that prefers darkness over brightness would respond better to the perceptually darker region (left image) than to the perceptually brighter region (right image) even though the physical luminance of the receptive field stimulus in the two cases is exactly the same [10]. Contextual Influences in Visual Processing. A complementary, right-border-preferring neuron exists at the same location, and both neurons could form a push–pull pair for every border orientation. The activity of a set of such pairs of ▶border-ownership neurons in various orientations along the border of each region in an image can encode the depth-order relation- ship between the different image regions or inferred surfaces. Secondly, it has been found that neurons in V2, but not in V1, are sensitive to the mismatch in features between the images from each eye at visual locations where one surface occludes another [9]. The emergence of sensitivity to this surface occlusion cue in V2, known as the ▶Da Vinci stereo, further suggests that 3D surfaces and their occlusions are explicitly represented in V2. The figure-ground context made explicit in V2 could feed back to constrain the computation in V1, resulting in, for example, the figure enhancement effect. However, it should be noted that the figure enhancement effect in V1 has not been conclusively demonstrated to depend solely on figure-ground organization. Contextual Influences in Visual Processing. g Figure 3 Neurophysiological evidence of surface inference in V2. (a) A left-border cell will respond more strongly when its receptive field (red ellipse) is analyzing the left border of a figure (left image) than when it is analyzing the right border of the figure (right image), even when the visual pattern on the receptive field and in its immediate surround is identical [8]. This class of cells, observed primarily in V2, is said to convey information about border-ownership or surface occlusion. Attentional effects in V1 are subtle and observable mostly when visual scenes are cluttered or in tasks that demand considerable spatial attention at precise locations From the perspective that vision is a process for inferring the various underlying environmental causes of visual patterns such as the 3D geometry of surfaces, the identities of objects and the illumination direction in the scene, the extrastriate areas in the visual hierarchical system might be conceptualized as modules that provide Contextual Influences in Visual Processing explicit representation of these decomposable causes. Each extrastriate module furnishes an explanation on some aspect of the visual scene. The inference of the underlying causes involves integration of information across space and over time by neurons in the higher-order visual areas, which in turn provide a variety of context in which visual processing in the earlier visual areas can be refined. V1, with its neurons arranged in a spatially precise ▶retinotopic map and endowed with small localized receptive fields capable of representing fine details in images, might serve as a high resolution buffer at which all the causes are combined together to synthesize an explanation of the visual input represented explicitly there. These interactive computations can bring about a very rich variety of contextual influences in V1 and the extrastriate cortex. The long latency of many of the contextual effects observed suggests that a substantial amount of recurrent interaction could have taken place. Computations involving such recurrent interaction will predict the simultaneous emergence of the perception-related signals in many visual and decision areas in the brain. 2. Angelucci A, Levitt JB, Walton EJ, Hupe JM, Bullier J, Lund JS (2002) Circuits for local and global signal integration in primary visual cortex. J Neurosci 22 (19):8633–8646 3. Knierim JJ, van Essen DC (1992) Neuronal responses to static texture patterns in area V1 of the alert macaque monkey. J Neurophysiol 67:961–980 4. Lamme, VAF (1995) The neurophysiology of figure- ground segregation in primary visual cortex. J Neurosci 15:1605–1615 5. Lee TS, Yang CF, Romero RD, Mumford D (2002) Neural activity in early visual cortex reflects behavioral experience and higher order perceptual saliency. Nat Neurosci 5:589–597 6. Samonds JM, Zhou Z, Bernard MR, Bonds A (2006) Synchronous activity in cat visual cortex encodes collinear and cocircular contours. J Neurophysiol 95 (4):2602–2616 7. Duncan RO, Albright TD, Stoner GR (2000) Occlusion and the interpretation of visual motion: perceptual and neuronal effects of context. Comp. by: ASaid Maraikayar Date:1/4/08 Time:15:28:25 Stage:First Proof File Path://spiina1001z/ womat/production/PRODENV/0000000005/0000006643/0000000016/0000767254.3D Proof by: QC by: J Neurosci 20:5885–5897 8. Zhou H, Friedman HS, von der Heydt R (2000) Coding of border ownership in monkey visual cortex. J Neurosci 20:6594–6611 9. Bakin JS, Nakayama K, Gilbert CD (2000). Visual responses in monkey areas V1 and V2 to three- dimensional surface configurations. J Neurosci 20:8188–8198 1. Albright TD, Stoner GR (2002) Contextual influences on visual processing. Annu Rev Neurosci 25:339–379 References 10. Hung CP, Ramsden BM, Roe AW (2007) A functional circuitry for edge-induced brightness perception. Nat Neurosci 10(9):1185–1190
https://openalex.org/W2969150076	https://jlse.springeropen.com/track/pdf/10.1186/s42825-019-0009-5	English	null	Regenerating leather waste for flexible pressure sensing applications	Journal of leather science and engineering/Journal of Leather Science and Engineering	2,019	cc-by	6,684	* Correspondence: iamsli@njtech.edu.cn; iambzheng@njtech.edu.cn; iamfwhuo@njtech.edu.cn Jie Lei and Binghua Zou are equally contributed to this work. Key Laboratory of Flexible Electronics (KLOFE) & Institute of Advanced Materials (IAM), Nanjing Tech University (NanjingTech), 30 South Puzhu Road, Nanjing 211816, People’s Republic of China Abstract Pressure sensor can be applied in a wide range of fields, such as voice recognition, human motions detection and artificial electronic skin, the sensing of which is greatly influenced by the flexibility and stretchability of substrate materials. Here, based on the piezoresistive effect, new kinds of flexible pressure sensors have been realized from a pair of flexible and biocompatible collagen films: one is coated by silver nanowires (Ag NWs) and the other by interdigital electrode, respectively. The collagen films are regenerated from leather waste and could bring economic benefits to society. The prepared pressure sensors are applied for voice recognition and human motion detection. Keywords: Leather waste, Collagen films, Pressure sensor, Flexible, Human motion detection © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Regenerating leather waste for flexible pressure sensing applications Jie Lei†, Binghua Zou†, Runan Zhang, Kang Zhang, Ruijie Xie, Weina Zhang, Jiansheng Wu, Sheng Li* , Bing Zheng* and Fengwei Huo* Binghua Zou†, Runan Zhang, Kang Zhang, Ruijie Xie, Weina Zhang, Jiansheng Wu, Sheng Li* , eng* and Fengwei Huo* Journal of Leather Science and Engineering Journal of Leather Science and Engineering (2019) 1:7 Lei et al. Journal of Leather Science and Engineering https://doi.org/10.1186/s42825-019-0009-5 1 Introduction Journal of Leather Science and Engineering Page 2 of 9 Page 2 of 9 the fabrication of their devices are still in high-cost and undesireable. which comprise tropocollagen molecule clusters with a diameter of about 1.5 nm and a length of ≈300 nm [33]. The hierarchical structures consisting of triple hel- ical structure endow the collagen with robust strength and high stability. ILs have been widely used for dissol- ution and regeneration of biomaterials, such as cellulose and protein, and the experimental process has also been widely explored [34–38]. In addition, ILs have been con- sidered as desirable green solvents to dissolve collagen fi- bers due to the excellent properties such as negligible vapor pressures, miscibility with water and organic sol- vents, and recyclability [39–41]. Here, the flexible collagen films can be obtained by the dissolution of collagen fibers in IL ([BMIM] Cl, 1-butyl-3-methylimidazolium chloride) and regenerated in precipitant (DI water) (Additional file 1: Figure S2). The dissolution mechanism of the col- lagen fibers in [BMIM] Cl can be found in Fig. 1a. Accord- ing to the electron donor-electron acceptor (EDA) theory, the [BMIM] Cl is dissociated at high temperature and the free Cl−acting as electron donor would associate with the hydrogen in the amino group, and the free [BMIM]+ act- ing as electron acceptor would complex with the oxygen in the carboxyl group, the process of which would disrupt hydrogen bond between the collagen macromolecule chains and lead to the dissolution of the whole collagen [33]. In addition, [BMIM] Cl can be completely miscible with the deionized water while collagen cannot, so during the regeneration process, the hydrogen bond can be re- built between collagen chains as the [BMIM] Cl has been washed away. However, the quantity and position of some hydrogen bond differ from that of the original collagen due to the changing position of the collagen chains during the regeneration process, resulting in the change of struc- ture and properties of regenerated collagen films. Considering that approximately 200 kg of leather would be manufactured from 1 ton of wet-salted skin/ hide, resulting in the generation of more than 600 kg lea- ther waste [28]. Collagen is an important component in leather waste. The usage of the above collagen can not only reduce the leather waste for the environment, but also promote the development of functional materials. 1 Introduction In recent years, there have been many researches on the applications of collagen fibers from leather, such as using the collagen fibers to fabricate microwave absorption materials, anode materials, conductive leather and mag- netic composites [29–32]. However, research on using collagen films as substrates from leather waste for flex- ible electronics has not been reported. p Herein, we present the use of regenerated collagen films from leather waste as the substrate to fabricate a flexible and transparent pressure sensor, which is piezo- resistive-type. It consists of two collagen films coating with Ag NWs as active layers and painted with Ag paste as the interdigital electrodes, respectively (Scheme 1). The sensor is capable of converting mechanical pressure into electrical signal effectively, and the pressure could be driven by different motion ranges of human body, such as voice recognition, finger and wrist bending-re- leasing motions. These properties and characteritics rely heavily on the flexible collagen film. Due to the collagen film is biocompatible and comfortable to human skin, there are only small mechanical property mismatches between the prepared device and human body, making it quite promising to the applications for electronic skins. 1 Introduction macromolecular materials can both serve as flexible sub- strates [10–27]. Comparing the two kinds of materials, the latter one has prominent advantages due to the bio- compatibility, biodegradability and sustainability, thus serving as desirable substrate for the next-generation flexible electronic devices [13]. Driven by the development of intelligent manufacturing, intelligent robotics, human-machine interaction and bio- medical diagnosis have received extensive attention [1–8], and they have high requirments for flexible bending per- formances. However, many traditional electronic devices are based on rigid semiconductor silicon materials. Thus the bending and elongation characteristics of the devices are significantly limited. The development of flexible sens- ing techonology could solve this problem effectively, such as new types of flexible pressure sensors, which have shown part of characteristics of human skin [4–10]. More- over, a flexible sensor with the property of external forces perception can be used for bionic electronic skin and vari- ous wearable electronic devices [9, 10]. Normally, many types industrial fabricated polymers, such as polydimethylsiloxane (PDMS) [10, 14, 15], poly (ethylene terephthalate) (PET) [16, 17], poly (ethylene naphthalate) (PEN) [11] and poly (imide) (PI) [15, 18] have been used as substrates for flexible electronic de- vices. Bao’s group fabricated a flexible and sensitive or- ganic thin film transistor based on PDMS, which can be used for pulse monitoring in cardiovascular surveillance [10]. Though these polymers have prominent advantages in flexibility, stretchability and even excellent fit between the device and tissue, their manufacturing process and fabricated devices are not eco-friendly in the long term. However, natural macromolecular materials, such as col- lagen [19], silk fibroin [20–22], cellulose [23–26], chitin [27] and so on, will overcome those difficulties and be the ideal choices in the development of electronic de- vices due to their flexibility, biodegradability and bio- compatibility. Roger’s group integrated thin-film silicon with transient circuits based on silk films [20]. Their de- vice can be used for versatile applications due to the tunable degradation time of silk-fibroin films. However, There are some key components which determine the basic performances of the electronics, such as active materilas, substrate materials and electrodes. Among these considerations, the substrate materials play an im- portant role [5, 7–12]. The substrate materials with ideal flexibility and robust mechanical strength can provide a sensitive response to the external pressure. Generally, in- dustrial fabricated polymers and some natural Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Lei et al. 2.1 Preparation and Properities of the flexible collagen film The collagen fibers possess an intensive band at 1659 cm−1 corresponding to Amide I in helical form. However, the peak of Amide I (1650 cm−1) in collagen film shifts to higher frequency, indicating that there are some changes with the triple helical structure of the collagen film. The surface appearance, transparency and crystal structure of collagen films are analyzed by scanning electron microscope (SEM), Ultraviolet–visible (UV-vis) spectra and X-ray diffraction (XRD). The top and cross- sectional-view SEM images of the collagen films reveal that the collagen films are compact (Fig. 2a, b). Figure 2c shows a magnified view of the cross-sectional image of the film, indicating the fabricated film is dense. The thickness of the collagen films can be controlled by changing the initial amount of the mixture solution dis- tributed on the glass sheet. And the transparency of the collagen film is higher than 55% in the visible spectrum (400–700 nm) by the UV-vis transmittance spectra (Fig. 2d), which is an attractive advantage for electronic skin, such as artificial cornea [45, 46]. The crystallization of the collagen films is investigated by XRD analysis (Fig. 2e). Three main diffraction peaks can be observed, and the dif- fraction peaks in 2θ = 7.8° produced in crystalline regions represent the distance between the molecular chains in collagen fibers [32]. Compared with that of collagen fibers, the diffraction peak of collagen film shifts to the left side, indicating the space between collagen molecular chains becomes lager and the chains are dispersed. It is notable that there exists a wide diffraction peak in 21.8° generated in amorphous regions in the profile of the collagen fibers, indicating a diffuse scattering due to the multilayer To investigate the thermal stability of the collagen fi- bers and regenerated collagen film, we obtained the Thermo Gravimetric Analyzer (TGA) profiles of the collagen fibers and film (10 °C min−1 ramp, N2 atmos- phere). Normally, The thermal stability of a large mol- ecule depends on its molecular weight, spatial structure and intermolecular forces. Here, there are two main stages of the weight loss of the collagen fibers and films: one is at 40~150 °C, representing the loss of absorbed and bound water of collagen fibers and films; the other is at 200~500 °C, representing the thermal decompos- ition of polypeptide chains in collagen [44]. 2.1 Preparation and Properities of the flexible collagen film The collagen fibers and regenerated collagen films are firstly studied by using Fourier-transform infrared (FT-IR) analysis (Fig. 1b). The peak of Amide A (3308 cm−1) rep- resents N-H stretching vibration absorption, which relates to the hydrogen bonds between the collagen chains. Amide I (1659 cm−1), Amide II (1550 cm−1) and the Amide III (1450 cm−1, 1241 cm−1) represent C=O The film is prepared by dissolving leather waste in ionic liq- uids (ILs). The pickled skin in this experiment is the source of collagen fibers, consisting of multiple collagen fibrils with a diameter of 100–500 nm (Additional file 1: Figure S1). The collagen fibrils are made up by several microfibrils Scheme 1 Schematic showing the source of collagen film and its application as the substrate for pressure sensors Scheme 1 Schematic showing the source of collagen film and its application as the substrate for pressure sensors Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Page 3 of 9 Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Lei et al. Journal of Leather Science and Engineering Page 3 of 9 Fig. 1 Properties of collagen fibers. a Dissolution and regeneration mechanism of collagen fibers. b FT-IR spectra of collagen fibers and collagen films. c TGA profiles of collagen fibers and collagen films (10 °C min−1 ramp, N2 atmosphere) Fig. 1 Properties of collagen fibers. a Dissolution and regeneration mechanism of collagen fibers. b FT-IR spectra of collagen fibers and collagen films. c TGA profiles of collagen fibers and collagen films (10 °C min−1 ramp, N2 atmosphere) breaking the forces of the hydrogen bonds and ionic bonds between the collagen molecules, and in the regen- eration process, the location and quantity of the hydrogen bonds have changed and even some hydrogen bonds have been disrupted due to the high dissolution temperature. stretching vibration adsorbtion, N-H bending vibration adsorbtion, C-N stretching vibration absorption and N-H bending vibration adsorbtion, respectively [42, 43]. The profile of the collagen film is similar to the collagen fibers, and this means they have similar chemical functions (Fig. 1b). However, there are also some differences which are clearly identified from the spectra. The peak of Amide A (3393 cm−1) in collagen film shifts to lower frequency and narrows down, indicating some hydrogen bonds have ruptured between the collagen chains, which is also con- sistent with the EDA theory. 2.1 Preparation and Properities of the flexible collagen film The decom- position temperature of the films (254 °C) is lower than that of the fibers (276 °C) (as shown in Fig. 1c). That is because the ILs dissolves collagen fibers mainly by Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Page 4 of 9 Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Page 4 of 9 Lei et al. Journal of Leather Science and Engineering Fig. 2 Properties of the collagen film. a top-view and b cross-sectional view of the collagen film. c A magnified view of (b). d UV-vis transmittance spectra of collagen film. e XRD patterns of collagen fiber and collagen film Fig. 2 Properties of the collagen film. a top-view and b cross-sectional view of the collagen film. c A magnified view of (b). d UV-vis transmittance spectra of collagen film. e XRD patterns of collagen fiber and collagen film structure in the collagen fibers. However, for the collagen film, the peak in 21.8° narrows down and the crystallinity increases. paste as the interdigital electrodes, respectively (Fig. 3c). Ag NWs have a concentration of 5 mg mL−1 and width of about 100 nm (Additional file 1: Figure S3). The Ag NWs are coated on the collagen films using a spin-coater at a rate of 2000 rpm, which results in a conductive film with a mean sheet resistance of 37 Ω sq.−1. The sensing principle of this pressure sensor is transducing the resistance change resulted from applied pressure into an electrical signal. When an applied pressure is put on the sensor, the amount of Ag NWs between interdigital electrode arrays was changed, leading to the change of elec- trical signal. Higher pressure would cause more Ag 2.2 Preparation and performances of the flexible pressure sensor Figure 4c shows the instant response of the sensor exhibits a response time of 349 ms and relaxation time of 147 ms under 636 Pa, indicating the current changes rap- idly respongding to the applied pressure, which can be expected to monitor human motions synchronously. To investigate the stability of the sensor, current changes ver- sus time are obtained for the repeated loading/unloading under 3182 Pa for 900 cycles (Fig. 4d). The magnified view in the bottom of Fig. 4d shows the middle and last 12 y g Based on the above working principle of the sensor, the sensitivity, transient response analysis and stability performances of the device are investigated. Typical current-voltage (I-V) curves of the sensor responding to different applied pressures are displayed in Fig. 4a. The resistance drops immediately along with the increasing pressure. As the applied pressure increasing, the amount of Ag NWs between interdigital electrode arrays increased, thereby enhancing electron transport and reducing the overall resistance of the sensor. The pressure signals can be read out when the operating voltage is 0.5 V, which is fa- vorable for portable diagnosis. And the applied pressures on the sensor are achieved by the equation: P = mg/A, where P is the pressure, m is the mass of a counterweight, g is the gravity unit, A is the actual contact area between the counterweight and the effective part of the sensor. Here, a 1.1 × 1.4 cm2 sheet glass is put on the device to apply a uniform pressure, so A is 1.1 × 1.4 cm2 in fact. When pressure is gradually applied to the device, the sensi- tivity S can be defined as S = (ΔI/I0)/ΔP, where ΔI is the relative change in the current, I0 is the initial current when a 1.1 × 1.4 cm2 glass is put on the device, and ΔP is the Fig. 4 Pressure sensing performances of the device. a Typical current–voltage curves of the sensor responding to different applied pressure. b Current changes of the sensor response to various applied pressure showing the pressure sensitivity value in a range of 64–1909 Pa is 13.33 KPa−1 and 1.27 KPa−1 in a range of 2545–6364 Pa. c Instant response of the sensor towards applied pressure showing a response time of 349 ms and relaxation time of 147 ms under 636 Pa. 2.2 Preparation and performances of the flexible pressure sensor A flexible pressure sensor has been fabricated based on the regenerated collagen film, demonstrating the possi- bility of the film as substrate for the pressure sensing ap- plications (Fig. 3a, b). In this work, the flexible pressure sensor is fabricated by two layers of collagen films coated with Ag NWs as the active layer and painted with Ag Fig. 3 Fabrication and working principle of the device. a Large-scale uniform and transparent collagen films (scale bar = 1.5 cm). b Photograph of the flexible and transparent pressure sensor (scale bar = 1.0 cm). c Fabrication of a flexible pressure senor based on collagen films. d The illustration of the work principle of the pressure sensor Fig. 3 Fabrication and working principle of the device. a Large-scale uniform and transparent collagen films (scale bar = 1.5 cm). b Photograph of the flexible and transparent pressure sensor (scale bar = 1.0 cm). c Fabrication of a flexible pressure senor based on collagen films. d The illustration of the work principle of the pressure sensor Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Page 5 of 9 Lei et al. Journal of Leather Science and Engineering Page 5 of 9 NWs contacting with the electrodes, leading to more conductive ways (Fig. 3d). change in applied pressure. To demonstrate the sensor’s sensitivity to the quick change of different applied pres- sures, we obtained the real-time current changes of the sensor upon the pressure values from 64 Pa to 6364 Pa. The current change responding to different applied pres- sure exhibits detecting and sensing capability of the sensor, and the sensitivity value in a range of 64 Pa −1909 Pa of our sensor is 13.33 KPa−1 and 1.27 KPa−1 in a range of 2545 Pa - 6364 Pa (Fig. 4b). As the collagen film is swelling in DI water, the surface of the film is uneven during drying processes, which causes the first stage of the sensitivity dia- gram is high. When higher pressure is applied to the sen- sor, the sensitivity is reduced because the most important influence factor in this stage is the conductive ways re- sulted from the contact between Ag NWs and the elec- trode. 2.2 Preparation and performances of the flexible pressure sensor d Working stability of the sensor for 900 cycles under a pressure of 3182 Pa and the magnified view of the middle and last 12 cycles Fig. 4 Pressure sensing performances of the device. a Typical current–voltage curves of the sensor responding to different applied pressure. b Current changes of the sensor response to various applied pressure showing the pressure sensitivity value in a range of 64–1909 Pa is 13.33 KPa−1 and 1.27 KPa−1 in a range of 2545–6364 Pa. c Instant response of the sensor towards applied pressure showing a response time of 349 ms and relaxation time of 147 ms under 636 Pa. d Working stability of the sensor for 900 cycles under a pressure of 3182 Pa and the magnified view of the middle and last 12 cycles Page 6 of 9 Page 6 of 9 Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Page 6 of 9 cycles, which shows the stable current changes responding to applied pressure, indicating the sensor has high stability, repeatability and durability. to perform the releasing motions, the current decreased due to the releasing deformation of the sensor. Notably, the current responses corresponding to bending and re- leasing motions exhibit stability and reproducibility, indi- cating the excellent reliability of the sensor. Furthermore, when the sensor is attached to the wrist of the volunteer to detect some large range of motions, apparent and re- peatable current signals can also be recorded during the wrist bending-releasing motions (Fig. 5c). The current in- creased when the volunteer performed bending motions of the wrist, and then it decreased rapidly responding to the releasing motions due to the recovery of the sensor. These results demonstrate the capabilities of the sensor in detecting human motions, which might be beneficial to access the training data of athletes and monitor some physical activities of patients synchronously. 2.3 Human physical motions sensing y The flexible pressure sensor is used to monitor different ranges of human motions, such as finger and wrist bending-releasing motions. First, we have investigated the electrical current values of the sensor recorded upon different applied pressures (Fig. 5a), which demonstrates the electrical signals generated upon applied pressure are sharp and reversible, indicating the steady and synchron- ous measurement of the human motions. Furthermore, the real-time current changes of the sensor are obtained upon different pressures. Responding to the increasing ap- plied pressure, the current increases apparently whereas the resistance decreases, which demonstrates the relia- bility of the sensor for real-time monitoring (Additional file 1: Figure S4). In order to verify this, we investigate the monitoring performances of the sensor. The sensor is attached to the joint of an index finger of the volunteer with adhesive tapes (Fig. 5b). And the volunteer is con- ducted to perform finger bending-releasing motions. In detail, the current increased sharply responding to the bending motions and when the volunteer was conducted 2.4 Voice recognition sensing The striking differences between these current signals indicate the good performance of the pressure sensor to act as a voice recognition device. 4.2 Characterization of the collagen films g The morphologies of blank collagen films and Ag NWs/ collagen film were characterized by scanning electron microscopy (SEM) (JSM-7800F, Japan) with an accelerat- ing voltage of 3 kV. FT-IR spectra were recorded with a FTIR 460 plus (JASCO, Japan). The thermal properties of collagen films were detected using TGA (TG209-F3, Netzsch) with a ramp of 10 °C min−1 under N2 atmos- phere. X-ray diffraction (XRD) patterns were recorded by Advance diffractometer (AXS D8, Bruker) using Nickel-filtered Cu Kα radiation (λ = 1.5406 Å). UV-vis spectra were measured using Shimadzu UV-1750. 4.3 Fabrication of the flexible pressure sensor Interdigital electrodes of silver paints were painted on a flexible collagen film. Ag NWs were dissolved in ethanol with a concentration of 5 mg mL−1 and the solution was dropped on the other collagen film. It was then spin-coated at a rate of 2000 rpm for 30 s for three times to obtain a mean sheet resistance of 37 Ω sq.−1. Then the two films coated with Ag NWs and silver paints were pressed together with edges and adhered by scotch tape. 2.4 Voice recognition sensing Apart from detecting some human physical motions, the pressure sensor is able to recognize different voices of human because of its high sensitivity to applied pressure and the piezoresistive characteristics (Fig. 5d, Additional file 1: Figure S5). A fabricated pressure sensor has been attached to the neck of the volunteer to detect the muscle motions. When the volunteer pronounces Fig. 5 Human motion detection and voice recognition. a Instant current changes of the sensor with different applied pressure. b Current signals of the sensor that was fixed on an index finger of the volunteer at bending-releasing motions. The top insets in b show a zoomed-in view of the red-boxed area. c Current signals of the sensor that was fixed on the wrist of the volunteer at bending-releasing motions. The top insets show the photograph of the sensor. d Recorded current signals versus time when the volunteer pronounced “bee” and “watermelon”. The top inset in d shows a photograph of the sensor attached to the neck of the volunteer Fig. 5 Human motion detection and voice recognition. a Instant current changes of the sensor with different applied pressure. b Current signals of the sensor that was fixed on an index finger of the volunteer at bending-releasing motions. The top insets in b show a zoomed-in view of the red-boxed area. c Current signals of the sensor that was fixed on the wrist of the volunteer at bending-releasing motions. The top insets show the photograph of the sensor. d Recorded current signals versus time when the volunteer pronounced “bee” and “watermelon”. The top inset in d shows a photograph of the sensor attached to the neck of the volunteer Lei et al. Journal of Leather Science and Engineering (2019) 1:7 Lei et al. Journal of Leather Science and Engineering Page 7 of 9 different words, such as “bee”, “apple”, “tomato” and “watermelon”, different current signals of the sensor are collected. Every graph in Additional file 1: Figure S5 represents one word. The current signal patterns in- crease corresponding to the increasing syllables of the word. When the volunteer repeats the same word every few seconds, the current signals keep consistent, which indicates the stability of the sensor and easy recovery from the deformation caused by the pressure of muscle motions. Also, when the volunteer pronounces different words, the current signal changes because different words give rise to different muscle motions. 3 Conclusions In conclusion, we have demonstrated the use of regener- ated collagen films from leather waste as the substrate for flexible pressure sensors. The prepared pressure sen- sor has a high sensitivity of 13.33 KPa−1 in a range of 64–1909 Pa and 1.27 KPa−1 in a range of 2545–6364 Pa. And the sensor exhibits a response time of 349 ms and relaxation time of 147 ms under 636 Pa. Moreover, the sensor has good stability and repeatability from the mea- surment of current changes responding to applied pres- sure for repeated loading/unloading operation under 3182 Pa. In addition, the sensor can monitor different ranges of human motions including voice recognition, finger and wrist bending-releasing, which demonstrates its potential applications in speech recognition and synchronously monitoring some physical activities of patients. This work provide a new thought for the development and application of bio-waste materials, such as leather waste in pressure sensing technology, biomedical diagnosis, human-machine interaction and so on. 4.4 Performances of the flexible pressure sensor The current-voltage (C-V) curves were collected using an electrochemical workstation (CHI 660E) between 0 and 0.5 V at a scan rate of 0.01V s-1. The other current mea- surements, including instant current-time curves recorded, human motion detection and voice recognition, were char- acterized by Keithley 4200 semiconductor characterization system. A 1.1×1.4 cm2 slight glass was covered on the pres- sure sensor to ensure homogeneous application of external pressure when the sensitivity and stability curves of the sensor were collected. And the pressure in experiments was applied by loading standard weights on the sensor and the size of the effective area is 1.1×1.4 cm2. To measure the performance of the sensor while it was bent, the sensor was attached to the joint of an index finger of the volunteer with adhesive tapes. When the volunteer performed finger bending-releasing motions, the current responses corre- sponding to bending and releasing motions can be re- corded. And when the sensor was attached to the wrist of the volunteer to detect some large range of motions, ap- parent and repeatable current signals can also be recorded during the wrist bending. The voice recognition tests were also conducted. The sensor was attached to the neck of the volunteer to detect the muscle motions. When the volunteer pronounced different words, such as “bee”, “apple”, “tomato” and “watermelon”, current signals were collected. The applied voltage between the interdigital electrodes was 0.5 V, and the above current signals recorded were current-time (I-t) curves in the measurements. 4.1 Preparation of the collagen films [BMIM] Cl (3.0 g) was placed into a 50 mL dried round flask with magnetic stirrer. Pickled skin powders (180 mg) were added in potions of 3 wt% of [BMIM] Cl each time and then microcrystalline cellulose (30 mg) were added as a cross-linker, the process of which was controlled by an oil bath at 100 °C for 5 h. When the solution is like syrup, the mixture solution was well spreaded on a glass sheet and then immersed into de- ionized water which served as a precipitator. Then the film was washed several times to ensure the ionic liquid washed away completely. Then a regenerated collagen film was obtained and then dried in a vac- cum drier to a constant weight. Page 8 of 9 Page 8 of 9 Lei et al. Journal of Leather Science and Engineering (2019) 1:7 (2019) 1:7 Lei et al. Journal of Leather Science and Engineering Acknowledgements 14. Wagner S, Bauer S. Materials for stretchable electronics. MRS Bull. 2012;37:207–13. The authors acknowledge all volunteers who participated in the project. 15. Kuribara K, Wang H, Uchiyama N, Fukuda K, Yokota T, Zschieschang U, Jaye C, Fischer D, Klauk H, Yamamoto T, Takimiya K, Ikeda M, Kuwabara H, Sekitani T, Loo YL, Someya T. Organic transistors with high thermal stability for medical applications. Nat Commun. 2012;3:723. Competing interests The authors declare that they have no competing interests. 22. Zhu B, Wang H, Leow WR, Cai Y, Loh XJ, Han MY, Chen X. Silk fibroin for flexible electronic devices. Adv Mater. 2016;28:4250–65. The authors declare that they have no competing interests. flexible electronic devices. Adv Mater. 2016;28:4250–65. Received: 10 June 2019 Accepted: 23 July 2019 Received: 10 June 2019 Accepted: 23 July 2019 Received: 10 June 2019 Accepted: 23 July 2019 23. Siegel AC, Phillips ST, Dickey MD, Lu N, Suo Z, Whitesides GM. Foldable printed circuit boards on paper substrates. Adv Funct Mater. 2010;20:28–35. 24. Nassar JM, Cordero MD, Kutbee AT, Karimi MA, Sevilla GAT, Hussain AM, Shamim A, Hussain MM. Paper skin multisensory platform for simultaneous environmental monitoring. Adv Mater Technol. 2016;1:1600004. Funding The project was supported by the National Key R&D Program of China (Grant No. 2017YFA0207201), National Natural Science Foundation of China (51702155, 21574065, 21604038, 21504043, 21604040), National Science Foundation for Distinguished Young Scholars (21625401), the Jiangsu Provincial Founds for Natural Science Foundation (BK20170975, BK20160975, BK20160981) and the Natural Science Fund for Colleges and Universities in Jiangsu Province (17KJB480007). 18. Sekitani T, Zschieschang U, Klauk H, Someya T. Flexible organic transistors and circuits with extreme bending stability. Nat Mater. 2010;9:1015–22. 19. Moreno S, Baniasadi M, Mohammed S, Mejia I, Chen YN, Lopez MAQ, Kumar N, Dimitrijevich S, Jolandan M. Biocompatible collagen films as substrates for flexible implantable electronics. Adv Electron Mater. 2015;1:1500154. 20. Hwang SW, Tao H, Kim DH, Cheng H, Song JK, Rill E, Brenckle MA, Panilaitis B, Won SM, Kim YS, Song YM, Yu KJ, Ameen A, Li R, Su Y, Yang M, Kaplan DL, Zakin MR, Slepian MJ, Huang Y, Omenetto FG, Rogers JA. A physically transient form of silicon electronics. Science. 2012;337:1640–4. 5 Additional file mapping with ultra-small passive sensors for health monitoring and critical care. Nat Commun. 2014;5:5028. 5. Wang X, Gu Y, Xiong Z, Cui Z, Zhang T. Silk-molded flexible, ultrasensitive, and highly stable electronic skin for monitoring human physiological signals. Adv Mater. 2014;26:1336. Additional file 1: Figure S1. (a) Photograph and (b) SEM image of the collagen fibers. Figure S2. Schematic illustration of the fabrication process of the collagen film from collagen fibers. Firstly, pickled skin powders were added in potions of 3 wt% of [BMIM] Cl each time and then microcrystalline cellulose were added as a cross-linker. Then, the mixed solution was washed by DI water. Finally, a regenerated collagen film was obtained by spreading the solution on glass and dried in a vaccum drier to a constant weight. Figure S3. SEM image of Ag NWs coated on the collagen film. Figure S4. Instant current–time curves of the device at different resistance states responding to an applied pressure between 318 Pa and 6364 Pa. Figure S5. Recorded current signals versus time when the volunteer pronounced (a) “bee”, (b) “apple”, (c) “tomato”, and (d) “watermelon”. The insets show magnified current signals. (DOCX 1830 kb) 6. Pang C, Lee GY, Kim TI, Kim SM, Kim HN, Ahn SH, Suh KY. A flexible and highly sensitive strain-gauge sensor using reversible interlocking of nanofibres. Nat Mater. 2012;11:795–801. 7. Mannsfeld SC, Tee BC, Stoltenberg RM, Chen CV, Barman S, Muir BV, Sokolov AN, Reese C, Bao Z. Highly sensitive flexible pressure sensors with microstructured rubber dielectric layers. Nat Mater. 2010;9:859–64. 8. Someya T, Sekitani T, Iba S, Kato Y, Kawaguchi H, Sakurai T. A large-area, flexible pressure sensor matrix with organic field-effect transistors for artificial skin applications. Proc Natl Acad Sci U S A. 2004;101:9966–70. 9. Zang Y, Zhang F, Di CA, Zhu D. Advances of flexible pressure sensors toward artificial intelligence and health care applications. Mater Horiz. 2015;2:140–56. 10. Schwartz G, Tee BC, Mei J, Appleton AL, Kim DH, Wang H, Bao Z. Flexible polymer transistors with high pressure sensitivity for application in electronic skin and health monitoring. Nat Commun. 2013;4:1859. Abbreviations [ ] l b [BMIM]Cl: 1-butyl-3-methylimidazolium chloride; A: Area; Ag NWs: Silver nanowires; AXS: Advance diffractometer; C-V: Current-voltage; EDA: Electron donor-electron acceptor; FT-IR: Fourier-transform infrared; g: Gravity unit; I0: The initial current; ILs: Ionic liquids; I-t: Current-time; I-V: Current-voltage; m: Mass; P: Pressure; PDMS: Polydimethylsiloxane; PEN: Poly (ethylene naphthalate); PET: Poly (ethylene terephthalate); PI: Poly (imide); S: Sensitivity; SEM: Scanning electron microscope; TGA: Thermo gravimetric analyzer; UV-vis: Ultraviolet–visible; XRD: X-ray diffraction; ΔI: The relative change in the current; ΔP: The change in applied pressure 11. Kaltenbrunner M, Sekitani T, Reeder J, Yokota T, Kuribara K, Tokuhara T, Drack M, Schwodiauer R, Graz I, Gogonea S, Bauer S, Someya T. An ultra-lightweight design for imperceptible plastic electronics. Nature. 2013;499:458–63. 12. Zou B, Chen Y, Liu Y, Xie R, Du Q, Zhang T, Shen Y, Zheng B, Li S, Wu J, Zhang W, Huang W, Huang X, Huo F. Repurposed leather with sensing capabilities for multifunctional electronic skin. Adv Sci. 2018;1801283. 13. Sun H-S, Chiu Y-C, Chen W-C. Renewable polymeric materials for electronic applications. Polym J. 2016;49:61–73. Availability of data and materials All data generated or analysed during this study are included in this published article and its supplementary information files. All data generated or analysed during this study are included in this published article and its supplementary information files. 21. Kim DH, Kim YS, Amsden J, Panilaitis B, Kaplan DL, Omenetto FG, Zakin MR, Rogers JA. Silicon electronics on silk as a path to bioresorbable, implantable devices. Appl Phys Lett. 2009;95:133701. Authors’ contributions JL and BZ contributed equally to this work. JL and BZ designed experiments, performed, analyzed the results, and drafted the manuscript. RZ, KZ and RX were responsible for preparation of some experiment materials and some characterizations. JW and WZ helped to revise the manuscript. FH, SL and BZ supervised the project, helped design the experiments, and revised the manuscript. All authors read and approved the final manuscript. 16. Han X, Chen X, Tang X, Chen YL, Liu JH, Shen QD. Flexible polymer transducers for dynamic recognizing physiological signals. Adv Funct Mater. 2016;26:3640–8. 17. Zhao Z, Yan C, Liu Z, Fu X, Peng LM, Hu Y, Zheng Z. Machine-washable textile triboelectric nanogenerators for effective human respiratory monitoring through loom weaving of metallic yarns. Adv Mater. 2016;28:10267–74. References Hierarchical structure and nanomechanics of collagen microfibrils from the atomistic scale up. Nano Lett. 2011;11:757–66. 34. Meng Z, Zheng X, Tang K, Liu J, Ma Z, Zhao Q. Dissolution and regeneration of collagen fibers using ionic liquid. Int J Biol Macromol. 2012;51:440–8. 35. Sun N, Rodriguez H, Rahman M, Rogers RD. Where are ionic liquid strategies most suited in the pursuit of chemicals and energy from lignocellulosic biomass? Chem Commun. 2011;47:1405–21. 36. Brogan AP, Hallett JP. Solubilizing and stabilizing proteins in anhydrous ionic liquids through formation of protein-polymer surfactant Nanoconstructs. J Am Chem Soc. 2016;138:4494–501. 37. Vancov T, Alston A-S, Brown T, McIntosh S. Use of ionic liquids in converting lignocellulosic material to biofuels. Renew Energy. 2012;45:1–6. 38. Sherman VR, Yang W, Meyers MA. The materials science of collagen. J Mech Behav Biomed Mater. 2015;52:22–50. 39. Weingartner H. Understanding ionic liquids at the molecular level: facts, problems, and controversies. Angew Chem Int Ed. 2008;47:654–70. 40. Huddleston JG, Visser AE, Reichert WM, Willauer HD, Broker GA, Rogers RD. Characterization and comparison of hydrophilic and hydrophobic room temperature ionic liquids incorporating the imidazolium cation. Green Chem. 2001;3:156–64. 41. Zhao H, Xia S, Ma P. Use of ionic liquids as ‘green’ solvents for extractions. J Chem Technol Biotechnol. 2005;80:1089–96. 41. Zhao H, Xia S, Ma P. Use of ionic liquids as ‘green’ solvents for extractions. J Chem Technol Biotechnol. 2005;80:1089–96. 42. Renugopalakrishnan V, Bhatnagar RS. Hydrogen-bonded water in collagen structure - a Ft-Ir spectroscopic corroboration. Biophys J. 1984;45:A163. 42. Renugopalakrishnan V, Bhatnagar RS. Hydrogen-bonded water in collagen structure - a Ft-Ir spectroscopic corroboration. Biophys J. 1984;45:A163. 43. Vidal Bde C, Mello ML. Collagen type I amide I band infrared spectroscopy. Micron. 2011;42:283–9. 44. He L, Mu C, Shi J, Lin W. Modification of collagen with a natural cross-linker, procyanidin. Int J Biol Macromol. 2011;48:354–9. 45. Zhao X, Long K, Liu Y, Li W, Liu S, Wang L, Ren L. To prepare the collagen-based artificial cornea with improved mechanical and biological property by ultraviolet-a/riboflavin crosslinking. J Appl Polym. 2017;134:45226. 46. Li WC, Long Y, Liu Y, Long K, Liu S, Wang Z, Wang Y, Ren L. Fabrication and characterization of chitosan-collagen crosslinked membranes for corneal tissue engineering. J Biomat Sci-Polym E. 2014;25:1962–197. References 1. Someya T, Kato Y, Sekitani T, Iba S, Noguchi Y, Murase Y, Kawaguchi H, Sakurai T. Conformable, flexible, large-area networks of pressure and thermal sensors with organic transistor active matrixes. Proc Natl Acad Sci U S A. 2005;102:12321–5. 25. Gong S, Schwalb W, Wang Y, Chen Y, Tang Y, Si J, Shirinzadeh B, Cheng W. A wearable and highly sensitive pressure sensor with ultrathin gold nanowires. Nat Commun. 2014;5:3132. 26. Zhang JM, Zhang J. Advanced functional materials based on cellulose. Acta Polym Sin. 2010;12:1376–98. 2. Wang C, Hwang D, Yu Z, Takei K, Park J, Chen T, Ma B, Javey A. User-interactive electronic skin for instantaneous pressure visualization. Nat Mater. 2013;12:899. 2. Wang C, Hwang D, Yu Z, Takei K, Park J, Chen T, Ma B, Javey A. User-interactive electronic skin for instantaneous pressure visualization. Nat Mater. 2013;12:899. 27. Jin J, Lee D, Im HG, Han YC, Jeong EG, Rolandi M, Choi KC, Bae BS. Chitin nanofiber transparent paper for flexible green electronics. Adv Mater. 2016;28:5169–75. 3. Wu W, Wen X, Wang ZL. Taxel-addressable matrix of vertical-nanowire piezotronic transistors for active and adaptive tactile imaging. Science. 2013;340:952. 3. Wu W, Wen X, Wang ZL. Taxel-addressable matrix of vertical-nanowire piezotronic transistors for active and adaptive tactile imaging. Science. 2013;340:952. 4. Chen LY, Tee BC, Chortos AL, Schwartz G, Tse V, Lipomi DJ, Wong HS, McConnell MV, Bao Z. Continuous wireless pressure monitoring and 28. Veeger L. Ecological procedure to solve the tannery waste problems. J Am Leather Chem Assoc. 1993;88:326–9. 4. Chen LY, Tee BC, Chortos AL, Schwartz G, Tse V, Lipomi DJ, Wong HS, McConnell MV, Bao Z. Continuous wireless pressure monitoring and Page 9 of 9 Lei et al. Journal of Leather Science and Engineering (2019) 1:7 29. Wang X, Huang X, Chen Z, Liao X, Liu C, Shi B. Ferromagnetic hierarchical carbon nanofiber bundles derived from natural collagen fibers: truly lightweight and high-performance microwave absorption materials. J Mater Chem C. 2015;3:10146–53. 30. Wang X, Liao X, Zhang W, Shi B. Bio-inspired fabrication of hierarchical Ni-Fe-P coated skin collagen fibers for high-performance microwave absorption. Phys Chem Chem Phys. 2015;17:2113–20. 31. Wegene JD, Thanikaivelan P. Conducting leathers for smart product applications. Ind Eng Chem Res. 2014;53:18209–15. 32. Thanikaivelan P, Narayanan NT, Pradhan BK, Ajayan PM. Collagen based magnetic nanocomposites for oil removal applications. Sci Rep. 2012;2:230. 33. Gautieri A, Vesentini S, Redaelli A, Buehler MJ. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
https://openalex.org/W2102697386	https://zenodo.org/records/2412933/files/article.pdf	English	null	Notes on the Nomenclature of certain African Tabanidae (Sub-family Pangoniinae), with Descriptions of a new Genus and new Species	Bulletin of entomological research	1,920	public-domain	8,310	NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE (SUB-FAMILY PANGONIINAE), WITH DESCRIPTIONS OF A NEW GENUS AND NEW SPECIES. By Major E. E. AUSTEN, D.S.O. Some six years ago, it was correctly pointed out by Brethes (Bull. Soc. Ent. de France, 1914, p. 59) that the generic designation Diatomineura, Rond. (Archiv. Zool. Anat. Fisiol. iii, p. 84, 1864), is a synonym of Osca, Walk. (Ins. Saund., 'Diptera, Part i, p. 10, 1850—nee Walk., 1864, nee Stal, 1871). The first species mentioned by both Walker and Rondani is Pangonia depressa, Macq. (=Erephop$is lata (Tabanus latus), Guer.), and this species is designated by Brethes as the genotype. Erephopsis lata, Guer., of Kertesz's Catalogus Dipterorum (iii, p. 165, 1908), must therefore be known henceforth as Osca lata, Guer., and the South African Tabanus barbatus, L. (Pangonia barbata, Auct.), and Pangonia fulvifascia, Walk., must be transferred to the genus Osca. It is not, however, simply a question of substituting Osca for Diatomineura, since Osca lata, Guer., is certainly not congeneric with any, or at any rate with the majority of the species at present grouped under Diatomineura, sub-genus Corizoneura. So far as regards Ethiopian and Oriental species, the difficulty can be met by, as is hereby, proposed, raising Corizoneura, Rond. (loc. cit., p. 85) to generic rank, desig- nating as its genotype the first species mentioned by Rondani under Corizoneura, viz., Tabanus aethiopicus, Thunb. (syn. Pangonia appendiculata, Macq.), and erecting a new genus, which may be termed Buplex* for certain Ethiopian species included by Kertesz under the Corizoneura division of Diatomineura, but not congeneric with C. aethiopica, Thunb. The arrangement indicated may be expressed in tabular form as follows. Eyes bare ; first posterior cell open. Ocelli absent; face, especially in $, markedly produced (forming a snout-like prolongation), with a shining callus, or at least a more or less shining area, on each side; proboscis long or very long, usually much longer than thorax including scutellum, and generally horizontal or nearly so ; distal extremities of first and second joints of front tarsus in <§ generally each produced above into a tongue- or lappet-like process, which, in case at least of second joint, is often of considerable relative length .. .. .. .. .. .. .. Corizoneura, Rond. (Genotype, Tabanus aethiopicus, Thunb.,—syn. Tanijglossa aethiopica, Thunb. ; Pangonia appendiculata, Macq.). (Genotype, Tabanus aethiopicus, Thunb.,—syn. Tanijglossa aethiopica, Thunb. ; Pangonia appendiculata, Macq.). 139 139 * /3<w-7rXi)l 6, an ox-goad. For names and illustrations of colours used for descriptive purposes in the present paper, see Eidgway, " Color Standards and Color Nomenclature" (Washington, D.C. Published by the Author, 1912). NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE (SUB-FAMILY PANGONIINAE), WITH DESCRIPTIONS OF A NEW GENUS AND NEW SPECIES. Ocelli usually present; face not produced, merely convex, and without shining callus or area on each side; proboscis of only moderate length, shorter than or about as long as thorax including scutellum, and generally slanting downwards at an angle of 45°; distal extremities of tarsal joints never produced p Buplex, gen. nov. (Genotype Pangonia suavis, Lw.). p Buplex, gen. nov. (Genotype Pangonia suavis, Lw.). /3<w-7rXi)l 6, an ox-goad. 140 E. E. AUSTEN. Among other species referable to the new genus thus briefly characterised are Pangonia brunnipennis, Lw., P. subfascia, Walk., Corizoneura albifacies, Ric... C. dissimilis, Ric, and the new species described below. The genus Corizoneura, Rond., as defined and restricted above, includes, in addition to the genotype and the new species described in this paper :—Diatomineura virgata, Austen ; D. inornata, Austen ; D. neavei, Austen ; Pangonia sagittaria, Sure; Corizoneura distincta, Ric. ; Pangonia lateralis, Wied. (Fabr. ?); Diatomineura hastata, Austen ; D. lineatithorax, Austen ; D. penetrabilis, Austen ; Corizoneura pallidipennis, Ric. ; and C. umbratipennis, Ric. The foregoing all belong to the Ethiopian Region, and congeneric with them are the Oriental Pangonia taprobanes, Walk. (syn. P. rufa, Macq.), and P. longirostris, Hardw. Another change which may as well be made now, since in any case it is almost certain to be effected sooner or later, is the replacement of Macquart's generic designation Cadicera by PJiara, Walk. (Ins. Saund., Diptera, Part i, p. 9, 1850). The latter name, accompanied by a brief diagnosis, was applied by its author to one of " several groups or subgenera " of Pangonia, Latr.; eighteen species were mentioned under Phara, the first three of-these being, in the order given, Pangonia melanopyga, Wied., P. chrysostigma, Wied., and P. crassipalpis, Macq. The present writer hereby proposes to regard Pangonia melanopyga, Wied., as the genotype of Phara, Walk. ; to raise the latter to generic rank; and to restrict it so as to include besides the genotype, among the species mentioned under Phara by Walker, only Pangonia chrysostigma, Wied., and P. crassipalpis, Macq., since these three species are obviously congeneric. Phara, Walk., therefore, as thus restricted, replaces and must be substituted for Cadicera, Macq. (Mem. Soc. Imp. des Sc. Lille, 1854, (2) p. 42, 1855), founded for C. rubramarginata, Macq. (loc. cit, p. 23). NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE (SUB-FAMILY PANGONIINAE), WITH DESCRIPTIONS OF A NEW GENUS AND NEW SPECIES. Except where otherwise stated, the types of the new species described in the following pages are in the British Museum (Natural History). NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE. NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE. Head : ocelli present; face, jowls and lower half of front light buff pollinose, clothed with pale yellowish hair, occiput smoke-grey pollinose, and likewise clothed with pale yellowish hair; upper part of front with a pair of narrow, blackish stripes, commencing at ocellar tubercle, descending somewhat lower than middle and diverging below, each stripe clothed with' short, curving black hairs, interspersed with some glistening yellowish hairs ; palpi clothed with yellowish hair, longer on proximal than on terminal joint, latter russet or ochraceous tawny, proximal joint mouse-grey, russet towards distal extremity; first and second joints of antennae clothed partly with yellowish hair, partly with black hair, first joint mouse-grey or deep neutral grey, second joint russet (third joint missing in case of type). Thorax: smoke-grey longitudinal stripes on dorsum extending from front to hind margin of scutum, confluent posteriorly ; transverse suture smoke-grey, forming a connection between lateral border and outer smoke-grey stripe on each side ; pleurae and pectus light greyish olive pollinose; thorax clothed with yellowish hair, black stripes on dorsum clothed posteriorly with fine black hair. Abdomen: black area on first (visible) and second tergites restricted to a transverse band occupying middle third of anterior half, with rounded posterior angles in each case, and, especially on second segment, indented in middle line behind; lateral margins of fourth and following tergites, and lateral extremities of hind border of third tergite ochraceous tawny, though normally concealed by hair and smoke-grey pollen ; seventh tergite entirely grey; lateral extremities and hind borders of all tergites clothed with glistening hair, longer and ochreous on lateral extremities of second and two or three following segments, shorter and paler elsewhere; black area on second tergite clothed with yellowish hair, corresponding areas on two following tergites clothed with short, erect black hair ; venter smoke-grey, clothed with appressed, glistening, cream- coloured hair ; second sternite with a transversely elongate, somewhat reniform, fuscous black blotch in middle line, resting on or close to anterior border and confined to anterior half of segment, third and three following sternites each with a sooty black, transverse band on front border, widely separated in each case from lateral margins, and from twice to four times deeper on third than on either of the other stemites, these bands clothed with short, erect, black hair. Buplex fuscinervis, sp. n. <£>.—Length (2 specimens) 12 to 13 mm. ; width of head 4 to 4-2 mm. ; width of front at vertex 0'8 mm. ; length of proboscis 3"4 to 3'5 mm. ; length of wing 12'4 to 12-5 mm. Body, except certain areas at sides of abdomen, black above, with scutellum, lateral borders and three narrow longitudinal stripes on scutum, and deep posterior borders to abdominal segments smoke-grey* or pale smoke-grey pollinose; lateral extremities of tergites of first (visible) and second abdominal segments smoke-grey, more or less tinged with ochraceous tawny owing to the subjacent ground-colour ; wings with base, costal and sub-costal cells and stigmatic area ochreous, veins after first longitudinal in distal half (from base of discal cell onwards) dark brown, and for most part more or less strongly suffused with mummy brown. 141 Corizoneura formosa, sp. n. cJ.—Length (4 specimens) 20 to 21-4, mm. ; width of head 6 to 6"4 mm.; distance from upper margin of occiput to anterior extremity of face 5 to 5"25 mm. ; length of proboscis 13 to 15 mm. ; length of wing 17'4 to 18 mm. ; wing-expanse 21-5 to 23 mm. Dorsum of thorax dark olivaceous black, with broad smoke-grey lateral borders clothed with cream-coloured hair, and with two narrow, smoke-grey, admedian, longitudinal stripes ; dorsum of abdomen ochraceous tawny, with distal extremity (fifth to seventh segments) infuscated (iron-grey to olivaceous black), a large black median blotch on each of the first four visible segments, and fourth segment bordered posteriorly with oppressed, silvery white and ochreous hair; venter cinnamon-buff or cinnamon-coloured, with last three segments sometimes mouse-grey and lateral extremities of first visible scute neutral grey ; wings with a sepiaceous tinge ; processess at tips of first and second joints of front tarsus in $ very long. Head : face bluntly conical, moderately produced ; front and face (except sides of facial prominence below, which are clove-brown and shining) pinkish bufE pollinose. front and borders of face adjacent to eyes clothed with longish, cream-buff hair ; occiput smoke-grey, clothed above with cream-buff, below with whitish hair ; jowls and basi-occipital region pale smoke-grey, clothed with long whitish hair; palpi russet, dark brown on outer side, proximal joint with a bunch of whitish hair at base below; first and second joints of antennae isabella-coloured pollinose, clothed above and below with longish hair, generally black or blackish but sometimes mainly cream- coloured on first joint, third joint ferruginous or vinaceous rufous, clove-brown at tip. Thorax: dorsum, including scutellum, clothed for most part with somewhat appressed, cream or cream-buff-coloured hair, often but little visible when regarded from above, hair above lateral borders between bases of wings often largely or mainly dark brown, pleurae and pectus clothed with fairly long, whitish or yellowish white hair. NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE. Wings with auxiliary and first longitudinal veins, second longitudinal from base to a little beyond origin of third longitudinal, praefurcal portion of fourth longitudinal, and extreme bases of fifth and sixth longitudinal veins ochreous ; veins otherwise more or less dark brown. Squamae cream-buff. Halteres maize-yellow, knobs sometimes darker (mummy-brown) at base. Legs : coxae and femora deep mouse-grey, clothed with yellowish hair ; tibiae and tarsi clove-brown, clothed with minute black hairs, tarsal joints sometimes paler towards base. South Africa (Cape Province); type and para-type from Bizana East Pondoland South Africa (Cape Province); type and para-type from Bizana, East Pondoland, 1912(—Goodall: presented by Mr. T. B. Goodall). The species described above shows a decided resemblance to Buplex (Pangonia) suavis, Lw., but, apart from its smaller size, is distinguishable, inter alia, by the presence of the two blackish stripes on the front; by the outer two of the three longitudinal smoke-grey stripes on the disk of the scutum being narrower ; by the absence of a brown blotch near the tip of the wing, above the fork of the third longi- tudinal vein ; and by the veins in the region of the discal cell being suffused with mummy-brown. The species described above shows a decided resemblance to Buplex (Pangonia) suavis, Lw., but, apart from its smaller size, is distinguishable, inter alia, by the presence of the two blackish stripes on the front; by the outer two of the three longitudinal smoke-grey stripes on the disk of the scutum being narrower ; by the absence of a brown blotch near the tip of the wing, above the fork of the third longi- tudinal vein ; and by the veins in the region of the discal cell being suffused with mummy-brown. 142 E E AUS 142 E. E. AUSTEN. Corizoneura formosa, sp. n. Abdomen : except on first (visible) tergite, on which median blotch extends to or is but narrowly separated from hind margin, black median blotches, which rest on front margin, do not reach hind border ; posterior angles of first and second tergites clothed with yellowish or whitish hair, hind border of second tergite with a patch of glistening, appressed, ivory-yellow or cream-buff-coloured hair in middle line, light-haired hind border of fourth tergite expanded in middle line and at each extremity, lateral extremities of sixth and seventh tergites generally clothed with glistening silvery white hair, lateral extremities of second to fourth tergites inclusive, except as already stated, clothed with black hair, ochraceous tawny area of first and second tergites clothed mainly with minute, glistening, appressed, ochreous hairs, corresponding area on third and fourth tergites clothed with minute black or blackish hairs, fifth and sixth tergites (except lateral extremities in case of latter) clothed with black hair ; venter clothed with minute, appressed, glistening, cream-buff- coloured hairs, fifth sternite sometimes largely clothed with minute, black or blackish hairs. Wings : veins sepia-coloured, adjacent membrane suffused with same colour at base, and to a less extent at level of proximal extremity of discal cell. Squamae waxen ochraceous bufE or pale orange-yellow, fringed with minute yellowish hairs. Halteres light chestnut-brown, tips of knobs light buff or light ochraceous buff. Halteres light chestnut-brown, tips of knobs light buff or light ochraceous buff. Legs : coxae neutral grey, clothed with whitish hair, which is longer on front pair; 143 NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE. femora russet (extreme tips ochraceous buff), clothed mainly with whitish or yellowish white hair, antero-inferior area of front pair with minute, erect, blackish hairs ; front and middle tibiae ochraceous buff, clothed with minute, glistening, yellowish or ochreous hairs, hind tibiae and hind tarsi russet-brown, clothed with black hair, distal extremities of hind tarsal joints blackish; front tarsi ochraceous buff or ochraceous tawny, distal extremities of last three joints dark brown, body of second joint very short, tongue-like process at its distal extremity very long (longer than following joint), process at end of first joint overlapping and closely applied to first two-thirds of following joint; first joint of middle tarsi cream-buff, remaining joints dark brown or mummy-brown, blackish brown at tips. Corizoneura formosa, sp. n. South Africa (Cape Province): type and three para-types from Deelfontein, 7.xii.l902 (presented by Colonel Sloggett, A.M.S.—now IAeut.-General Sir A. T. Sloggett, K.C.B., K.G.M.G.). This handsome species presents a decided resemblance to Corizoneura aethiopica, Thunb., another inhabitant of the South African portion of the Ethiopian Region, but can at once be distinguished, inter alia, by the presence of the conspicuous, smoke-grey, admedian, longitudinal stripes on the dorsum of the thorax. Corizoneura schwetzi, sp. n. (J 9-—Length, (J (6 specimens) 16'4 to 18 mm., 9 (6 specimens) 16 to 18"6 mm. ; width of head, $ just over 5 to 5'4 mm., 9 5'2 to 5"75 mm.; distance from upper margin of occiput to anterior extremity of face, <J 4-2 to 4-5 mm., $'4"5 to 4"8 mm.; width of front of 9 at vertex 06 to O75 mm. ; length of proboscis, £ 15 to 15-25 mm., 9 11-75 to 13-75 mm. ; length of wing, $ 15 to 15"2 mm., 9 15"5 to 17-25 mm. Dusky species, with base of abdomen paler; dorsum of thorax dark brownish olive, with a broad, indistinct, blackish brown longitudinal stripe along each side above latera margin, a narrow, elongate, pale spot {composed of Naples yellow or cream-buff hair) above base of each wing in front of postalar callus, and a short but conspicuous stripe of black hair between base of costa and humeral callus on each side; dorsum of abdomen with first two (visible) segments cinnamon-buff or tawny olive in <$, ochraceous tawny in 2, {? with a blackish brown median triangular spot at base of second segment, tergites of third and following abdominal segments blackish brown or black, fourth segment posteriorly with a conspicuous transverse band of appressed, glistening, silvery white hair; venter abruptly bicoloured, proximal portion as far as base of third (visible) stemite, ivory-yellow, cream or cream-buff-coloured, remainder blackish brown; wings strongly and uniformly tinged with sepia. Head drab pollinose, posterior orbits paler (light greyish olive above, smoke- grey or pale smoke-grey below), shining callus on each side of face below antenna black, sharply denned, sides of face below caHi more or less dark brown or blackish brown owing to ground-colour showing through pollinose covering, lower border of sides of face in $ shining blackish brown, deeper posteriorly, distal extremity of face in 9 shining blackish brown, sparsely clothed with drab pollen, in both sexes a more or less distinct, somewhat triangular, olive-buff or pale olive-buff, pollinose spot on each side of lower part of front, between base of antenna, margin of eye and shining callus ; front in $ with following series of dark brown marks— (687) D D 141 E. E. AUSTEN. Corizoneura schwetzi, sp. n. a transversely oblong spot occupying ocellar region of vertex but not extending to eyes, a pair of somewhat guttate spots in centre of front, indistinctly connected with foregoing spot but likewise not in contact with eyes, and an oblique spot on each side below, extending from base of antenna to eye, above the light (olive-buff or pale olive-buff) spot already mentioned ; front in <$ clothed with black hair in $ with shorter blackish hair, mingled with paler hair in upper portion, posterior orbits in both sexes fringed above with yellowish hair and below with longer whitish hair, basi-occipital region and jowls densely clothed with fairly long, whitish hair, lower borders of sides of face clothed posteriorly with black or blackish hair, more conspicuous in <J than in $ ; palpi russet-brown in $, russet in $, terminal segment {elongate and very narrow in $) clothed with minute black or blackish hairs, proximal segment, especially in $, clothed below with longer hair, brownish or blackish brown anteriorly, yellowish or whitish posteriorly ; antennae russet, first and second joints more or less smoke-grey pollinose, clothed with black hair. Thorax : dorsum clothed anteriorly with silky, tawny olive or ochreous hair, posteriorly, including base of scutellum, with fine, erect, black hair, postalar calli, and scutellum, except as stated, clothed with Naples yellow or cream-buff hair, hairy covering of central portion of dorsum visible only when viewed from side ; pleurae and pectus clothed with longish hair, for most part cream-buff (whitish on propleurae) in colour, a tuft of black hair (less conspicuous in §, in which sex it is often much reduced) below base of wing on each side ; sternopleurae in (J sometimes clothed mainly with black or blackish hair. Corizoneura schwetzi, sp. n. Legs : front coxae neutral grey, clothed with cream-coloured or cream-buff hair, middle and hind coxae deep neutral grey or dark neutral grey, clothed with black hair, hind coxae in 2 also with some yellowish hairs ; hind femora in $ and bases of front and middle femora in same sex blackish brown, front and middle femora in $ except at base russet-brown or cinnamon- brown, all femora in £ clothed with black hair, femora in 2 paler, and clothed largely with ochraceous buff or ochreous hair as well as with black hair ; coloration of tibiae and tarsi and of their hairy covering alike in both sexes, front and middle tibiae ochraceous buff or ochraceous tawny, clothed with minute, appressed glisten- ing ochraceous buff hair, hind tibiae and hind tarsi blackish brown, clothed with black hair, tips of second and two following tarsal joints clothed below with ferruginous hair, similarly coloured hair also largely present, at least in 2, on under side of first tarsal joint, front and middle tarsi ochraceous tawny, clothed above with black hair, last joint in each case, as well as distal extremities of preceding joints mummy-brown, processes at tips of first and second joints of front tarsi in (J of moderate size, in neither case reaching distal extremity of following joint. B l i C (N h K ) d T ik T i T i l i Belgian Congo (North Katanga) and Tanganyika Territory. Typical series from Kakanu (between 15 and 16 miles south of Kisengwa, E. Lomami), N. Katanga, vi. 1918 (Dr. J. Schwetz). Type of <$, type of 2, 36 <J and 2 2 para-types, in Mus Eoyal d'Histoire Naturelle de Belgique (Brussels) ; 6 <$ and 6 2 para-types, in British Museum (Natural History)—presented by M. G. Severin ; 1 2, from Tangan- yika Territory (formerly German East Africa), 30° 55' E. Long., 2° 5' S. Lat., 16. vi. 1916 (Dr. G. D. H. Carpenter), in British Museum (Natural History), presente by Imperial Bureau of Entomology. This fine species, with which the author has much pleasure in connecting the name of its discoverer (the well-known student of tsetse-fly bionomics in North Katanga), was met with by the investigator in question in large numbers in the vicinity of Kakanu.* On 6th June 1918, in a belt of forest several hundred metres in breadth and about 3f miles from Kakanu, Dr. Corizoneura schwetzi, sp. n. Abdomen : hind border (distal third or rather less) of second (visible) tergite clothed with minute, appressed hairs, which are silvery white or whitish at and towards lateral extremities of segment, and sparser and more yellow- ish (often glistening cream-buff) in vicinity of median line (owing to paler ground colour, whitish transverse band thus formed is, except at lateral extremities, less •conspicuous than corresponding band on fourth segment); base of first (visible) tergite with a blackish brown, median area, extending beyond scutellum, but not or barely reaching hind margin except in middle line in <J, in which sex it is more or less distinctly connected with triangular spot on second segment; blackish brown median triangle on second tergite in $ with its base resting on or close to anterior margin, and its apex reaching or extending somewhat beyond middle of segment; third tergite sometimes irregularly paler (cinnamon-brown) at base and on hind border ; hairs in silvery white transverse band on fourth tergite somewhat yellowish in vicinity of middle line; seventh tergite in <j>, as well as frequently hind borders of fifth and sixth tergites, often dull fuscous; lateral extremities, or at least posterior angles, of fifth and sixth tergites each clothed with a prominent tuft of silvery white hair ; posterior angles, as well as in $ hind margin of first (visible) tergite clothed with shining ochreous or pale ochraceous orange hair; dorsum of abdomen except as already stated clothed with minute, appressed, black hairs ; first (visible) sternite bare, second sternite clothed with minute, appressed, glistening cream-coloured hairs (anterior border of second sternite fringed with fine, erect, black or blackish hair, shorter and less developed in $ than in $, in which sex fine, recumbent, black hairs are also present among the cream-coloured hair, especially towards lateral extremities of the scute), blackish brown portion of venter clothed with black hair, among which a few pale hairs are occasionally present NOTES ON THE NOMENCLATURE OP CERTAIN AFRICAN TABANIDAE. 145 on one or more of fourth and following segments. Wings : appendix to anterior branch of third longitudinal vein, though of variable length, as a rule relatively somewhat long. Squamae isabella-coloured. Halteres mummy-brown, knobs in 2 often paler (cream-buff) at tip. * Of. Schwetz, " Dix Jours d'Observations sux les Moeurs de la ' Pangonia zonata ' et de la 'Pangoniaoldii' (Deuxieme Note)" : BevueZoologique Africaine, vii,pp. 92-106 (1919).—Cf. also the earlier paper by the same author, " Quelques Observations Pr6- liminaires sur les Moeurs de la ' Pangonia zonata ' " : ibid., pp. 46-54. In both of the memoirs cited Gorizoneura schwetzi is referred to as Pangonia oldii, while the species termed Pangonia zonata is really Gorizoneura inornata, Austen. * Of. Schwetz, " Dix Jours d'Observations sux les Moeurs de la ' Pangonia zonata ' et de la 'Pangoniaoldii' (Deuxieme Note)" : BevueZoologique Africaine, vii,pp. 92-106 (1919).—Cf. also the earlier paper by the same author, " Quelques Observations Pr6- liminaires sur les Moeurs de la ' Pangonia zonata ' " : ibid., pp. 46-54. In both of the memoirs cited Gorizoneura schwetzi is referred to as Pangonia oldii, while the species termed Pangonia zonata is really Gorizoneura inornata, Austen. t Of. Schwetz, loo. dt., p. 103. J See below, p. 147. (687) D2 t Of. Schwetz, loo. dt., p. 103. Corizoneura schwetzi, sp. n. Schwetz's native carriers succeeded in catching some 2,000 specimens of C. schwetzi, from 80 to 90 per cent- of which were males.f Unlike Corizoneura inornata, Austen (see below), which is found in the open, C. schwetzi does not occur outside the forest. According to Dr. Schwetz, in the case of the present and the following species (C. inornata, Austen) at least, the labium itself is the piercing organ, and in the act of biting is thrust by the insect deeply into the skin of its victim. It is therefore interesting and possibly suggestive to note that, in the 2 taken by Dr. Carpenter in Tanganyika Territory, the labium D2 146 E. E. AUSTEN. only projects 7"5 mm. beyond the extremity of the clypeus, exceeding the length of the labrum-epipharynx and the other mouth-parts by little more than the labella, while the proximal portion of the labium is bent backwards underneath the head, and beneath the cleft between head and thorax forms an angle of 45° with the distal portion. Although superficially presenting a decided resemblance to Pangonia oldii, Austen, Corizoneura schwetzi can be distinguished from that species inter alia by the processe (entirely wanting in P. oldii) at the tips of the first and second joints of the front tarsi in the $; by the much greater development of the dark spot (often scarcely more than vestigial in P. oldii) at the base of the second abdominal tergite in the same sex; and in both sexes by the short stripe of black hair on each side of the dorsum of the thorax, in front of the base of the wing. From Corizoneura inornata, Austen, apart from obvious differences in coloration especially the sharply bicoloured venter of the species just described, C. schwetzi is distinguished by the inferior development of the processes at the tips of the first and second joints of the front tarsi in the <$. Whereas in 0. inornata $ each of these processes is so long as to project beyond the tip (excluding the process in the case of the second segment) of the following joint, in C. schwetzi <$ neither process reach the tip of the succeeding joint. P. 145, note. Corizoneura inornata, Austen. Diatomineura inornata, Austen, Bull. Ent. Res. i, p. 282 (1911) Diatomineura inornata, Austen, Bull. Ent. Res. i, p. 282 (1911) This species was described from a single 2, obtained in September 1907, in S. Katanga, Belgian Congo, between Bunkeya and Kambove, at an altitude of 3,500 ft., by Dr. Sheffield Neave. The kindness of Dr. J. Schwetz, in presenting to the British Museum (Natural History) a large number of specimens of both sexes, now renders it possible to indicate the distinctive characters of the <$. C. inornata, Austen, £.—Apart from ordinary secondary sexual characters, agreeing generally with the 2 except as follows. Head: hair clothing jowls and basi-occipital region often hoary or nearly white; first two joints of antennae clothed mainlv with black hair. Thorax : pleurae on each side with a tuft of black hair below base of wing. Abdomen : first (visible) tergite with a median brownish black area at base, projecting somewhat beyond scutellum, but not reaching hind margin ; second tergite with a conspicuous, median, brownish black, triangular spot resting on front margin, and varying in size in different individuals, but not extending beyond middle of segment, if so far ; fifth and sixth tergites mainly brownish black, blackish brown or clove-brown, each of the two preceding tergites often with an ill-defined median blotch of same colour occupying anterior two-thirds. Legs : front and middle femora clothed largely with black hair, at least towards base, hind legs clothed mainly with black hair, processes at tips of first and second joints of front tarsi very long, in each case projecting beyond end of following joint (excluding process in case of second segment). In the papers already referred to,* under the name " Pangonia zonata," Dr. Schwetz has furnished a series of interesting field notes on the behaviour of this NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE. 147 species, which, at the end of May 1914 and at the close of the same month and beginning of June 1918, was found by him in great abundance in the vicinity of Kakanu, N. Katanga, at and about the flowers of Acanthus montanus (termed by the natives " Nafimbia"), which grows in sheets in open spaces outside the forest. According to Dr. Schwetz, the existence of C. Corizoneura inornata, Austen. inornata is " intimately connected " with that of the flowers in question, the nectar of which is imbibed by both sexes, though the females also suck blood on occasion. It is interesting to note that, having in a number of cases observed the process of sucking blood by females of this and the foregoing species (Corizoneura schwetzi, Austen), Dr. Schwetz states tha in these PANGONIINAE at any rate, the actual piercing organ is the proboscis (labium itself, which is " evidently sufficiently rigid " to penetrate the human integument, and is driven by the insect " fairly and squarely into the skin to the extent of one-third or one-half of its length. . . ." * * Of. Schwetz, Rev. Zool. Afrioaine, vii, pp. 101-102 (1919) Pangonia discors, sp. n. g p $.—Length (1 specimen) 19'6 mm. ; width of head 56 mm. ; width of front at vertex 0-75 mm. ; distance from upper margin of occiput to anterior extremity of face 4'5 mm. ; length of proboscis 5'5 mm. ; length of wing 17"6 mm. Deep black, somewhat shining ; first (visible) and second abdominal tergifes densely covered with pale gull-grey pollen, and thickly clothed with closely oppressed, silvery white hair ; area beneath scutellum free from pollen and clothed with black hairs, some black hairs also present in middle line on anterior border of second tergite, latter likewise exhibiting a broad, median, triangular area (its base resting on front margin, its apex directed backwards and reaching beyond middle of segment) which, like front border of same segment, has a blackish look owing to pollen on it being thinner than elsewhere ; lateral extremities of second tergite clothed with black hair; wings mummy-brown, proximal half (as far as base of discal cell) and stigma pale orange-yellow or light orange- yeUow. Head: ocelli wanting ; face moderately prominent, front above relatively some- what narrow; area from middle of front to anterior margin of clypeus olive-buff, upper half of front fuscous black, sides of face clove-brown, occiput and basi-occipital region smoke-grey or pale smoke-grey, jowls clothed with yellowish cream-coloured hair ; palpi clove-brown, terminal joint elongate, tapering to a point and somewhat curved; first and second joints of antennae dark mummy-brown, clothed with minute black hairs mixed with some minute yellowish hairs, first joint short, expanded portion of terminal joint cinnamon-brown (terminal portion missing in case of type). Thorax: postalar calli chestnut-brown; dorsum including scutellum clothed with short black hair, which on front border of scutellum is interspersed with minute, glistening, appressed, golden hairs, a few golden hairs also on upper portion of swelling occupying depression at each end of transverse suture, hair on pleurae and pectus entirely black or fuscous black. Abdomen: third (visible) and following tergites clothed with appressed, black hair ; venter, except second sternite, clothed with appressed, black or brownish black hair, ventral surface of second segment 148 E. E. AUSTEN. clothed with minute, appressed, glistening, yellowish white hairs, fore border and lateral extremities clothed with black hairs. Wings : transition from orange- yellow proximal to mummy-brown distal portion sharply marked, at least in case of type. Squamae ivory-yellow. Pangonia discors, sp. n. Halteres ochraceous buff, stalks and knobs towards base brownish. Legs : coxae and femora dark brown or blackish brown, clothed with black hair, which on under side of hind femora is mixed with a certain number of minute, glistening tawny hairs, and on anterior surface of front coxae with a few golden hairs ; tibiae and tarsi russet, clothed for most part with minute, appressed, glistening, ochraceous tawny hairs, last three joints of tarsi dark brown above. Angola (J. J. Monteiro). Angola (J. J. Monteiro). Pangonia discors resembles and is closely allied to the East African P. beckeri, Bezzi, but is readily distinguishable owing to, inter alia, its more prominent face and narrower upper part of the front, the entire absence of white hair on pleurae, postalar calli and front coxae, and the distal portion of the abdomen being entirely black and covered with black hair, instead of having the tip ochraceous tawny and clothed with glistening ochraceous orange or ochreous hair. Pangonia lautissima, sp. n. Abdomen : first six (visible) tergites in both sexes each with its lateral fourth on each side (rather more in case of $) clothed with close-set, appressed, minute, glistening smoke grey hairs (extreme lateral extremities of seventh tergite in $ clothed with hairs of same kind), so that dorsum of abdomen exhibits on each side a broad grey stripe, which in certain aspects contrasts fairly sharply with remainder of surface, which is clothed with minute appressed black hairs and thus forms a broad, longi- tudinal, median, black stripe ; first (visible) tergite clothed with greyish pollen on each side at base, second tergite in both sexes on each side with a roughly semi- circular whitish pollinose spot on hind margin, from most points of view concealed by the smoke-grey hair, but clearly visible when abdomen is looked at obliquely from behind, in 9 a Pair °f similar but smaller whitish pollinose spots on hind margin of each of the two following tergites also ; venter clothed with minute, appressed, glistening, smoke-grey hair, last two sternites, and sometimes median area of hind border of preceding sternite also, clothed mainly or entirely with black hair. Wings : extreme base of costa and first longitudinal vein brownish black ; ochraceous orange base extending into bases of basal and anal cells, similarly coloured anterior border including upper margin of first basal and proximal fourth of first submarginal cell, thence tapering obliquely to end of third costal cell; veins within ochraceous • orange area similarly coloured, elsewhere dark brown. Squamae light orange yellow. Halteres: stalks and knobs mummy-brown above an below, tips of knobs cream-buff. Legs black and clothed with black hair, front coxae anteriorly dark neutral grey pollinose, thinly clothed towards base with longish pale orange-buff hair. Tanganyika Territory (formerly German East Africa) : Itigi, iv, 1917 (Dr. G. D. H. Carpenter). Type of £, type of $, and 2 $ para-types, taken 18.iv. 1917 ; 7 £ para-types, taken 15.iv. 1917, "on composite flower"; 1 £ para-type, taken 6.iv. 1917, " among low herbage." All foregoing presented by Imperial Bureau of Entomology, which retains .possession of six additional para-tj^pes, taken by Dr. Carpenter at same time and place as specimens already enumerated. The extremely striking and unusual-looking Tabanid just described belongs to the group of the genus Pangonia that includes P. elongata, Bic, P. becleeri, Bez and P. Pangonia lautissima, sp. n. Pangonia lautissima, sp. n. $2.—Length, $ (9 specimens) 17 to 19 mm., 2 (3 specimens) 16 to 17"4 mm. ; width of head, <§ just under 5 to 5"4 mm., 2 5 to 5"6 mm. ; width of front of 2 at vertex just under 1 to 1'2 mm. ; length of proboscis, (J just under 4 to 4-25 mm., 2 3-4 to 4-2 mm. ; length of wing, <J 14 to 15"2 mm., 2 14 to 15"6 mm. Shining black ; basi-occipital region clothed and lower halves of posterior orbits fringed with orange-buff hair, and patches of similarly coloured hair on pleurae ; wings with base and a deep anterior border extending to end of third costal cell ochraceou orange, and remainder of surface uniformly brownish black, with a strong purplish metallic sheen. Head black, frontal triangle in <§ and region of subcallus (area immediately above bases of antennae) in 2 shimmering silvery white pollinose, a similar pollinose patch (clothed with a few whitish hairs, and usually more distinct and sharply defined in 2 than in $) on each side of upper part of face in both sexes ; face in both sexes tumid below antefmal prominence, then indented or somewhat receding, not produced into a nose-like prolongation, front in 2 deeply furrowed ; occiput pallid neutral grey pollinose, clothed with whitish hair, posterior orbits silvery white, their upper halves fringed behind with minute blackish hairs ; palpi and antennae black, proximal joints of both sparsely clothed with short, black or blackish hairs, third joint of antennae from certain aspects appearing dark olive-grey, mouse-grey or brownish grey pollinose. Thorax : dorsum including scutellum clothed with short black hair, humeral calli inconspicuously neutral grey pollinose and clothed on sides in front with pale orange-buff hair, postalar calli fringed posteriorly below with orange-buff hair ; pleurae on each side with a thick tuft of orange-buff hair below humeral callus, and more posteriorly with two further tufts of similar hair arranged somewhat in the shape of a wide V, of which the anterior branch fringes the hind margin of the mesopleura, while the posterior runs back to the squamae ; pleurae except as stated, and pectus except in front of front coxae clothed with black hair, pectus in front of front coxae neutral grey pollinose, clothed with pale orange-buff 149 NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE. hair. Pangonia lautissima, sp. n. discors, Austen, but is readily distinguishable from all three species by, apart from its sharply defined wing-markings and other characters, its unbanded abdomen and the patches of orange-buff hair on the pleurae. The coloration of the wings, in conjunction with the shining black body, would seem to suggest that P. lavtissima possibly mimics some species of wasp (perhaps Khynchium cyanopterum Sauss.), and it is worthy of note that the same colours, though differing widely in extent and arrangement in the case of the wings, are exhibited by " Pangonia " mesembrinoides, Sure, of which the type was also obtained in Tanganyika Territor (Amani). The latter species, however, was incorrectly assigned by its describer to the genus Pangonia, and really belongs to a new genus allied to Thriambeutes, Griinb E. E. AUSTEN. E. E. AUSTEN. In $ sex, at any rate, a somewhat sombre-coloured species looking more like a small or medium-sized Corizoneura than a Pangonia, and not unlike a smaller and more dusky form of Corizoneura hastata, Austen, of Portuguese E. Africa.—Face moderately produced ; dorsum of thorax olivaceous black, with a pair of broad, paler (greyish olive), narrowly separated or sometimes confluent, admedian, longitudinal stripes ; dorsum of abdomen shining blackish brown, on each side of base rather more than lateral third of first (visible) tergite ochraceous tawny, a similar area on each side of second tergite, or of second and third tergites, usually russet or chestnut-brown, lateral borders, posterior angles, and (at least in part) hind borders of second and fourth tergites clothed with glistening silvery white hair; wings strongly tinged with sepia ; legs for most part ochraceous tawny. Head : ocelli wanting ; face and front clothed with yellowish grey or isabella- coloured pollen, and front thinly covered with yellowish hair, lower half of front, above antennal prominence, with an ill-defined, shining black, median longitudinal mark, upper half of front usually streaked with mummy-brown or dark brown along each side, while a narrow mummy-brown streak, which starts from base of antenna on each side, runs obliquely upwards and outwards, and joins corresponding eye just above base of antennal prominence ; below antennae, a shining black transverse band unites and includes the shining black facial calli'; jowls and basi-occipital region clothed with long, whitish hair ; occiput greyish olive or smoke-grey pollinose, sparsely clothed with cream-coloured hair, which also forms a short fringe behind posterior orbits above, lower portion of posterior orbits fringed with whitish hair ; palpi elongate, proximal joint russet-brown or deep mouse-grey, sparsely clothed with whitish hair on outer side and below, terminal joint russet, somewhat expanded in middle, clothed with minute black hairs ; first and second joints of antennae. pinkish cinnamon or cinnamon-buff pollinose, both sometimes clothed with black hair though in some specimens hair on first joint is for most part yellowish, second joint usually with outstanding black hair above and below, third joint orange- cinnamon. Pangonia carpenteri, sp. n. ?.—Length (8 specimens) 15 to 16-6 mm. ; width of head 4-25 to just under 5 mm. ; width of front at vertex 0'6 mm. ; distance from upper margin of occiput to anterior extremity of face 3'5 to 4 mm. ; length of proboscis 1T75 to 13-75 mm. ; length of wing 13'5 to 15 mm. ?.—Length (8 specimens) 15 to 16-6 mm. ; width of head 4-25 to just under 5 mm. ; width of front at vertex 0'6 mm. ; distance from upper margin of occiput to anterior extremity of face 3'5 to 4 mm. ; length of proboscis 1T75 to 13-75 mm. ; length of wing 13'5 to 15 mm. 150 E. E. AUSTEN. Thorax : admedian stripes on dorsum entire, extending from front margin to prescutellar groove, outer border of each stripe paler in front of transverse suture ; dorsum including scutellum thickly clothed with fine yellowish hair, with which in some specimens on central part of posterior portion of scutum a few fine black or blackish hairs are intermixed ; postalar calli and lateral margins of dorsum behind transverse suture fringed with whitish hair, pleurae and pectus thickly clothed with similar hair. Abdomen : central portion of base of first (visible) tergite olivaceous black, basal angles of same tergite more or less distinctly neutral grey or dark neutral grey ; rather less than median third of second tergite and a broad median triangle on each of following tergites, in each case with its base resting on hind margin and its apex in contact with front margin, and on fourth and following tergites with its base expanded laterally so as to include entire hind border of segment, •dull dark olive-grey ; olive-grey median area on second tergite (sometimes that on third tergite also) indented on each side ; second tergite at base with a median, semicircular, dead black spot, or in some cases with a pair of smaller spots, narrowly separated in middle line by a dark olive-grey longitudinal stripe; first (visible) tergite with a median patch of glistening yellowish or pale yellowish hair, and clothed elsewhere with ochraceous hair; second tergite, except on black spot 151 NOTES ON THE NOMENCLATURE OF CERTAIN AFRICAN TABANIDAE. (or spots) on which hair is usually black or ochreous, clothed with minute, appressed, glistening silvery white hairs, which however, at least on each side of median olive- grey area, are often largely replaced by similar ochreous hairs ; lateral borders of fourth and following tergites, as well as (at least in part) hind border of fourth tergite, clothed with glistening, silvery white hair, similar hair also usually visible on lateral margins of third tergite, towards posterior angles ; dorsum except as stated clothed with appressed black hair ; venter isabella-coloured or light brownish olive, clothed with minute, appressed, whitish or yellowish white hairs. Wings : veins mummy-brown ; first posterior cell variable as usual as regards distance from hind margin at which it is closed, sometimes closed on margin itself, or even in one or other wing narrowly open. Squamae cream-buff. Halteres : knobs fuscous, stalks cinnamon-buff. E. E. AUSTEN. Legs : coxae neutral grey, clothed with whitish hair, anterior and inferior surfaces of femora clothed, at least in part, with black hair, femora elsewhere clothed with yellowish hair, tibiae and tarsi clothed with minute, appressed, ochreous hairs, extensor surfaces of hind tibiae and hind tarsi clothed, at least in part, with black hair ; front femora blackish brown at base and sometimes also on greater part of under side, joints of front tarsi often mummy- brown or dark brown at tips above, those of hind tarsi similarly marked, or some- times entire upper surface of hind tarsi, except base of first joint, dark brown. Tanganyika Territory : Itigi, iv, 1917 (Dr. G. D. H. Carpenter). Type and 3 para-types, taken 18.iv.1917 ; 1 para-type, taken 15, iv, 1917, " on low herbage "; 1 para-type, taken 6.iv.l917, " hovering while feeding from composite flower ; looking much like a Bombylius." All foregoing presented by Imperial Bureau of Entomology, in whose possession are two other para-types, taken by Dr. Carpenter at same time and place as specimens already mentioned. So far as it is possible to judge from the $ alone, this species, which is named in honour of its discoverer and does not resemble any African Pangonia hitherto described, presents, apart from the venation, all the characteristics of a Corizoneura, to which genus it would have been assigned were it not that its posterior cell seems normally to be closed before reaching the wing-margin. Should the <J prove to have processes at the tips of the first and second joints of the front tarsi, Pangonia carpenteri, despite the transitional character of its venation, would more fittingly be placed under Corizoneura, so long as the independence of the latter be maintained. E. E. AUSTEN. E. E. AUSTEN. Head black or blackish brown, occiput dark neutral grey pollinose, basi-occipital region thinly clothed with fine blackish brown hair ; ocelli present, enlarged facets of eyes very coarse, area occupied thereby same as in $ of genotype (Thriambeutes singularis, Griinb., of Togoland and S. Nigeria), i.e., small facets confined to a deep lower border and a narrower hind border of uniform width running up to ocelli; palpi blackish brown and clothed with fine hair of same colour, terminal joint elongate and curved but not conspicuously swollen, considerably smaller and narrower than in <J of genotype ; antennal protuberance large and prominent, considerably larger and more prominent than in <$ of genotype ; first joint of antennae blackish brown, short, swollen, cylindrical, and clothed like second joint with blackish brown hair, second and third joints sepia-coloured, expanded portion of third joint rather broad. Thorax and abdomen thinly clothed with fine blackish brown hair. Wings : anal angle and lower region of distal extremity short of actual tip paler than elsewhere with exception of clear, transverse streak, a close scrutiny, when wing is viewed against a light back-ground, revealing beyond clear streak an ill-defined transverse band, which appears somewhat darker than remainder of surface ; stigma well developed, elongate, cinnamon-brown when seen against a light background. Squamae blackish brown. Halteres : knobs ivory yellow, stalk sepia-coloured. Legs : coxae, femora and tibiae clothed with blackish brown or blackish hair, middle as well as front tibiae swollen (front and hind tarsi, and hind tibiae missing in case of type). South Africa, Bechuanaland Protectorate : N'Gami Country, 1897 (Sir Frederick Lugard, G.C.M.G., C.B., D.S.O.). South Africa, Bechuanaland Protectorate : N'Gami Country, 1897 (Sir Frederick Lugard, G.C.M.G., C.B., D.S.O.). The species characterised above is readily distinguishable by its wing-markings alone, apart from all other characters, from Thriambeutes singularis, Griinb. (the only other member of its genus as yet described), in which moreover the body as well as the head and its appendages are in the $ uniformly tawny. So far as it is possible to judge from a photograph, which is all that is at present available for comparison, what appears to be another £ of Thriambeutes fuscus is in the possession of Mr. R. W. Jack, Government Entomologist, Southern Rhodesia, and was taken by him in November 1914, in Sebungwe District, Southern Rhodesia, on the jacket of a companion. In Mr. Genus Thriambeutes, Griinb. Thriambeutes fuscus, sp. n. $.—Length (1 specimen) 1114 mm.; width of head 4-25 mm.; length of wing 105 mm. Thriambeutes fuscus, sp. n. $.—Length (1 specimen) 1114 mm.; width of head 4-25 mm.; length of wing 105 mm. $.—Length (1 specimen) 1114 mm.; width of head 4-25 mm.; length of wing *5 mm. Dorsum of thorax sepia-coloured, with traces of a faintly marked, paler, longitudinal median stripe in front of transverse suture, pleurae and pectus mummy-brown ; abdomen uniformly blackish brown ; wings mummy-brown, with a clear oblique transverse streak, commencing on anterior transverse vein (its base extending from commencement of lower border of distal fourth of first basal cell to proximal extremity of first posterior cell), including rather more than proximal third of' discal cell, proximal extremity of fourth posterior cell, distal extremity of second basal cell, and upper border of proximal two- thirds of fifth posterior cell, but not reaching hind margin ; legs blackish brown or black, middle tarsi cream-buff, last joint and tips of preceding joints cinnamon-brown. 152 E. E. AUSTEN. Jack's specimen, however, the clear streak in the wing reaches the hind margin, while the margin of the anal angle, and a further portion of the hind border embracing part of the distal extremity of the second submarginal cell and the distal extremities of the first three posterior cells are also hyaline.
https://openalex.org/W2130235085	https://research.vu.nl/files/2913753/264182.pdf	English	null	Prevention of anxiety and depression in the age group of 75 years and over: a randomised controlled trial testing the feasibility and effectiveness of a generic stepped care programme among elderly community residents at high risk of developing anxiety and depression versus usual care [ISRCTN26474556]	BMC public health	2,006	cc-by	6,819	VU Research Portal Prevention of anxiety and depression in the age group of 75 years and over: a randomised controlled trial testing the feasibility and effectiveness of a generic stepped care programme among elderly community residents at high risk of developing anxiety and depression versus usual care [ISRCTN26474556] Tazelaar, P.J.; van Marwijk, H.W.J.; van Oppen, P.C.; Nijpels, G.; van Hout, H.P.J.; Cuijpers, P.; Stalman, W.A.B.; Beekman, A.T.F. published in BMC Public Health 2006 DOI (link to publisher) 10.1186/1471-2458-6-186 document version Publisher's PDF, also known as Version of record Link to publication in VU Research Portal citation for published version (APA) Tazelaar, P. J., van Marwijk, H. W. J., van Oppen, P. C., Nijpels, G., van Hout, H. P. J., Cuijpers, P., Stalman, W. A. B., & Beekman, A. T. F. (2006). Prevention of anxiety and depression in the age group of 75 years and over: a randomised controlled trial testing the feasibility and effectiveness of a generic stepped care programme among elderly community residents at high risk of developing anxiety and depression versus usual care [ISRCTN26474556]. BMC Public Health, 6. https://doi.org/10.1186/1471-2458-6-186 citation for published version (APA) Tazelaar, P. J., van Marwijk, H. W. J., van Oppen, P. C., Nijpels, G., van Hout, H. P. J., Cuijpers, P., Stalman, W. A. B., & Beekman, A. T. F. (2006). Prevention of anxiety and depression in the age group of 75 years and over: a randomised controlled trial testing the feasibility and effectiveness of a generic stepped care programme among elderly community residents at high risk of developing anxiety and depression versus usual care [ISRCTN26474556]. BMC Public Health, 6. https://doi.org/10.1186/1471-2458-6-186 citation for published version (APA) Tazelaar, P. J., van Marwijk, H. W. J., van Oppen, P. C., Nijpels, G., van Hout, H. P. J., Cuijpers, P., Stalman, W. A. B., & Beekman, A. T. F. (2006). Prevention of anxiety and depression in the age group of 75 years and over: a randomised controlled trial testing the feasibility and effectiveness of a generic stepped care programme among elderly community residents at high risk of developing anxiety and depression versus usual care [ISRCTN26474556]. BMC Public Health, 6. https://doi.org/10.1186/1471-2458-6-186 General rights C i h d General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. ral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright own n of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain You may freely distribute the URL identifying the publication in the public portal • Users may download and print one copy of any publication from the public portal for the purpose of private study • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may not further distribute the material or use it for any profit making activity or • You may freely distribute the URL identifying the publication in the public portal Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. document breaches copyright please contact us providing details, and we will remove access to the work immediately aim E-mail address: vuresearchportal.ub@vu.nl E-mail address: vuresearchportal.ub@vu.nl E-mail address: vuresearchportal.ub@vu.nl Download date: 24. Oct. 2024 Download date: 24. Oct. 2024 BioMed Central y p Prevention of anxiety and depression in the age group of 75 years and over: a randomised controlled trial testing the feasibility and effectiveness of a generic stepped care programme among elderly community residents at high risk of developing anxiety and depression versus usual care [ISRCTN26474556] Nelleke van 't Veer-Tazelaar1,2, Harm van Marwijk1,2, Patricia van Oppen1,2,3, Giel Nijpels1,2, Hein van Hout1,2, Pim Cuijpers3,4, Wim Stalman1,2 and Aartjan Beekman1,2,3 Prevention of anxiety and depression in the age group of 75 years and over: a randomised controlled trial testing the feasibility and effectiveness of a generic stepped care programme among elderly community residents at high risk of developing anxiety and depression versus usual care [ISRCTN26474556] Nelleke van 't Veer-Tazelaar1,2, Harm van Marwijk1,2, Patricia van Oppen1,2,3, Giel Nijpels1,2, Hein van Hout1,2, Pim Cuijpers3,4, Wim Stalman1,2 and Aartjan Beekman1,2,3 Address: 1Department of General Practice, VU University Medical Centre, Amsterdam, The Netherlands, 2Institute for Research in Extramural Medicine, VU University Medical Centre, Amsterdam, The Netherlands, 3Department of Psychiatry, VU University Medical Centre, Amsterdam, The Netherlands and 4Department of Clinical Psychology, VU University Medical Centre, Amsterdam, The Netherlands Email: Nelleke van 't Veer-Tazelaar* - pj.vantveer@vumc.nl; Harm van Marwijk - hwj.vanmarwijk@vumc.nl; Patricia van Oppen - pvanoppen@ggzba.nl; Giel Nijpels - g.nijpels@vumc.nl; Hein van Hout - hpj.vanhout@vumc.nl; Pim Cuijpers - p.cuijpers@psy.vu.nl; Wim Stalman - w.stalman@vumc.nl; Aartjan Beekman - aartjanb@ggzba.nl * Corresponding author Published: 18 July 2006 BMC Public Health 2006, 6:186 doi:10.1186/1471-2458-6-186 Received: 07 June 2006 Accepted: 18 July 2006 This article is available from: http://www.biomedcentral.com/1471-2458/6/186 © 2006 van 't Veer-Tazelaar et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Received: 07 June 2006 Accepted: 18 July 2006 This article is available from: http://www.biomedcentral.com/1471-2458/6/186 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. BMC Public Health Open Access Methods/design: Design: randomised controlled trial. (See figure 1: organisation chart) together with two other projects, this project is part of a national consortium that investigates the prevention of anxiety and depressive disorders in later life using a stepped care programme. The three projects have their own particular focus. This project is aimed at elderly living in the community. Inclusion: subjects with a high risk for depression and anxiety without clinical evidence of these syndromes. The participants are 75 years of age and over and have subthreshold symptoms of depression and or anxiety: they score above the cut-off point on the self-report Centre for Epidemiologic Studies Depression (CES-D) scale, but the criteria for a major depressive disorder or anxiety disorder (panic disorder, agoraphobia, social phobia, generalized anxiety disorder) according to a validated interview, the Mini International Neuropsychiatric Interview (MINI) are not fulfilled. Abstract Background: In frail elderly, the effects of depression and anxiety are deep encroaching. Indicated prevention studies, aimed at subjects with subthreshold disorder, have shown that well designed interventions are capable of reducing the incidence of depression and anxiety. In this randomised prevention trial for elderly, living in the community and suffering from subthreshold depression and anxiety, a stepped care programme was put together to be tested versus usual (GP) care. Methods/design: Design: randomised controlled trial. Methods/design: Design: randomised controlled trial. Outcomes: Outcomes: primary outcome: incidence of a depressive or anxiety disorder over a period of two years (MINI); Page 1 of 8 (page number not for citation purposes) BMC Public Health 2006, 6:186 http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 secondary outcome: a positive influence of the intervention, a stepped care programme, on symptoms of depression and anxiety and on quality of life as assessed with the CES D, the HADS A and the SF36 respectively (i.e. stabilisation or improvement of symptoms) [see table 1]. secondary outcome: a positive influence of the intervention, a stepped care programme, on symptoms of depression and anxiety and on quality of life as assessed with the CES D, the HADS A and the SF36 respectively (i.e. stabilisation or improvement of symptoms) [see table 1]. Take place at baseline and at 3, 6, 9, 12, 18 and 24 months. Trained independent evaluators assess depression and anxiety status, the primary end point (6, 12, 18, 24 months) [see table 2]. Discussion: Late-life depression and anxiety are characterised by high prevalence, unfavourable prognosis, reduced quality of life, excess mortality and substantial societal costs. No health service, however well equipped, will be able to effectively treat all elderly with depression and anxiety. Therefore, development of (cost) effective means to prevent these disorders is very important. Page 2 of 8 (page number not for citation purposes) http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 BMC Public Health 2006, 6:186 ness and efficiency of decisions about allocation of resources in therapy. Not all elderly with subthreshold anxiety or depression need the same type and intensity of preventive intervention. Frequently, complaints will dis- appear without active intervention. Some may be helped by reading a self-help book or watching an instructional video. Others could benefit from a brief psycho-educa- tional group conducted by a paraprofessional or problem solving treatment, and still others may require a form of pharmacotherapy or more intensive individual psycho- logical treatment. Although stepped care seems a logical approach from the clinical perspective, surprisingly few studies have actually examined the effects of stepped care programmes. Flowchart Figure 2 Flowchart piko D patients with a CESD score ≥ 16 Assessment for eligibility: (n= ) and approach for study participation CESD assessment (n= ) MINI assessment (n= ) Randomisation (n= ) Allocation to intervention (n= ) Allocation to control group (n= ) Refusal to participate (n= ) Not meeting inclusion criteria (n= ) Refusal to participate (n= ) Not meeting inclusion criteria (n= ) The proposed generic stepped care programme is based on the following four assumptions: (a) empowerment; participants with subthreshold anxiety and depression need personalised, repeated education, counselling and confrontation with their symptom levels to be able to acknowledge them, (b) different persons require different levels of preventive activities; (c) determining the right level of preventive intervention is critically dependent on proactively monitoring outcome; and (d) stepping up from lower, less intensive to higher, more intensive levels of preventive activities based on monitored outcomes may increase effectiveness and lower costs overall. Flowchart Figure 2 Flowchart Figure 2 Flowchart Flowchart Study design (See figure 2: flow chart) The study is a prospective ran- domised two-armed intervention study. Subjects are ran- domised by an independent statistician and assigned to intervention (stepped care programme) or control group (care as usual). This project aims at the latter; the design and testing of a stepped care intervention to prevent the onset of full blown anxiety and depressive disorder among elderly people with subthreshold anxiety or depression disorder. Methods/design Aim of the study tionally, the aims of indicated preventive interventions might be to shorten the duration of the persistence of the early symptoms and to stop the progression of severity so that the subject does not meet DSM IV diagnostic levels [11]. Indicated prevention studies have shown that well designed interventions are capable of reducing the inci- dence of depression and anxiety [12]. Moreover, they are likely to be more cost-effective than alternative approaches [13]. So, favourable prevention should be well designed and should be aimed at selected groups of elderly, at high risk of developing anxiety or depression. Selected risk factors for the onset of late life depression are [10]: loss of a partner, chronic illness, neuroticism, family history, lack of social support and subthreshold anxiety or depression disorders. a) To put together an indicated prevention intervention for participants 75 years of age and over who live in the community and suffer from subthreshold anxiety or depression (but have no evidence of the clinical disor- ders) for use by home care in collaboration with mental health care, and b) To evaluate the effects of this programme versus usual general practitioner (GP) care in the prevention of depres- sive or anxiety disorders. Background depression or anxiety will develop a major depressive or anxiety disorder in three years [8]. Therefore, preventing the onset or development of these disorders has a high priority [9,10]. Three types of prevention can be distin- guished; universal prevention (aimed at the general pop- ulation regardless of risk status), selective prevention (aimed at high-risk groups), and indicated prevention (aimed at subjects who have early symptoms of a disorder but do not meet diagnostic criteria). The aims of these types of preventive intervention are reduction of occur- rence of new cases and the delay of onset of illness. Addi- g Depression and anxiety lead to a serious impairment of daily functioning and quality of life. In frail elderly, the effects of depression and anxiety are especially deep encroaching. Besides a deleterious effect on daily func- tioning and quality of life, a large number of studies dem- onstrate excess mortality, disability, handicap and service utilisation [1-7]. The number of elderly is rapidly growing. Almost a third of elderly subjects in the community with subthreshold Organisation chart Figure 1 Organisation chart Consortium / nationally Title: Prevention of anxiety and depression in later life: A programme testing the feasibility and effectiveness of a generic stepped-care programme among elderly at high risk of developing anxiety and depression PIKO / locally (Amsterdam) “Primary Care” PIKO = Preventive Intervention Frail Elderly PIKO D (depression and anxiety) PIKO M (mantlecare / dementia) PIKO (frail) project 2 Prevention of anxiety and depression in later life focus on “co-morbidity” Groningen project 1 Prevention of anxiety and depression in later life focus on “primary care” Amsterdam project 3 Prevention of anxiety and depression in later life focus on “recurrent depression” Leiden Title: Prevention of anxiety and depression in later life: A programme testing the feasibility and effectiveness of a generic stepped-care programme among elderly at high risk of developing anxiety and depression PIKO / locally (Amsterdam) “Primary Care” Page 2 of 8 (page number not for citation purposes) http://www.biomedcentral.com/1471-2458/6/186 Page 3 of 8 (page number not for citation purposes) Participants Adult persons, aged 75 years or older, with subthreshold anxiety or depression capable to give informed consent and with sufficient knowledge of the Dutch language are eligible. Subjects meeting criteria for major depression and/or clinical anxiety, subjects with insufficient mastery of the Dutch language and subjects unwilling or unable to The chosen intervention is a stepped care programme [14- 16]. In an environment of limited resources, it makes sense to provide expertise, and all the time and individual attention a patient needs, but not more. Stepped care models represent endeavours to maximize the effective- Page 3 of 8 (page number not for citation purposes) Page 3 of 8 (page number not for citation purposes) BMC Public Health 2006, 6:186 http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 -reduction of depressive and/or anxiety symptoms as measured with the CES-D and the HADS-A give informed consent are excluded. A subject is defined to have subthreshold depression and/or anxiety when he or she scores above the cut-off score of 16 or more on the Centre for Epidemiologic Studies Depression scale (CES- D, a selfrating inventory for depression and/or anxiety), but does not meet criteria for a full-blown depressive or anxiety disorder as assessed with the Mini International Neuropsychiatric Interview (MINI) [17]. -improvement of quality of life (MOS-SF-36) -mortality -mortality MINI diagnostic interview [17,20] MINI diagnostic interview [17,20] The Mini International Neuropsychiatric Interview (M.I.N.I.) is a short structured diagnostic interview devel- oped in 1990 by psychiatrists and clinicians in the United States and Europe for DSM-IV and ICD-10 psychiatric dis- orders. With an administration time of approximately 20 minutes, the MINI has become the structured interview of choice for psychiatric evaluation and outcome tracking in clinical psychopharmacology trials and epidemiological studies. The MINI interviews in this project entail the following modules: depressive disorder, dysthymia, panic disorder, agoraphobia, social phobia and generalized anxiety disor- der. For this project we used MiniManager 2.0 (created by E.de Beurs, Leiden University Medical Centre), an elec- tronic version of the Dutch MINI interview (5.0). Both the intervention and control participants will receive blinded independent assessments by telephone of depression and anxiety status at baseline, 6, 12, 18 and 24 months. Instruments h i The instruments used in this study are frequently applied in international studies and well validated. Data collection/settings and locations The data are collected from subjects who live independ- ently on their own in West-Friesland (North-Western part of the Netherlands). Outcome measures The primary outcome is the incidence of major depressive or anxiety (panic disorder, agoraphobia, social phobia and generalised anxiety) disorder (after two years) as measured with the Mini International Neuropsychiatric Interview (MINI) CES-D [21-24] The Centre for Epidemiologic Studies Depression scale (CES-D) will be used for the screening of subthreshold depression and anxiety. It consists of 20 items and its total score has a range between 0 and 60. Scores = 16 indicate clinically significant levels of depressive symptoms. At this cutoff the sensitivity is 100% and the specificity is 88% for major depressive disorder in the elderly Dutch population [22]. The CES-D has also been found to be a satisfactory screener for anxiety disorders [23]. However, as the CES- D was designed specifically for the screening of depression and as the criterion validity for depression was considera- bly better than for anxiety disorders, for follow-up pur- poses, the CES-D will be combined with additional items from the Hospital Anxiety and Depression scale – anxiety section (HADS-A) [25]. The aim is to optimise the sensi- tivity and specificity for both subthreshold depression and anxiety. Recruitment Eligible subjects for the present study are identified among the study population of the larger project for frail elderly -the PIKO project [19] in which this study is embedded. This PIKO project is based on the results of a self-rating health inventory, including the Centre for Epi- demiologic Studies Depression scale (CES-D). The inven- tory has been completed by general practice patients of 75 years or older (see organisation chart). Subjects in this database with a CES-D score above the cut-off point of 16 are approached regarding compliance for the present study. Sample size Prospective data derived from the Longitudinal Ageing Study Amsterdam (LASA) study show that 27% of elderly subjects in the community with subthreshold depression develop a major depressive disorder within three years [8]. We assume that this percentage is comparable for sub- threshold anxiety, and for combined anxiety and depres- sion. The expected incidence rate of DSM-IV depressive or anxiety disorder within two years is estimated conserva- tively at 35%. Based on previous work in indicated pre- vention [12,18], we expect the stepped care programme to reduce the incidence rate of major depression and/or anx- iety disorder from 35% to 20%. We aim for 110 partici- pants in each arm (α = 0.05 and a power of (1-β) = 0.80). If after final inclusion the total of participants unexpect- edly falls short of the required number to reach a power of 0.8 in our calculations, pooling with the two other projects (see organisation chart) will be considered. HADS [25] The Hospital Anxiety and Depression Scale has been found to perform well in assessing the symptom severity and caseness of anxiety disorders and depression in both Secondary outcomes are Page 4 of 8 (page number not for citation purposes) BMC Public Health 2006, 6:186 http://www.biomedcentral.com/1471-2458/6/186 Table 1: Instruments used to establish primary and secondary outcomes Outcome measures Instrument Primary outcome depressive/anxiety disorder MINI diagnostic interview Secondary outcome reduction of depressive symptoms CES-D reduction of anxiety symptoms HADS-A quality of life MOS-SF-36 health care utilisation TIC-P patient satisfaction GGZ thermometer Table 1: Instruments used to establish primary and secondary outcomes months (T5) and then at eighteen = follow-up1 (T6) and twenty-four = follow-up2 (T7) months. See table 2 Institution of Mental Health – thermometer (Geestelijke GezondheidsZorg thermometer) Institution of Mental Health – thermometer (Geestelijke GezondheidsZorg thermometer) An instrument to measure patient treatment satisfaction. It focuses on the appreciation of treatment explanation, the social worker and the result of the coaching. table 1 displays the outcome parameters, the above men- tioned interview and instruments and additionally used questionnaires Step 2, after randomisation. When spontaneous recovery does not occur, participants with subthreshold illness receive mini- SF (Short Form) 36 – quality of life [26] The SF36 is the short form questionnaire to measure qual- ity of life. It was used in the original health inventory of the 'embedding' PIKO study. The stepped care programme as applied in this study entails: TiC-P health care utilisation [27] Trimbos/iMTA questionnaire for Costs associated with Psychiatric Illness (TiC-P). This questionnaire will be applied to monitor costs. It is developed by the Trimbos Institute Utrecht in combination with the institute for Medical Technology Assessment Rotterdam Step 1: Elderly with subthreshold anxiety or depression are actively followed-up by watchful waiting whether participants recover spontaneously. Subjects with a previously high CESD-score are invited to complete a second CESD questionnaire. The interval between the two CESD measurements is at least three months. In case of another score higher than the cut-off score, the MINI diagnostic interview takes place. When the result of the interview is negative (no symptoms of depressive/anxiety disorder during the past year) and the additional inclusion criteria are met, subjects are ran- domised. Intervention/the stepped care approach The stepped care approach in this project is based on sev- eral protocols with known effectiveness [14-16] and dis- cussions with the local homecare agency, regional mental health organisation and other careproviders such as GPs. Two interventions with a strong emphasis on psycho-edu- cation and the acquisition of skills with respect to the pre- vention of depression will be tested. First; the group course "Coping with Depression" (CWD) [28,28-30] which is revised for anxiety and adapted for this specific population (75 years of age and older) and individual use and second; Problem Solving Treatment, which is a brief cognitive behavioral therapy (PST) [31]. Both interven- tion protocols have come about in collaboration with the Trimbos Institute (the Netherlands) and the Mynors-Wal- lis (UK) group respectively. A valuable contribution to evidence-based treatments in the elderly might be consti- tuted by the present study in the production of empirically supported treatment protocols. somatic, psychiatric and primary care patients and in the general population. Of the HADS, in this project only the seven anxiety items will be used to complement the anxi- ety-screening ability of the CES-D questionnaire. SF (Short Form) 36 – quality of life [26] SF (Short Form) 36 – quality of life [26] Discussion Late-life depression and anxiety are characterised by high prevalence, unfavourable prognosis, reduced quality of life, excess mortality and substantial societal costs [13]. Fortunately, there are hopeful indications that the preven- tion of new cases of mental disorders seems to be possible [12]. Subthreshold anxiety and depression (i.e., the pres- ence of symptoms of anxiety or depression without evi- dence of the actual psychiatric disorder) are prognostic variables for major anxiety disorder and depression [13]. Interventions in the subthreshold disorders may prevent the onset of new cases of major depression and anxiety disorder [18]. Demonstration projects of shared care models for indicated prevention of new cases of mental disorders may be an important step forward to reduce the enormous burden of these disorders as was shown by a recent meta-analysis of randomised trials of preventive interventions [12] In this study the indicated prevention of depression and anxiety is attempted by way of a stepped-care programme offering the aged subthreshold depressed or anxious participant, several interventions ranging from noncommittal to the requirement of some engagement. During the second visit the folder will be briefly discussed and a reader, the self-help course "coping with depres- sion" will be supplied to the client. Subjects are stimu- lated to do the course in their own tempo. The nurse will regularly ring or visit the client to discuss progress and questions. Obviously to read both folder and reader and make the exercises offered in the latter, is not obligatory. Subjects are free to read the information and make the exercises. An important aspect of the role of the nurse is advice and encouragement. After three months, at the end of the first phase an evaluation form is completed by the nurse. Step 3. When symptoms persist, participants are introduced to a brief cognitive behavioural therapy intervention (Problem Solving Treatment) A preliminary reflection on the limitations and strengths of our design: If -after three months- there is still a CES-D score of 16 or higher, subjects will receive a telehone call in which is explained to them that they now qualify for Problem Solv- ing Treatment. The method and approach are explained to them. The short-term Problem Solving Treatment is given by specially trained Community Psychiatric Nurses, who will also complete an evaluation form at the end of the maximum 7 sessions. Assessments Assessments take place at baseline (T0), start of the pro- gramme (T1), at three (T2), six (T3), nine (T4) and twelve Table 2: Assessments health inventory T0 T1-a T1-b T1-c T2 3 mnths T3 6 mnths T4 9 mnths T5 12 mnths T6 18 mnths T7 24 mnths CES-D x x - - x x x x x x MINI - x - - x - x x x extra identification information - - x - - - - - - HADS-A - - x - x - x x x SF36 - - x - - - x - x TIC-P health care utilisation - - x - x - x x x pt.satisfaction - - x x x x x x x Table 2: Assessments Page 5 of 8 (page number not for citation purposes) Table 2: Assessments health inventory T0 T1-a T1-b T1-c T2 3 mnths T3 6 mnths T4 9 mnths T5 12 mnths T6 18 mnths T7 24 mnths CES-D x x - - x x x x x x MINI - x - - x - x x x extra identification information - - x - - - - - - HADS-A - - x - x - x x x SF36 - - x - - - x - x TIC-P health care utilisation - - x - x - x x x pt.satisfaction - - x x x x x x x http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 BMC Public Health 2006, 6:186 mal interventions (folder, self-help book "Coping with Depres- sion") This step is coached by home care nurses. step. A result below the cut-off score means; 'wait and see' for three months, after which a new evaluation takes place. At that point, again, a CES-D score of 16 or higher means that participation in the next step is offered after all. So, in the programme, postponement or cancellation of steps is possible. When a major depression or an anxi- ety disorder develops, subjects are referred back to their GP. Subjects receive a telephone call with explanation about the intervention. After that, they receive the first question- naire together with again -but this time; written- informa- tion. Shortly afterwards subjects are seen by a specially trained home care nurse who hands over a folder with information about and tips how to deal with subthresh- old anxiety and/or depression. Assessments The aims of this first visit are: a) to assess type, the level of severity and causes of symptoms, b) to reflect on emotional symptoms through an initial brief nondirectional exploration and discussion of everyday problems, which also helps to establish a working relationship. c) to educate participants about their symptoms. The nurse makes an appointment for the next visit. Discussion A limitation of this pragmatic design is that it will be dif- ficult to weigh the specific contributions of the various elements of this project. Another limitation may be that the quantity and type of questions (questions about mood may be confronting and depressing on their own) can lead to drop-out. The stimulating ideas and activities in the self-help course trying to achieve a positive and active attitude may have an opposite effect since the reader might instead be confronted with his/her limita- tions. Step 4. At this final step, participants who still suffer from sub- threshold depression or anxiety, are referred to their GP to dis- cuss the appropriateness of specific medication. Timeframe of the study Timeframe of the study See table 2 See table 2 Strong aspects of this design: In case of a continuously elevated CES-D score, subjects will receive written advice to discuss suitable medication with their GP (i.e., antidepressants). - It is a unique practice-based project in which several evi- dence-based aspects of for instance the highly successful IMPACT treatment study (Improving Mood: Promoting Access to Collaborative Care Treatment) [16] are incorpo- rated, now for indicated prevention purposes and among a very old population. Where IMPACT provides its partic- Summarising, the CES-D score is checked every three months during a year. Subjects with a result at or above the cut-off score of 16 are offered participation in the next Page 6 of 8 (page number not for citation purposes) BMC Public Health 2006, 6:186 http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 Description of risks g This study is financed by ZonMw – the Netherlands Organisation for Health Research and Development, grant-number 2620.00003 g This study is financed by ZonMw – the Netherlands Organisation for Health Research and Development, grant-number 2620.00003 g This study is financed by ZonMw – the Netherlands Organisation for Health Research and Development, grant-number 2620.00003 To our knowledge, serious risks or undesired effects of completing questionnaires are not reported in the litera- ture. There are no specific risks related to this study. Edwin de Beurs gave permission to use MiniManager, his electronic version of the Dutch MINI diagnostic interview. Data security ipants of 60 years of age or older with education, brief psy- chotherapy, medication and stepped care while outcomes and progress are regularly and closely monitored by a care manager, in this Dutch project the participants are older but very similar items compared to the IMPACT model are employed. However, the focus now is on symptoms and on indicated prevention, not on treatment of persons who drew up a specific treatment contract with their physician. Confidential information and participant names are secured by the medical confidentiality rules and are treated according to the code of conduct for medical research, developed by the FMWV (the Federation of Bio- medical Scientific Societies). The results of the participant questionnaires are not acces- sible to the General Practitioners. All study related docu- ments and data are stored on a protected central server of the Institute for Research in Extramural Medicine, VU University Medical Centre, Amsterdam, the Netherlands. Only members of the study have access to the respective files. - A further strength of our study is that it encompasses anxiety: that even less studies among the elderly have looked at indicated prevention of anxiety [12,32], and at co-morbidity of anxiety and depression [12]. There are several community studies which show that anxiety is as important a problem for elderly as depression [2,3]. Ethical principles The group course "Coping with Depression" (CWD) was revised for anx- iety and adapted for 75+ readers and individual use by the Trimbos Insti- tute. The group course "Coping with Depression" (CWD) was revised for anx- iety and adapted for 75+ readers and individual use by the Trimbos Insti- tute. The participation in the study is voluntarily. Participants are informed that they can cancel their participation at any time without disclosing reasons for their cancellation and without negative consequences for their future medi- cal care. Competing interests The third strength is that we will also be able to reflect on the ethical dilemma of whether or not to include persons in programmes such as these who are at risk (i.e., not par- ticipants) but have no explicit request for care. The per- centage uptake of the intervention in the various stepped care stages and the participants' preferences will show us whether an intervention aimed at symptoms and not at syndromes connects to the reality of distressed old people in the community. The questions would be whether these elderly are willing to acknowledge distress and to take active steps to feel better. The author(s) declare that they have no competing inter- ests. Authors' contributions NvtV, HvM, PvO perform the study NvtV, HvM, PvO perform the study HvM participated in the draft manuscript of the proposal, the study design AB and PC participated in the draft manuscript and the conceiving of this study Strengths and limitations taken into account, the develop- ment and research of (cost) effective means to prevent depressive and anxiety disorder in the rapidly growing group of elderly is a matter of the greatest importance. Timeframe of the study See table 2 Page 7 of 8 (page number not for citation purposes) Vote of the ethics committee Lewinsohn PM: The Coping with Depression course: A psychoeducational intervention for unipolar depression. 1984. p p 29. Cuijpers P: Bibliotherapy in unipolar depression: a meta-anal- ysis. J Behav Ther Exp Psychiatry 1997, 28:139-147. 6. Katon WJ, Lin E, Russo J, Unutzer J: Increased medical costs of a population-based sample of depressed elderly patients. Arch Gen Psychiatry 2003, 60:897-903. 30. Haringsma R, Engels GI, Cuijpers P, Spinhoven P: Effectiveness of the Coping With Depression (CWD) course for older adults provided by the community-based mental health care sys- tem in the Netherlands: a randomized controlled field trial. Int Psychogeriatr 2005:1-19. y y 7. Cuijpers P: Mortality and depressive symptoms in inhabitants of residential homes. Int J Geriatr Psychiatry 2001, 16:131-138. 8. Beekman AT, Geerlings SW, Deeg DJ, Smit JH, Schoevers RS, de Beurs E, Braam AW, Penninx BW, van Tilburg W: The natural his- tory of late-life depression: a 6-year prospective study in the community. Arch Gen Psychiatry 2002, 59:605-611. y g 31. Mynors-Wallis L, Davies I, Gray A, Barbour F, Gath D: A ran- domised controlled trial and cost analysis of problem-solving treatment for emotional disorders given by community nurses in primary care. Br J Psychiatry 1997, 170:113-119. y y y 9. Beekman AT, Deeg DJ, Geerlings SW, Schoevers RA, Smit JH, van Tilburg W: Emergence and persistence of late life depression: a 3-year follow-up of the Longitudinal Aging Study Amster- dam. J Affect Disord 2001, 65:131-138. p y J y y 32. Cuijpers P: Examining the effects of prevention programs on the incidence of new cases of mental disorders: the lack of statistical power. Am J Psychiatry 2003, 160:1385-1391. J 10. de Beurs E, Beekman A, Geerlings S, Deeg D, van Dyck R, van Tilburg W: On becoming depressed or anxious in late life: similar vul- nerability factors but different effects of stressful life events. Br J Psychiatry 2001, 179:426-431. Vote of the ethics committee 1. Beekman AT, de Beurs E, van Balkom AJ, Deeg DJ, van Dyck R, van Tilburg W: Anxiety and depression in later life: Co-occurrence and communality of risk factors. Am J Psychiatry 2000, 157:89-95. 2. van Hout HP, Beekman AT, de Beurs E, Comijs H, van Marwijk H, de Haan M, van Tilburg W, Deeg DJ: Anxiety and the risk of death in older men and women. Br J Psychiatry 2004, 185:399-404. 3. de Beurs E, Beekman AT, van Balkom AJ, Deeg DJ, van Dyck R, van Tilburg W: Consequences of anxiety in older persons: its 1. Beekman AT, de Beurs E, van Balkom AJ, Deeg DJ, van Dyck R, van Tilburg W: Anxiety and depression in later life: Co-occurrence and communality of risk factors. Am J Psychiatry 2000, 157:89-95. The design and conduct of the study was approved by the ethics committee of VU University Medical Centre, Amsterdam. y J y y 2. van Hout HP, Beekman AT, de Beurs E, Comijs H, van Marwijk H, de Haan M, van Tilburg W, Deeg DJ: Anxiety and the risk of death in older men and women. Br J Psychiatry 2004, 185:399-404. J y y 3. de Beurs E, Beekman AT, van Balkom AJ, Deeg DJ, van Dyck R, van Tilburg W: Consequences of anxiety in older persons: its Page 7 of 8 (page number not for citation purposes) Page 7 of 8 (page number not for citation purposes) http://www.biomedcentral.com/1471-2458/6/186 http://www.biomedcentral.com/1471-2458/6/186 BMC Public Health 2006, 6:186 25. Zigmond AS, Snaith RP: The hospital anxiety and depression scale. Acta Psychiatr Scand 1983, 67:361-370. effect on disability, well-being and use of health services. Psy- chol Med 1999, 29:583-593. y 26. Ware JJ, Sherbourne CD: The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Medical Care 1992:473-483. 4. Schoevers RA, Deeg DJ, van Tilburg W, Beekman AT: Depression and generalized anxiety disorder: co-occurrence and longi- tudinal patterns in elderly patients. Am J Geriatr Psychiatry 2005, 13:31-39. 27. Hakkaart-van Roijen L, Van Straten A, Donker M, Tiemens B: Trim- bos/iMTA questionnaire for Costs associated with Psychiat- ric Illness (TiC-P). 2002. 5. Beekman AT, Penninx BW, Deeg DJ, de Beurs E, Geerling SW, van Tilburg W: The impact of depression on the well-being, disa- bility and use of services in older adults: a longitudinal per- spective. Acta Psychiatr Scand 2002, 105:20-27. ( ) 28. Pre-publication history The pre-publication history for this paper can be accessed here: The pre-publication history for this paper can be accessed here: J y y 11. Haggerty M, Medicine I: Reducing risks for mental disorders Washinton DC, National Academy Press; 1994. 12. Cuijpers P, Van Straten A, Smit F: Preventing the incidence of new cases of mental disorders: a meta-analytic review. J Nerv Ment Dis 2005, 193:119-125. http://www.biomedcentral.com/1471-2458/6/186/pre pub 13. Smit F, Ederveen A, Cuijpers P, Deeg D, Beekman A: Opportunities for Cost-effective Prevention of Late-Life Depression: An Epidemiological Approach. Arch Gen Psychiatry 2006, 63:290-296. p g pp y y 14. Otto MW, Pollack MH, Maki KM: Empirically supported treat- ments for panic disorder: costs, benefits, and stepped care. J Consult Clin Psychol 2000, 68:556-563. y 15. Haaga DA: Introduction to the special section on stepped care models in psychotherapy. J Consult Clin Psychol 2000, 68:547-548. 6 16. Hunkeler EM, Katon W, Tang L, Williams JWJ, Kroenke K, Lin EH, Harpole LH, Arean P, Levine S, Grypma LM, Hargreaves WA, Unut- zer J: Long term outcomes from the IMPACT randomised trial for depressed elderly patients in primary care. BMJ 2006, 332:259-263. 17. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, Hergueta T, Baker R, Dunbar GC: The Mini-International Neu- ropsychiatric Interview (M.I.N.I.): the development and vali- dation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry 1998, 59 Suppl 20:22-33. J y y pp 18. Willemse GR, Smit F, Cuijpers P, Tiemens BG: Minimal-contact psychotherapy for sub-threshold depression in primary care. Randomised trial. Br J Psychiatry 2004, 185:416-421. J y y 19. van Hout HP, Nijpels G, van Marwijk HW, Jansen AP, Van't Veer PJ, Tybout W, Stalman WA: Design and pilot results of a single blind randomized controlled trial of systematic demand-led home visits by nurses to frail elderly persons in primary care [ISRCTN05358495]. BMC Geriatr 2005, 5:11. [ ] 20. Pinninti NR, Madison H, Musser E, Rissmiller D: MINI International Neuropsychiatric Schedule: clinical utility and patient acceptance. Eur Psychiatry 2003, 18:361-364. Pre-publication history Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Page 8 of 8 (page number not for citation purposes) Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Publish with BioMed Central and every scientist can read your work free of charge p y y 21. Radloff LS: The CES-D Scale: a self-report depression scale for research in the general population. J Applied Psychol Measure- ment 1977:385-401. 22. Beekman AT, Deeg DJ, Van Limbeek J, Braam AW, De Vries MZ, van Tilburg W: Criterion validity of the Center for Epidemiologic Studies Depression scale (CES-D): results from a commu- nity-based sample of older subjects in The Netherlands. Psy- chol Med 1997, 27:231-235. 23. Breslau N: Depressive symptoms, major depression, and gen- eralized anxiety: a comparison of self-reports on CES-D and results from diagnostic interviews. Psychiatry Res 1985, 15:219-229. 24. Radloff LS, Teri L: Use of the Center for Epidemiological Stud- ies-Depression Scale with older adults. Clinical Gerontologist 1986, 5:119-136.
https://openalex.org/W4362429131	https://figshare.com/articles/journal_contribution/Supplementary_Figure_2_from_Somatic_Profiling_of_the_Epidermal_Growth_Factor_Receptor_Pathway_in_Tumors_from_Patients_with_Advanced_Colorectal_Cancer_Treated_with_Chemotherapy_Cetuximab/22445664/1/files/39896706.pdf	English	null	Supplementary Figure 3 from Somatic Profiling of the Epidermal Growth Factor Receptor Pathway in Tumors from Patients with Advanced Colorectal Cancer Treated with Chemotherapy ± Cetuximab	null	2,023	cc-by	446	A 0.75 1.00 OS PIK3CA wild type B OS PIK3CA mutant HR (95% CI) = 1.01 (0.88, 1.16) HR (95% CI) = 1.04 (0.72, 1.50) Supplementary Figure 2 0 6 12 18 24 30 36 42 48 0 6 12 18 24 30 36 42 48 Time (months) 0.00 0.25 0.50 Survival Ti ( th ) P=0.84 P=0.83 Time (months) Arm A (OxFp) Arm B (OxFp+cetuximab) 1.00 C PFS PIK3CA wild type D PFS PIK3CA mutant Interaction HR (95% CI) = 1.01 (0.68, 1.49) ; P=0.96 Time (months) 0 0.25 0.50 0.75 Survival HR (95% CI) = 0.99 (0.88, 1.12) P=0.90 HR (95% CI) = 1.01 (0.73, 1.40) P=0.94 0.00 0 6 12 18 24 30 36 42 48 Time (months) 0 6 12 18 24 30 36 42 48 Time (months) Interaction HR (95% CI) = 1.04 (0.73, 1.47) ; P=0.84 A 0.75 1.00 OS PIK3CA wild type B OS PIK3CA mutant HR (95% CI) = 1.01 (0.88, 1.16) HR (95% CI) = 1.04 (0.72, 1.50) Supplementary Figure 2 0 6 12 18 24 30 36 42 48 0 6 12 18 24 30 36 42 48 Time (months) 0.00 0.25 0.50 Survival Ti ( th ) P=0.84 P=0.83 Time (months) Arm A (OxFp) Arm B (OxFp+cetuximab) 1.00 C PFS PIK3CA wild type D PFS PIK3CA mutant Interaction HR (95% CI) = 1.01 (0.68, 1.49) ; P=0.96 Time (months) 0 0.25 0.50 0.75 Survival HR (95% CI) = 0.99 (0.88, 1.12) P=0.90 HR (95% CI) = 1.01 (0.73, 1.40) P=0.94 0.00 0 6 12 18 24 30 36 42 48 Time (months) 0 6 12 18 24 30 36 42 48 Time (months) Interaction HR (95% CI) = 1.04 (0.73, 1.47) ; P=0.84 Supplementary Figure 2 A 0.75 1.00 OS PIK3CA wild type B OS PIK3CA mutant HR (95% CI) = 1.01 (0.88, 1.16) HR (95% CI) = 1.04 (0.72, 1.50) pp y g 0 6 12 18 24 30 36 42 48 0 6 12 18 24 30 36 42 48 Time (months) 0.00 0.25 0.50 Survival Ti ( th ) P=0.84 P=0.83 Time (months) Time (months) A B Survival Interaction HR (95% CI) = 1.01 (0.68, 1.49) ; P=0.96 Arm A (OxFp) Arm B (OxFp+cetuximab) 1.00 C PFS PIK3CA wild type D PFS PIK3CA mutant Interaction HR (95% CI) = 1.01 (0.68, 1.49) ; P=0.96 0 0.25 0.50 0.75 Survival HR (95% CI) = 0.99 (0.88, 1.12) P=0.90 HR (95% CI) = 1.01 (0.73, 1.40) P=0.94 0.00 0 6 12 18 24 30 36 42 48 Time (months) 0 6 12 18 24 30 36 42 48 Time (months) D Survival Time (months) Interaction HR (95% CI) = 1.04 (0.73, 1.47) ; P=0.84
https://openalex.org/W3206585457	https://repository.kulib.kyoto-u.ac.jp/dspace/bitstream/2433/274291/1/s00028-022-00778-7.pdf	English	null	Maximal $$L^1$$-regularity for parabolic initial-boundary value problems with inhomogeneous data	Journal of evolution equations	2,022	cc-by	37,791	J. Evol. Equ. (2022) 22:30 © 2022 The Author(s) https://doi.org/10.1007/s00028-022-00778-7 J. Evol. Equ. (2022) 22:30 © 2022 The Author(s) https://doi.org/10.1007/s00028-022-00778-7 Journal of Evolution Equations Maximal L1-regularity for parabolic initial-boundary value problems with inhomogeneous data Takayoshi Ogawa and Senjo Shimizu Abstract. End-point maximal L1-regularity for parabolic initial-boundary value problems is considered. For the inhomogeneous Dirichlet and Neumann data, maximal L1-regularity for initial-boundary value problems is established in time end-point case upon the homogeneous Besov space ˙Bs p,1(Rn +) with 1 < p < ∞and −1 + 1/p < s ≤0 as well as optimal trace estimates. The main estimates obtained here are sharp in the sense of trace estimates and it is not available by known theory on the class of UMD Banach spaces. We utilize a method of harmonic analysis, in particular, the almost orthogonal properties between the boundary potentials of the Dirichlet and the Neumann boundary data and the Littlewood-Paley dyadic decomposition of unity in the Besov and the Lizorkin–Triebel spaces. Keywords: Parabolic equations with variable coefﬁcients, Maximal L1-regularity, End-point estimate, Initial-boundary value problems, The Dirichlet problem, The Neumann problem. Mathematics Subject Classiﬁcation: Primary 35K20; Secondary 42B25 1 0123456789().: V,-vol 1. Introduction and main results In this paper, we are concerned with maximal L1-regularity for initial-boundary value problems of parabolic equations in the half-space Rn + with inhomogeneous boundary data. Let X beaBanachspaceand A beaclosedlinearoperatorin X withadenselydeﬁned domain D(A). For an initial data u0 ∈X and an external force f ∈Lρ(0, T ; X) (1 ≤ρ ≤∞), let u be a solution to the abstract Cauchy problem: ⎧ ⎨ ⎩ d dt u + Au = f, t > 0, u(0) = u0. (1.1) (1.1) Then, A has maximal Lρ-regularity if there exists a unique solution u of (1.1) such that d dt u, Au ∈Lρ(0, T ; X) satisfy the estimate: Then, A has maximal Lρ-regularity if there exists a unique solution u of (1.1) such that d dt u, Au ∈Lρ(0, T ; X) satisfy the estimate: d dt u Lρ(0,T ;X) + ∥Au∥Lρ(0,T ;X) ≤C ∥u0∥(X,D(A))1−1ρ ,ρ + ∥f ∥Lρ(0,T ;X) , under the restriction u0 ∈(X, D(A))1−1 ρ ,ρ, where (X, D(A))1−1 ρ ,ρ denotes the real interpolation space between X and D(A), and C is a positive constant independent under the restriction u0 ∈(X, D(A))1−1 ρ ,ρ, where (X, D(A))1−1 ρ ,ρ denotes the real interpolation space between X and D(A), and C is a positive constant independent 30 Page 2 of T. Ogawa and S. Shimizu J. Evol. Equ. 67 of u0 and f . Maximal regularity for parabolic equations was ﬁrst considered by Ladyzhenskaya–Solonnikov–Ural’tseva [29]. Then, the research of maximal regular- ity has developed immensely in these last few decades by many authors; [5,8,9,13,16– 19,21,22,25,31,43,44]. In the general framework on Banach spaces X that satisfy the unconditional martingale differences (called as UMD), the general theory of maxi- mal regularity was well established, especially by Amann [2,3], Denk–Hieber–Prüss [14,15], Weis [53] (see also [26,28,40]). On the other hand, maximal regularity on non-UMD Banach spaces, for instance, non-reﬂexive Banach space such as L1 or L∞requires independent arguments. For example, one can observe some results for maximal L1-regularity for the Cauchy problem on the homogenous Banach spaces in Chemin [7]. Danchin [10,11], Giga– Saal [20], Iwabuchi [23], Ogawa–Shimizu [33,34] in various non-UMD settings. In general, maximal time L1-regularity fails over the Lebesgue spaces in spatial variables and we need to introduce restrictive function classes such as homogeneous and inho- mogeneous Besov spaces for the spatial variables. 1. Introduction and main results Hence, maximal L1-regularity for the initial boundary value problems is not well established in the general framework. 1.1. The Dirichlet boundary condition Then, the problem (1.2) admits a unique solution u ∈W 1,ρ I; L p(Rn +) ∩Lρ I; W 2,p(Rn +) if and only if f ∈Lρ I; L p(Rn +) , u0 ∈B2(1−1/ρ) p,ρ (Rn +), f ∈Lρ I; L p(Rn +) , u0 ∈B2(1−1/ρ) p,ρ (Rn +), g ∈F1−1/2p ρ,p I; L p(Rn−1) ∩Lρ I; B2−1/p p,p (Rn−1) , if 1 −1/(2p) > 1/ρ, then u0(x′, xn)\|xn=0 = g(t, x′)\|t=0. (1.3) f ∈Lρ I; L p(Rn +) , u0 ∈B2(1−1/ρ) p,ρ (Rn +), g ∈F1−1/2p ρ,p I; L p(Rn−1) ∩Lρ I; B2−1/p p,p (Rn−1) , if 1 −1/(2p) > 1/ρ, then u0(x′, xn)\|xn=0 = g(t, x′)\|t= f ∈L I; L (R+) , u0 ∈Bp,ρ (R+), g ∈F1−1/2p ρ,p I; L p(Rn−1) ∩Lρ I; B2−1/p p,p (Rn−1) , if 1 −1/(2p) > 1/ρ then u0(x′ x )\| 0 = g(t x′)\|t g ∈F1−1/2p ρ,p I; L p(Rn−1) ∩Lρ I; B2−1/p p,p (Rn−1) , (1.3) Besides there exists a constant CT > 0 depending on n, p, ρ, {ai j}, T such that the solution u is subject to the inequality: Besides there exists a constant CT > 0 depending on n, p, ρ, {ai j}, T such that the solution u is subject to the inequality: ∥∂tu∥Lρ(I;L p(Rn +)) + ∥∇2u∥Lρ(I;L p(Rn +)) ∥∂tu∥Lρ(I;L p(Rn +)) + ∥∇2u∥Lρ(I;L p(Rn +)) ∥∂tu∥Lρ(I;L p(Rn +)) + ∥∇2u∥Lρ(I;L p(Rn +)) ≤CT ∥u0∥B2(1−1/2p) p,ρ (Rn +) + ∥f ∥Lρ(I;L p(Rn +)) + ∥g∥F1−1/2p ρ,p (I;L p(Rn−1)) + ∥g∥Lρ(I;B2−1/p p,p (Rn−1)) , ≤CT ∥u0∥B2(1−1/2p) p,ρ (Rn +) + ∥f ∥Lρ(I;L p(Rn +)) + ∥g∥F1−1/2p ρ,p (I;L p(Rn−1)) ≤CT ∥u0∥B2(1−1/2p) p,ρ (Rn +) + ∥f ∥Lρ(I;L p(Rn +)) + ∥g∥F1−1/2p ρ,p (I;L p(Rn−1)) ∥∥ + ∥g∥Lρ(I;B2−1/p p,p (Rn−1)) , where\|∇2u\| = ( 1≤i, j≤n \|∂i∂ju\|2)1/2, Lρ(I; X)denotestheρ-thpoweredLebesgue- Bochner space upon a Banach space X and B2−1/p p,p (Rn−1) and F1−1/2p ρ,p (I; X) denote the interpolation spaces of the Besov and the Lizorkin–Triebel type, respectively. where\|∇2u\| = ( 1≤i, j≤n \|∂i∂ju\|2)1/2, Lρ(I; X)denotestheρ-thpoweredLebesgue- Bochner space upon a Banach space X and B2−1/p p,p (Rn−1) and F1−1/2p ρ,p (I; X) denote the interpolation spaces of the Besov and the Lizorkin–Triebel type, respectively. where\|∇2u\| = ( 1≤i, j≤n \|∂i∂ju\|2)1/2, Lρ(I; X)denotestheρ-thpoweredLebesgue- Bochner space upon a Banach space X and B2−1/p p,p (Rn−1) and F1−1/2p ρ,p (I; X) denote the interpolation spaces of the Besov and the Lizorkin–Triebel type, respectively. 1.1. The Dirichlet boundary condition 1.1. The Dirichlet boundary condition We ﬁrst recall non-endpoint maximal regularity for the initial boundary value prob- lems to parabolic equations. Let I = (0, T ) with 0 < T ≤∞. Let u be a solution of the initial-boundary value problem of the second-order parabolic equation with variable coefﬁcients and the inhomogeneous Dirichlet boundary condition in the half-space Rn + = {x = (x′, xn); x′ ∈Rn−1, xn > 0}: ⎧ ⎪⎪⎪⎪⎨ ⎪⎪⎪⎪⎩ ∂tu − 1≤i, j≤n ai j(t, x)∂i∂ju = f, t ∈I, x ∈Rn +, u(t, x′, xn) xn=0 = g(t, x′), t ∈I, x′ ∈Rn−1, u(t, x) t=0 = u0(x), x ∈Rn +, (1.2) where ∂t and ∂i ≡∂xi are partial derivatives with respect to t and xi, u = u(t, x) denotes the unknown function, u0 = u0(x), f = f (t, x) and g = g(t, x′) are given initial, external force and boundary data, respectively. The coefﬁcient matrix {ai j(t, x)}1≤i, j≤n satisﬁes uniformly elliptic condition and have enough regularity. Namely {ai j} is a real-valued symmetric matrix such that for some constant c > 0, 1≤i j≤n ai j(t, x)ξiξ j ≥c\|ξ\|2 for all ξ ∈Rn with sufﬁcient regularity in both t and x. For 1 ≤ρ ≤∞and a Banach space X, we denote the Lebesgue-Bochner space Lρ(I; X) and the inhomogeneous and homogeneous Sobolev-Bochner spaces as W 1,ρ(I; X), ˙W 1,ρ(I; X), respectively. We denote a set of all continuous bounded X-valued functions over an interval I by Cb(I; X) and on ⊂Rn by Cb(). We also denote a set of all X-valued functions which is bounded uniformly continuous over an interval I by BUC(I; X) and on ⊂Rn by BUC(). J. Evol. Equ. Page 3 of 67 30 Maximal L1-regularity for parabolic Page 3 of 67 30 Maximal L1-regularity for parabolic Page 3 of 67 30 In this context, the following results were obtained by Weidemaier [51] and Denk– Hieber–Prüss [15]. Proposition 1.1 (The Dirichlet boundary condition [15,51]). Let 1 < ρ, p < ∞with 1 −1/(2p) ̸= 1/ρ, I = (0, T ) with T < ∞. 1.1. The Dirichlet boundary condition Weidemaier [49] ﬁrst obtained a trace theorem for functions in anisotropic Sobolev spaces. Then, he extended his result to a boundary trace of a solution of parabolic equations in the Bochner space and obtained the optimal trace estimates ([50–52]) with introducing the Lizorkin–Triebel space in the time variable. In the proof of the results, he employed a solution formula with respect to the time variable and the proof is involved the maximal function for a test function. The results in [48–52] are obtained under the restriction of exponents of the Bochner spaces to space p and time ρ variables as 3 2 < p ≤ρ < ∞of for the Dirichlet problem and 3 < p ≤ρ < ∞ for the Neumann problem, respectively. Denk–Hieber–Prüss [15] obtained the above necessary and sufﬁcient condition of unique existence of solutions to initial-boundary value problems including higher order parabolic operators subject to general boundary conditions in a domain in Rn with a compact boundary. Their ingenious idea is regarding the problem as an evolution equation with respect to the spatial variable xn and the boundary condition as an initial data. However, the proof in [15] is based on the vector valued version of Mikhlin’s Fourier multiplier theorem, and accordingly the result is restricted in the cases 1 < ρ < ∞. Their result is essentially in a time local estimate because the elliptic operator considered there has strictly positive spectrum, and hence, the boundary conditions are limited in the inhomogeneous real interpolation spaces. In this paper, we show time global maximal L1-regularity for initial-boundary value problems of the evolution equation of parabolic type in both the inhomoge- neous Dirichlet and the Neumann boundary conditions. For 0-Dirichlet boundary T. Ogawa and S. Shimizu 30 Page 4 of J. Evol. Equ. 67 data, Danchin–Mucha [12] obtained maximal L1-regularity for the initial-boundary value problem (1.2) in the half-space with ai j(t, x) = δi j and g(t, x′) ≡0. On the other hand, maximal L1-regularity for the case of nonzero boundary condition re- quires very different treatment and it is far from obvious since the time exponent is the end-point and it is neither clear nor straightforward if a natural extension from the known result Proposition 1.1 holds in a natural exponent. 1.1. The Dirichlet boundary condition Here, we consider in- homogeneous initial-boundary value problems in both the Drichlet and the Neumann boundary conditions and show that a natural extension from Proposition 1.1 does not hold in general (namely the conditions for the boundary data in (1.3) with ρ = 1) and we explicitly prove it by showing the end-point time exponent invites the end-point interpolation exponent similar to the case of the initial value problem. Beforestatingourresults,wedeﬁneBesovspacesandLizorkin–Triebelspacesinthe half-spaceandthehalf-line. Sincetheglobal estimaterequires thebasespacefor spatial variable x in the homogeneous Besov space, we introduce the homogeneous Besov space over Rn + (see for details Bergh-Löfström [6], Lizorkin [30], Peetre [38,39], Triebel [45,47]). Deﬁnition (The Besov and the Lizorkin–Triebel spaces). Let s ∈R, 1 ≤p, σ ≤∞. Let {φ j} j∈Z be the Littlewood-Paley dyadic decomposition of unity for x ∈Rn, namely φ is the Fourier transform of a smooth radial function φ with φ(ξ) ≥0 and supp φ ⊂{ξ ∈Rn \| 2−1 < \|ξ\| < 2} and φ ⊂{ξ ∈Rn \| 2−1 < \|ξ\| < 2}, and φ j(ξ) = φ(2−jξ), j∈Z φ j(ξ) = 1 for any ξ ∈Rn \ {0}, j ∈Z φ ⊂{ξ ∈Rn \| 2−1 < \|ξ\| < 2}, and φ j(ξ) = φ(2−jξ), j∈Z φ j(ξ) = 1 for any ξ ∈Rn \ {0}, j ∈Z and φ0(ξ) + j≥1 φ j(ξ) = 1 for any ξ ∈Rn, (1.4) (1.4) where φ0(ξ) ≡ζ(\|ξ\|) with a low frequency cutoff ζ(r) = 1 for 0 ≤r < 1 and ζ(r) = 0 for 2 < r. For s ∈R and 1 ≤p, σ ≤∞, ˙Bs p,σ(Rn) be the homogeneous Besov space with norm where φ0(ξ) ≡ζ(\|ξ\|) with a low frequency cutoff ζ(r) = 1 for 0 ≤r < 1 and ζ(r) = 0 for 2 < r. For s ∈R and 1 ≤p, σ ≤∞, ˙Bs p,σ(Rn) be the homogeneous Besov space with norm ∥˜f ∥˙Bsp,σ ≡ ⎧ ⎨ ⎩ j∈Z 2sσ j∥φ j ∗˜f ∥σ p 1/σ , 1 ≤σ < ∞, sup j∈Z 2sj∥φ j ∗˜f ∥σ p, σ = ∞, where f ∗g denotes the convolution between Schwartz class functions f and g ∈ S(Rn) given by f ∗g(x) = c−1 n Rn f (x −y)g(y)dy, cn = (2π)−n 2 and for f , g ∈S′ as the distribution sense, where S∗is the tempered distributions. 1.1. The Dirichlet boundary condition Bs p,σ(Rn) be the inhomogeneous Besov space with norm and for f , g ∈S′ as the distribution sense, where S∗is the tempered distributions. Bs p,σ(Rn) be the inhomogeneous Besov space with norm ∥˜f ∥Bsp,σ ≡ ⎧ ⎪⎪⎨ ⎪⎪⎩ ∥φˆ0 ∗˜f ∥p + j∈Z 2sσ j∥φ j ∗˜f ∥σ p 1/σ , 1 ≤σ < ∞, ∥φˆ0 ∗˜f ∥p + sup j∈Z 2sj∥φ j ∗˜f ∥σ p, σ = ∞. σ = ∞. Maximal L1-regularity for parabolic Page 5 of 67 30 Maximal L1-regularity for parabolic Page 5 of 67 30 J. Evol. Equ. Page 5 of 67 30 For s ∈R, 1 ≤p < ∞and 1 ≤σ ≤∞, ˙Fs p,σ(Rn) be the homogeneous Lizorkin– Triebel space with norm ∥˜f ∥˙Fsp,σ ≡ ⎧ ⎪⎨ ⎪⎩ j∈Z 2sσ j\|φ j ∗˜f (·)\|σ1/σ p, 1 ≤σ < ∞, sup j∈Z 2sj\|φ j ∗˜f (·)\| p, σ = ∞. We deﬁne the homogeneous Besov space ˙Bs p,σ(Rn +) as the set of all measurable functions f in Rn + satisfying ∥f ∥˙Bsp,σ (Rn +) ≡inf ∥˜f ∥˙Bsp,σ (Rn) < ∞; & ˜f = f (x′, xn) (xn > 0) any extension (xn < 0) , ˜f = j∈Z φ j ∗˜f in S′ . (1.5) (1.5) Analogously, we deﬁne the inhomogeneous Besov space Bs p,σ(Rn +) as the set of all measurable functions f in Rn + satisfying Analogously, we deﬁne the inhomogeneous Besov space Bs p,σ(Rn +) as the set of all measurable functions f in Rn + satisfying ∥f ∥Bsp,σ (Rn +) ≡inf ∥˜f ∥Bsp,σ (Rn) < ∞; ˜f = f (x′, xn) (xn > 0) any extension (xn < 0) . Deﬁnition (The Bochner–Lizorkin–Triebel spaces). Let s ∈R, 1 ≤p, σ ≤∞and X(Rn +) be a Banach space on Rn + with the norm ∥·∥X. Let {ψk}k∈Z be the Littlewood- Paleydyadicdecompositionofunityfort ∈R.Fors ∈Rand1 ≤p < ∞, ˙Fs p,σ(R; X) be the Bochner–Lizorkin–Triebel space with norm ∥˜f ∥˙Fsp,σ (R;X) ≡ ⎧ ⎪⎨ ⎪⎩ k∈Z 2sσk∥ψk ∗˜f (t, ·)∥σ X 1/σ L p(Rt), 1 ≤σ < ∞, supk∈Z 2sk∥ψ j ∗˜f (t, ·)∥X L p(Rt), σ = ∞. 1.1. The Dirichlet boundary condition Analogously above, we deﬁne the Bochner–Lizorkin–Triebel spaces ˙Fs p,σ(I; X) as the set of all measurable functions f on X satisfying ∥f ∥˙Fsp,σ (I;X) ≡inf ∥˜f ∥˙Fsp,σ (R;X) < ∞; ˜f = f (t, x) (t ∈I) any extension (t ∈R \ I) . We note that all the spaces of homogeneous type are understood as the Banach spaces by introducing the quotient spaces identifying all polynomial differences. We note that all the spaces of homogeneous type are understood as the Banach spaces by introducing the quotient spaces identifying all polynomial differences. We assume that the real-valued coefﬁcients {ai j(t, x)}1≤i, j≤n satisfy the following conditions. T. Ogawa and S. Shimizu 67 J. Evol. Equ. J. Evol. Equ. Assumption 1. For 1 ≤i, j ≤n and (t, x) ∈I × Rn +, Assumption 1. For 1 ≤i, j ≤n and (t, x) ∈I × Rn +, Assumption 1. For 1 ≤i, j ≤n and (t, x) ∈I × Rn +, (1) ai j(t, x) = δi j + ai j + bi j(t, x), where δi j and ai j denote the Kronecker delta and components of a positive constant matrix, respectively, p p p y (2) ai j = a ji and bi j(t, x) = b ji(t, x) for all t > 0 and x ∈Rn +, (3) there exists a constant c > 0 such that for any ξ ∈Rn, (2) ai j = a ji and bi j(t, x) = b ji(t, x) for all t > 0 and x ∈Rn +, (3) there exists a constant c > 0 such that for any ξ ∈Rn, 1≤i, j≤n ai j(t, x)ξiξ j ≥c\|ξ\|2, (t, x) ∈I × Rn +, (4) bi j ∈BUC R+; ˙B n q q,1(Rn +) for some 1 ≤q < ∞, where BUC(I; X) denotes a set of all bounded uniformly continuous functions on I. (4) bi j ∈BUC R+; ˙B n q q,1(Rn +) for some 1 ≤q < ∞, where BUC(I; X) denotes a set of all bounded uniformly continuous functions on I. 1.1. The Dirichlet boundary condition It is known that for 1 ≤q < ∞, ˙B n q q,1(Rn) satisﬁes S0(Rn) →˙B n q q,1(Rn) → Cv(Rn), where S0 denotes the rapidly decreasing smooth functions with vanishing at the origin of its Fourier transform and Cv denotes the set of all continuous functions with vanishing at inﬁnity, respectively (cf. [34, Proposition 2.3]). It also satisﬁes the product inequality, namely the following inequality holds: ∥bi jh∥˙Bs p,1 ≤C∥bi j∥˙B nq q,1 ∥h∥˙Bs p,1, 1 ≤p ≤∞, −n q < s ≤0 (cf. Adibi-Paicu [1,4]). Hence to ensure the uniformly elliptic condition (3) in As- sumption 1, we split the coefﬁcients ai j(t, x) into the constant parts δi j + ai j and the decreasing functions bi j(t, x) as \|x\| →∞. Then our main result for the end-point case of maximal regularity to the problem (1.2) now read as the following: Theorem 1.2 (The Dirichlet boundary condition). Let 1 < p < ∞, −1+1/p < s ≤ 0 and assume that the coefﬁcients {ai j}1≤i, j≤n satisfy Assumption 1. (1) S h b ( ) 0 f ll 1 i j Th h bl (1 2) d i Theorem 1.2 (The Dirichlet boundary condition). Let 1 < p < ∞, −1+1/p < s ≤ 0 and assume that the coefﬁcients {ai j}1≤i, j≤n satisfy Assumption 1. (1) Suppose that bi j(t, x) ≡0 for all 1 ≤i, j ≤n. Then, the problem (1.2) admits a unique solution j ≤, j≤ (1) Suppose that bi j(t, x) ≡0 for all 1 ≤i, j ≤n. Then, the problem (1.2) admits a unique solution u ∈˙W 1,1 R+; ˙Bs p,1(Rn +) , u ∈L1 R+; ˙Bs p,1(Rn +) , if and only if the external, initial and boundary data in (1.2) satisfy f ∈L1 R+; ˙Bs p,1(Rn +) , u0 ∈˙Bs p,1(Rn +), (1.6) g ∈˙F1−1/2p 1,1 R+; ˙Bs p,1(Rn−1) ∩L1 R+; ˙Bs+2−1/p p,1 (Rn−1) , (1.7) (1.6) (1.7) respectively. Besides, the solution u satisﬁes the following estimate for some constant CM > 0 depending only on p, s and n respectively. Besides, the solution u satisﬁes the following estimate for some constant CM > 0 depending only on p, s and n ∥∂tu∥L1(R+; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(R+; ˙Bs p,1(Rn +)) ≤CM ∥u0∥˙Bs p,1(Rn +) + ∥f ∥L1(R+; ˙Bs p,1(Rn +)) + ∥g∥˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥g∥L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) . 1.1. The Dirichlet boundary condition ∥∂tu∥L1(R+; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(R+; ˙Bs p,1(Rn +)) ≤CM ∥u0∥˙Bs p,1(Rn +) + ∥f ∥L1(R+; ˙Bs p,1(Rn +)) + ∥g∥˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥g∥L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) . + ∥g∥˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥g∥L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) . + ∥g∥˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥g∥L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) . Maximal L1-regularity for parabolic Page 7 of 67 30 Maximal L1-regularity for parabolic Maximal L1-regularity for parabolic Page 7 of 67 30 Page 7 of 67 30 Page 7 of 67 30 J. Evol. Equ. J. Evol. Equ. (2) Let 1 ≤q < ∞, −n/q < s ≤0 and let I = (0, T ) for T < ∞. Then, the problem (1.2) admits a unique solution (2) Let 1 ≤q < ∞, −n/q < s ≤0 and let I = (0, T ) for T < ∞. Then, the problem (1.2) admits a unique solution (2) Let 1 ≤q < ∞, −n/q < s ≤0 and let I = (0, T ) for T < ∞. Then, the problem (1.2) admits a unique solution u ∈˙W 1,1 I; ˙Bs p,1(Rn +) , u ∈L1 R+; ˙Bs p,1(Rn +) , if and only if the external force, the initial data and the boundary data satisfy (1.6) and (1.7) with replacing R+ into I. Besides, the solution u satisﬁes the following estimate for some constant CM = CM(n, p, q, { ai j}) > 0 ∥∂tu∥L1(I; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(I; ˙Bs p,1(Rn +)) ≤CM T 0 eμ(T −s)∥f (s)∥˙Bs p,1ds +CM 1 + max 1≤i, j≤n ∥bi j∥L∞(I; ˙Bn/q q,1 ) eμT −1 × ∥u0∥˙Bs p,1 + ∥g∥˙F1−1/2p 1,1 (I; ˙Bs p,1(Rn−1)) + ∥g∥L1(I; ˙Bs+2−1/p p,1 (Rn−1)) , (1.8) (1.8) where μ = C2 M log(1 + CM). where μ = C2 M log(1 + CM). where μ = C2 M log(1 + CM). Remarks. (i) The solution in Theorem 1.2 satisﬁes u ∈Cb [0, T ); ˙Bs p,1(Rn +) u ∈Cb [0, T ); ˙Bs p,1(Rn +) with T = ∞for the solution u in (1) and with T < ∞in (2). The linear evolution gen- erated by the elliptic operator generates C0-semigroup in ˙Bs p,1(Rn +) and the estimate of maximal L1-regularity ensures that the absolute continuity in t to the solution. 1.1. The Dirichlet boundary condition In such a case, the restriction on the initial data u0 can be relaxed into the class of inhomogeneous Besov spaces Bs p,1(Rn +) ⊃˙Bs p,1(Rn +) and the constant appeared in the estimate can be estimated as CM ≃O(log T ) (T →∞) (see [34]). The function class for the x-variable in Theorem 1.2 is restricted in ˙Bs p,1(Rn +) ⊊ ˙W s,p(Rn +) and this restriction is necessary for obtaining maximal L1-regularity (see [34]). Connecting this fact, the conditions on the coefﬁcients ai j(t, x) are so far best available. In general, the elliptic and parabolic type estimates in L p-setting allow us to treat much general coefﬁcients such as ai j(t, x) ∈V MO(I ×Rn) for the whole space case, where V MO stands for the vanishing mean oscillation (see e.g. Krylov [27]). However, since maximal regularity in L1 generally fails for the Lebesgue spaces L p, we need to restrict the spatial function class to the Besov space ˙B0 p,1 ⊊L p even for the whole space Rn case. Such a restriction limits the condition on the coefﬁcients as given in Assumption 1. On the other hand, Theorem 1.2 does not cover the end-point spatial exponent p = 1 nor p = ∞. Instead of the above result, we show a substituting estimate holds in L p(Rn +). Theorem 1.3 (The Dirichlet boundary condition). Assume that the coefﬁcients {ai j}1≤i, j≤n be constants that satisfy Assumption 1, i.e., bi j(t, x) ≡0. (1) Let 1 ≤p < ∞. If the external force, the initial data and the boundary data satisfy f ∈L1 R+; ˙B0 p,1(Rn +) , u0 ∈˙B0 p,1(Rn +), g ∈˙F1−1/2p 1,1 (R+; L p(Rn−1)) ∩L1(R+; B2−1/p p,1 (Rn−1)), (1.10) (1.10) then there exists a unique solution u to (1.2) in W 1,1(R+; L p(Rn +)) ∩L1(R+; ˙W 2,p(Rn +)) and which satisﬁes the following estimate: and which satisﬁes the following estimate: and which satisﬁes the following estimate: 1.1. The Dirichlet boundary condition (ii) Since 1 1/(2p) < 1 for all 1 < p < ∞the compatibility condition with T = ∞for the solution u in (1) and with T < ∞in (2). The linear evolution gen- erated by the elliptic operator generates C0-semigroup in ˙Bs p,1(Rn +) and the estimate of maximal L1-regularity ensures that the absolute continuity in t to the solution. (ii) Since 1 −1/(2p) < 1 for all 1 < p < ∞, the compatibility condition u0(x′, xn)\|xn=0 = g(t, x′)\|t=0 (1.9) (1.9) is not necessarily required for (1.8). is not necessarily required for (1.8). (iii) If p = ∞, the corresponding result holds for the homogeneous Besov space ˙Bs ∞,1(Rn) ≡C∞ 00(Rn) ˙Bs ∞,1(Rn), where C∞ 00(Rn) denotes all compactly supported smooth functions with vanishing at the origin of its Fourier image and (iii) If p = ∞, the corresponding result holds for the homogeneous Besov space ˙Bs ∞,1(Rn) ≡C∞ 00(Rn) ˙Bs ∞,1(Rn), where C∞ 00(Rn) denotes all compactly supported smooth functions with vanishing at the origin of its Fourier image and ∥f ∥˙Bs ∞,1(Rn +) ≡inf ∥f ∥˙Bs p,1(Rn) < ∞; & ˜f = f (x′, xn) (xn > 0) any extension (xn ≤0) , ˜f = j∈Z φ j ∗˜f in S′ , instead of the Besov space ˙Bs ∞,1(Rn +) with imposing the compatibility condition (1.9) if (s, p) = (0, ∞). Note that ˙B0 ∞,1(Rn +) ⊂Cv(Rn +) for the endpoint case (s, p) = (0, ∞). instead of the Besov space ˙Bs ∞,1(Rn +) with imposing the compatibility condition (1.9) if (s, p) = (0, ∞). Note that ˙B0 ∞,1(Rn +) ⊂Cv(Rn +) for the endpoint case (s, p) = (0, ∞). T. Ogawa and S. Shimizu J. Evol. Equ. 67 We only show the estimate for R+ in time but a similar estimate for the ﬁnite time interval I = (0, T ) with T < ∞is also available. In such a case, the restriction on the initial data u0 can be relaxed into the class of inhomogeneous Besov spaces Bs p,1(Rn +) ⊃˙Bs p,1(Rn +) and the constant appeared in the estimate can be estimated as CM ≃O(log T ) (T →∞) (see [34]). We only show the estimate for R+ in time but a similar estimate for the ﬁnite time interval I = (0, T ) with T < ∞is also available. and which satisﬁes the following estimate: ∥∂tu∥L1(R+;L p(Rn +)) + ∥∇2u∥L1(R+;L p(Rn +)) ≤C ∥u0∥˙B0 p,1(Rn +) + ∥f ∥L1(R+; ˙B0 p,1(Rn +)) + ∥g∥˙F1−1/2p 1,1 (R+;L p(Rn−1)) +∥g∥L1(R+;B2−1/p p,1 (Rn−1)) , (1.11) (1.11) where C is depending only on p and n. (2) F ∞if th t l f th i iti l d th b d d t ti f where C is depending only on p and n. where C is depending only on p and n. (2) For p = ∞, if the external force, the initial and the boundary data satisfy p g y p (2) For p = ∞, if the external force, the initial and the boundary data satisfy (2) For p = ∞, if the external force, the initial and the boundary data satisfy f ∈L1 R+; ˙B0 ∞,1(Rn +) , u0 ∈˙B0 ∞,1(Rn +), g ∈˙F1 1,1(R+; Cv(Rn−1)) ∩L1(R+; B2 ∞,1(Rn−1)), with imposing the compatibility condition (1.9), then there exists a unique solution u to (1.2) in with imposing the compatibility condition (1.9), then there exists a unique solution u to (1.2) in W 1,1(R+; L∞(Rn +)) ∩L1(R+; ˙W 2,∞(Rn +)) Page 9 of 67 30 Maximal L1-regularity for parabolic Page 9 of 67 30 Maximal L1-regularity for parabolic J. Evol. Equ. J. Evol. Equ. d only if the external, initial and boundary data in (1.12) satisfy if and only if the external, initial and boundary data in (1.12) satisfy f ∈L1(R+; ˙Bs p,1(Rn +)), u0 ∈˙Bs p,1(Rn +), (1.13) g ∈˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) ∩L1(R+; ˙Bs+1−1/p p,1 (Rn−1)), (1.14) and the corresponding estimate to (1.11) holds as and the corresponding estimate to (1.11) holds as ∥∂tu∥L1(R+;L∞(Rn +)) + ∥∇2u∥L1(R+;L∞(Rn +)) ≤C ∥u0∥˙B0 ∞,1(Rn +) + ∥f ∥L1(R+; ˙B0 ∞,1(Rn +)) + ∥g∥˙F1 1,1(R+;Cv(Rn−1)) + ∥g∥L1(R+; ˙B2 ∞,1(Rn−1)) . The main difference between Theorem 1.2 and 1.3 is not only on the condition of the end-point exponents p = 1, ∞but also the difference of boundary regularity. The estimate in Theorem 1.3 only requires the Lebesgue and the inhomogeneous Besov regularity in the spatial direction and this is closer result to the known estimate in (1.3) due to Denk–Hieber–Prüss [15]. Note that the inhomogeneous Besov space B0 p,1(Rn +) is slightly wider than the one for homogeneous space ˙B0 p,1(Rn +). However, it does not stand for the strict sense of maximal regularity since the regularity for the initial and external force is more than the regularity for the solution itself by ˙B0 p,1(Rn +) ⊊L p(Rn +) for all 1 ≤p ≤∞. Remarks. (i) In the second statement of Theorem 1.3, ˙F1 1,1(R+; Cv(Rn−1)) is em- bedded into the continuous functions Cb([0, ∞); Cv(Rn−1)) and the compatibility condition (1.9) is required. (ii) Shibata–Shimizu [41,42] showed a boundary estimate by extending the bound- ary data g into g : R+ × Rn + →R and assume that the extended function satisﬁes g ∈W 1,ρ(R+; L p(Rn +)) ∩Lρ(R+; W 2,p(Rn +)) for 1 < ρ, p < ∞. However, the estimate there does not include the endpoint exponent ρ = 1. (iii) In the case 1 < ρ ≤∞, one can extend Proposition 1.1 into a time global version. Indeed, the case p = 1 for (1) in Theorem 1.3 is the corresponding estimate to Proposition 1.1. 1.2. The Neumann boundary condition 2. The Neumann boundary condition Similar to the initial-boundary value problem with the Dirichlet condition, we con- sider the initial-boundary value problem of the Neumann boundary condition. We assume Assumption 1 for the coefﬁcients {ai j}1≤i, j≤n. ⎧ ⎪⎪⎪⎪⎨ ⎪⎪⎪⎪⎩ ∂tu − 1≤i, j≤n ai j(t, x)∂i∂ju = f, t ∈I, x ∈Rn +, ∂nu(t, x′, xn) xn=0 = g(t, x′), t ∈I, x′ ∈Rn−1, u(t, x) t=0 = u0(x), x ∈Rn +, (1.12) where x = (x′, xn) ∈Rn + and ∂n denotes the normal derivative ∂/∂xn at any boundary point of Rn +. Denk-Hieber-Prüss [15] showed the following: 30 Page 10 of 67 30 Page 10 of 67 T. Ogawa and S. Shimizu J. if 1/2 −1/(2p) > 1/ρ, then assume further that the compatibility condition if 1/2 −1/(2p) > 1/ρ, then assume further that the compatibility condition (∂nu0)(x′, xn)\|xn=0 = g(t, x′)\|t=0. For the case of Neumann boundary problem (1.12), we obtain end-point maximal L1-maximal regularity as follows: Theorem 1.5 (The Neumann boundary condition). Let 1 < p < ∞, −1 + 1/p < s ≤0 and assume that the coefﬁcients {ai j}1≤i, j≤n satisfy Assumption 1. (1) Suppose that bi j(t, x) ≡0 for all 1 ≤i, j ≤n. Then, the problem (1.12) admits a unique solution u ∈˙W 1,1 R+; ˙Bs p,1(Rn +) , u ∈L1 R+; ˙Bs p,1(Rn +) , and the corresponding estimate to (1.11) holds as Evol. Equ. J. Evol. Equ. Proposition 1.4 (The Neumann boundary condition [15]). Let 1 < ρ, p < ∞with 1/2 −1/(2p) ̸= 1/ρ, T < ∞and set I = (0, T ) for T < ∞. The initial-boundary value problem (1.12) has a unique solution u in W 1,ρ(I; L p(Rn +))∩Lρ(I; W 2,p(Rn +)) if and only if f ∈Lρ(I; L p(Rn +)), u0 ∈B2(1−1/ρ) p,ρ (Rn +), g ∈F1/2−1/2p ρ,p (I; L p(Rn−1)) ∩Lρ(I; B1−1/p p,p (Rn−1)), if 1/2 −1/(2p) > 1/ρ, then assume further that the compatibility condition respectively. Moreover end-point maximal L1-regularity holds: respectively. Moreover end-point maximal L1-regularity holds: ∥∂tu∥L1(R+; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(R+; ˙Bs p,1(Rn +)) ≤C ∥u0∥˙Bs p,1(Rn +) + ∥f ∥L1(R+; ˙Bs p,1(Rn +)) + ∥g∥˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥g∥L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) , where C is depending only on p, s and n. (2) Let 1 ≤q < ∞and −n/q < s ≤0. For any T < ∞, let I = (0, T ). Then, there exists a unique solution u to (1.12) where C is depending only on p, s and n. (2) Let 1 ≤q < ∞and −n/q < s ≤0. For any T < ∞, let I = (0, T ). Then, there exists a unique solution u to (1.12) where C is depending only on p, s and n. (2) Let 1 ≤q < ∞and −n/q < s ≤0. For any T < ∞, let I = (0, T ). Then, there exists a unique solution u to (1.12) u ∈˙W 1,1 I; ˙Bs p,1(Rn +) , u ∈L1 I; ˙Bs p,1(Rn +) , u ∈˙W 1,1 I; ˙Bs p,1(Rn +) , u ∈L1 I; ˙Bs p,1(Rn +) , u ∈˙W 1,1 I; ˙Bs p,1(Rn +) , u ∈L1 I; ˙Bs p,1(Rn +) , if and only if the external force, the initial data and the boundary data satisfy (1.13) and (1.14) with replacing R+ into I. Besides, the solution u satisﬁes the following estimate for some constant CM = CM(n, p, q, { ai j}) > 0 J. Evol. Equ. Maximal L1-regularity for parabolic Page 11 of 67 30 Maximal L1-regularity for parabolic Page 11 of 67 30 Maximal L1-regularity for parabolic Page 11 of 67 30 Page 11 of 67 30 J. Evol. Equ. J. Evol. Equ. ∥∂tu∥L1(I; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(I; ˙Bs p,1(Rn +)) ≤CM T 0 eμ(T −s)∥f (s)∥˙Bs p,1ds +CM 1 + max 1≤i, j≤n ∥bi j∥L∞(I; ˙Bn/q q,1 ) eμT −1 × ∥u0∥˙Bs p,1 +∥g∥˙F1/2−1/2p 1,1 (I; ˙Bs p,1(Rn−1))+∥g∥L1(I; ˙Bs+1−1/p p,1 (Rn−1)) , (1.15) (1.15) where μ = C2 M log(1 + CM). where μ = C2 M log(1 + CM). where μ = C2 M log(1 + CM). where μ = C2 M log(1 + CM). For the case of the Neumann boundary condition, we have an analogous compensated end-point result as in Theorem 1.3 with L p type boundary data. Theorem 1.6 (The Neumann boundary condition). respectively. Moreover end-point maximal L1-regularity holds: Assume that the coefﬁcients {ai j}1≤i, j≤n satisfy Assumption 1 and be constants, i.e., bi j(t, x) ≡0. (1) Let 1 ≤p < ∞. Assume that the external, initial and boundary data satisfy f ∈L1 R+; ˙B0 p,1(Rn +) , u0 ∈˙B0 p,1(Rn +), (1.16) g ∈˙F1/2−1/2p 1,1 (R+; L p(Rn−1)) ∩L1(R+; B1−1/p p,1 (Rn−1)), (1.17) then there exists a unique solution u to (1.12) in then there exists a unique solution u to (1.12) in ˙W 1,1(R+; L p(Rn +)) ∩L1(R+; ˙W 2,p(Rn +)) and which satisﬁes the following estimate: and which satisﬁes the following estimate: and which satisﬁes the following estimate: ∥∂tu∥L1(R+;L p(Rn +)) + ∥∇2u∥L1(R+;L p(Rn +)) ≤C ∥u0∥˙B0 p,1(Rn +) + ∥f ∥L1(R+; ˙B0 p,1(Rn +)) + ∥g∥˙F1/2−1/2p 1,1 (R+;L p(Rn−1)) ∥∂tu∥L1(R+;L p(Rn +)) + ∥∇2u∥L1(R+;L p(Rn +)) ≤C ∥u0∥˙B0 p,1(Rn +) + ∥f ∥L1(R+; ˙B0 p,1(Rn +)) + ∥g∥˙F1/2−1/2p 1,1 (R+;L p(Rn−1)) + ∥g∥L1(R+;B1−1/p p,1 (Rn−1)) , + ∥g∥L1(R+;B1−1/p p,1 (Rn−1)) , where C is depending only on p and n. (2) For p = ∞, the corresponding result to (1) holds, i.e., where C is depending only on p and n. (2) For p = ∞, the corresponding result to (1) holds, i.e., f ∈L1 R+; ˙B0 ∞,1(Rn +) , u0 ∈˙B0 ∞,1(Rn +), g ∈˙F1/2 1,1 (R+; Cv(Rn−1)) ∩L1(R+; ˙B1 ∞,1(Rn−1)), then there exists a unique solution u to (1.12) in then there exists a unique solution u to (1.12) in ˙W 1,1(R+; L∞(Rn +)) ∩L1(R+; ˙W 2,∞(Rn +)) and which satisﬁes the following estimate: and which satisﬁes the following estimate: ∥∂tu∥L1(R+;L∞(Rn +)) + ∥∇2u∥L1(R+;L∞(Rn +)) ≤C ∥u0∥˙B0 ∞,1(Rn +) + ∥f ∥L1(R+; ˙B0 ∞,1(Rn +)) + ∥g∥˙F1/2 1,1 (R+;Cv(Rn−1)) + ∥g∥L1(R+;B1 ∞,1(Rn−1)) , + ∥g∥L1(R+;B1 ∞,1(Rn−1)) , where C is depending only on n. where C is depending only on n. T. Ogawa and S. Shimizu 30 Page 12 of 67 age 12 of 67 J. Evol. Equ. Remark. In the case p = ∞, the compatibility condition (1.9) appeared in Theo- rem 1.3 (2) is redundant in Theorem 1.6 (2) since the regularity for the boundary data (1.17) is weaker than the case in Theorem 1.3 and the boundary data is not the continuous function in t-variable. Therestofthispaperisorganizedasfollows.Wepresentthebasicformulationforthe proof in particular the reduction to the boundary value problems of the heat equations in the next section. We construct explicit solution formulas of the fundamental solutions in Sect. 3. Section 4 is devoted to prove for the Dirichlet condition case and Sect. 5 for the Neumann boundary condition case. In Sect. 6, we devote the proof of the key estimate almost orthogonality (2.12). Finally, we show the optimality of the main result by showing the sharp boundary trace estimate in Sect. 7. Throughout this paper we use the following notations. Let R+ = (0, ∞) and Rn + denote the n-dimensional Euclidean half-space; {(x′, xn); x′ ∈Rn−1, xn ∈R+}. For x ∈Rn, ⟨x⟩≡(1 + \|x\|2)1/2. and which satisﬁes the following estimate: The Fourier and the inverse Fourier transforms are deﬁned for any rapidly decreasing function f ∈S(Rn) with cn = (2π)−n/2 by f (ξ) = F[ f ](ξ) ≡cn Rn e−ix·ξ f (x)dx, F−1[ f ](x) ≡cn Rn eix·ξ f (ξ)dξ. For f ∈S′(Rn),wealsodenote f (ξ′) = F[ f ](ξ′).Foranyfunctions f = f (t, x′, xn) and g = g(t, x′, xn), f ∗ (t) g, f ∗ (t,x′) g and f ∗ (xn) g stand for the convolution between f and g with respect to the variable indicated under ∗, respectively. In the summation k∈Z, the parameter k runs for all integers k ∈Z and for k≤j, k runs for all integers less than or equal to j ∈Z. In the norm for the Bochner spaces on ˙Fs p,ρ I; X(Rn−1) , we use ∥f ∥˙Fsp,ρ(I;X) = ∥f ∥˙Fsp,ρ(I;X(Rn−1)) unless it may cause any confusion. For a ∈Rn, we denote BR(a) as the open ball centered at a with its radius R > 0. We also denote the compliment of BR(0) by Bc R. Various constants are simply denoted by C unless otherwise stated. 2. Reduction to the heat equation and outline of the proofs The outline of the proof of Theorems 1.2 and 1.3 is summarized as follows: We decompose the initial-boundary value problem (1.2) into the following three problems and reduce the problem into the inhomogeneous boundary value problem with 0 initial and external force: t ∈I, x ∈Rn, x′ ∈Rn−1, (2.1) ⎧ ⎨ ⎩ ∂tu1 − u1 = 0, t ∈I, x ∈Rn, u1(t, x) t=0 = u0(x′, xn), xn > 0 −u0(x′, −xn), xn ≤0, x′ ∈Rn−1, (2.1) ⎧ ⎪⎨ ⎪⎩ ∂tu2 − u2 = 0, t ∈I, x ∈Rn +, u2(t, x′, xn) xn=0 = g(t, x′) −u1(t, x′, xn) xn=0, t ∈I, x′ ∈Rn−1, u2(t, x) t=0 = 0, x ∈Rn +, (2.2) ⎧ ⎨ ⎩ ∂tu1 − u1 = 0, t ∈I, x ∈Rn, u1(t, x) t=0 = u0(x′, xn), xn > 0 −u0(x′, −xn), xn ≤0, x′ ∈Rn−1, (2.1) ⎧ ⎪⎨ ⎪⎩ ∂tu2 − u2 = 0, t ∈I, x ∈Rn +, u2(t, x′, xn) xn=0 = g(t, x′) −u1(t, x′, xn) xn=0, t ∈I, x′ ∈Rn−1, u2(t, x) t=0 = 0, x ∈Rn +, (2.2) (2.1) Maximal L1-regularity for parabolic Page 13 of 67 30 Maximal L1-regularity for parabolic Page 13 of 67 30 Maximal L1-regularity for parabolic J. Evol. Equ. Page 13 of 67 30 ⎧ ⎪⎨ ⎪⎩ ∂tu3 − i, j ai j(t, x)∂i∂ju3 = f + i, j ai j +bi j(t, x) ∂i∂j(u1+u2)≡F, t ∈I, x ∈Rn +, u3(t, x′, xn) xn=0 =0, t ∈I, x′ ∈Rn−1, u3(t, x) t=0 =0, x ∈Rn +, (2.3 (2.3) where the coefﬁcients ai j and bi j(t, x) are deﬁned in Assumption 1 (1) and f = f (t, x), g = g(t, x′), u0(x) are given external force, the Dirichlet boundary data, and initial data, respectively. where the coefﬁcients ai j and bi j(t, x) are deﬁned in Assumption 1 (1) and f = f (t, x), g = g(t, x′), u0(x) are given external force, the Dirichlet boundary data, and initial data, respectively. Then u(t, x′, xn) = u1(t, x′, xn) xn>0 + u2(t, x′, xn) + u3(t, x′, xn) is the solution of (1.2). is the solution of (1.2). 2. Reduction to the heat equation and outline of the proofs The external force F in (2.3) contains not only the given data f but the solutions u1 and u2 to (2.1) and (2.2) and we need to verify the regularity of u1 and u2 in order to solve (2.3). Indeed, if u1 and u2 have maximal L1-regularity: For 1 < p < ∞and −1 + 1/p < s ≤0, u1, u2 ∈W 1,1 R+; ˙Bs p,1(Rn +) , u1, u2 ∈L1 R+; ˙Bs p,1(Rn +) , (2.4) hen under the assumption bi j ∈BUC R+; ˙B n q q,1(Rn−1 + ) (1 ≤q < −n s ) it holds that (2.4) i, j ai j + bi j(t, x) ∂i∂j(u1 + u2) L1(R+; ˙Bs p,1(Rn +)) i, j ai j + bi j(t, x) ∂i∂j(u1 + u2) L1(R+; ˙Bs p,1(Rn +)) ≤C 1 + sup t>0 ∥bi j(t, ·)∥˙B nq q,1(Rn +) ∥ u1∥L1(R+; ˙Bs p,1) + i, j ai j + bi j(t, x) ∂i∂j(u1 + u2) L1(R+; ˙Bs p,1(Rn +)) ≤C 1 + sup t>0 ∥bi j(t, ·)∥˙B nq q,1(Rn +) ∥ u1∥L1(R+; ˙Bs p,1) + ∥ u2∥L1(R+; ˙Bs p,1) . ≤C 1 + sup t>0 ∥bi j(t, ·)∥˙B nq q,1(Rn +) ∥ u1∥L1(R+; ˙Bs p,1) + ∥ u2∥L1(R+; ˙Bs p,1) . We set g−u1 as h and assume that it is given. When (s, p) = (0, ∞), the compatibility condition h(t, x′) = g(t, x′) −u1(t, x′, 0) on t = 0 is required and it coincides with u2(t, x) t=0 = 0, namely g(t, x′) t=0 = u0(x′, xn) xn=0. The requirement for the compatibility condition is natural for maximal Lρ-regularity in the cases 1 < ρ < ∞. However, such a restriction is not adopted for maximal L1-regularity except the endpoint case (s, p) = (0, ∞). For the problem (2.3), we ﬁrst notice that assuming the regularity (2.4), the external force F is in L1(I; ˙Bs p,1) under the regularity assumption on bi j(t, x) and ai j being constant coefﬁcients. Then, we extend the problem (2.3) into the whole space by an appropriate extension of data and coefﬁcients and maximal regularity (1.8) follows from the estimate for the Cauchy problem in Rn shown in [34]. Note that the result in [34] treats only the case ai j = 0; however, the analogous result follows for the constant positive coefﬁcient case. Besides the solution u is subject to the estimate: ∥∂tu∥L1(R+; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(R+; ˙Bs p,1(Rn +)) ≤C ∥h∥˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥h∥L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) , where C is depending only on p, s and n. 2. Reduction to the heat equation and outline of the proofs Therefore, our main issue is to consider the initial boundary value problem (2.2) ⎧ ⎪⎨ ⎪⎩ ∂tu − u = 0, t ∈I, x ∈Rn +, u(t, x′, xn) xn=0 = h(t, x′), t ∈I, x′ ∈Rn−1, u(t, x)\|t=0 = 0, x ∈Rn +, (2.5) (2.5) 30 Page 14 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. 67 where the boundary function h(t, x′) is given by the function after a proper linear transformed function of g(t, x′)−u1(t, x′, xn) xn=0 in (2.2). Once we obtain maximal L1-regularity to (2.5) with the boundary trace, then the original problem can be reduced into the initial value problem, and it can be reduced into the Cauchy problem in the whole Euclidian space Rn. Note that the solution u1(t, x) that has regularity (2.4) has the boundary trace estimate (see Theorem 7.1 in Section 7) and the condition on the boundary data h is the same as the condition on the original data g. In what follows, we rewrite u1 into u and consider the initial-boundary value problem (2.5). If we obtain the following theorem, then the main result Theorem 1.2 also follows: If we obtain the following theorem, then the main result Theorem 1.2 also follows: Theorem 2.1 (Maximal L1-regularity by the Dirichlet boundary data). Let 1 < p < ∞and −1 + 1/p < s ≤0. there exists a unique solution Theorem 2.1 (Maximal L1-regularity by the Dirichlet boundary data). Let 1 < p < ∞and −1 + 1/p < s ≤0. there exists a unique solution u ∈˙W 1,1 R+; ˙Bs p,1(Rn +) , u ∈L1 R+; ˙Bs p,1(Rn +) , to (2.5) if and only if h satisﬁes to (2.5) if and only if h satisﬁes h ∈˙F1−1/2p 1,1 R+; ˙Bs p,1(Rn−1) ∩L1 R+; ˙Bs+2−1/p p,1 (Rn−1) . (2.6) (2.6) Besides the solution u is subject to the estimate: where C is depending only on p, s and n. where C is depending only on p, s and n. When p = ∞, the analogous result holds under arranging the function classes as in Theorem 1.3 with the compatibility condition h(t, x′)\|t=0 = 0. (2.7) (2.7) Similarly, Theorem 1.3 can be reduced into the following: Similarly, Theorem 1.3 can be reduced into the following: Theorem 2.2 (L p-estimate by the Dirichlet boundary data). (1) Let 1 ≤p < ∞. If h satisﬁes h ∈˙F1−1/2p 1,1 R+; L p(Rn−1) ∩L1 R+; B2−1/p p,1 (Rn−1) , then there exists a unique solution u to (2.5) which satisﬁes the following estimate: then there exists a unique solution u to (2.5) which satisﬁes the following estimate: ∥∂tu∥L1(R+;L p(Rn +)) + ∥∇2u∥L1(R+;L p(Rn +)) ≤C ∥h∥˙F1−1/2p 1,1 (R+;L p(Rn−1)) + ∥h∥L1(R+;B2−1/p p,1 (Rn−1)) , where C is depending only on p and n. where C is depending only on p and n. where C is depending only on p and n. (2) For p = ∞, the corresponding result to (1) holds under the analogous arrangement for the function classes as in Theorem 1.3 with imposing the compatibility condition (2.7). (2) For p = ∞, the corresponding result to (1) holds under the analogous arrangement for the function classes as in Theorem 1.3 with imposing the compatibility condition (2.7). Maximal L1-regularity for parabolic Page 15 of 67 30 Maximal L1-regularity for parabolic Page 15 of 67 30 Maximal L1-regularity for parabolic J. Evol. Equ. Page 15 of 67 30 For the case of the Neumann boundary condition, we decompose the problem (1.12) into the following problems and reduce the boundary condition into the case of heat equation. Besides, it holds the estimate: Besides, it holds the estimate: t holds the estimate: ∥∂tu∥L1(R+; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(R+; ˙Bs p,1(Rn +)) ≤C ∥h∥˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥h∥L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) , where C is depending only on p, s and n. where C is depending only on p, s and n. ⎧ ⎪⎨ ⎪⎩ ∂tu1 − u1 = 0, t ∈I, x ∈Rn, u1(t, x) t=0 = u0(x′, xn), xn > 0, u0(x′, −xn), xn ≤0, x ∈Rn, ⎧ ⎨ ⎩ ∂tu2 − u2 = 0, t ∈I, x ∈Rn +, ∂nu2(t, x′, xn) xn=0 = g(t, x′) −∂nu1(t, x′, xn) xn=0, t ∈I, x′ ∈Rn−1, u2(t, x) t=0 = 0, x ∈Rn +, ⎧ ⎪⎨ ⎪⎩ ∂tu3 − 1≤i, j≤n ai j(t, x)∂i∂ju3 ≡F, t ∈I, x ∈Rn +, ∂nu3(t, x′, xn) xn=0 = 0, t ∈I, x′ ∈Rn−1, u3(t, x) t=0 = 0, x ∈Rn +, ⎧ ⎪⎨ ⎪⎩ ∂tu1 − u1 = 0, t ∈I, x ∈Rn, u1(t, x) t=0 = u0(x′, xn), xn > 0, u0(x′, −xn), xn ≤0, x ∈Rn, where F is similarly deﬁned as in (2.3) and f = f (t, x), g = g(t, x′) are given external data and the Neumann boundary condition, respectively, and u0(x) is the initial data. Therefore, the problem can be reduced by setting h(t, x′) = g(t, x′) − ∂nu1(t, x′, xn) xn=0 into the following problem: where F is similarly deﬁned as in (2.3) and f = f (t, x), g = g(t, x′) are given external data and the Neumann boundary condition, respectively, and u0(x) is the initial data. Therefore, the problem can be reduced by setting h(t, x′) = g(t, x′) − ∂nu1(t, x′, xn) xn=0 into the following problem: ⎧ ⎪⎪⎨ ⎪⎪⎩ ∂tu − u = 0, t ∈I, x ∈Rn +, ∂nu(t, x′, xn) xn=0 = h(t, x′), t ∈I, x′ ∈Rn−1, u(t, x) t=0 = 0, x ∈Rn +. (2.8) (2.8) Then, the following result yields our main result for the Neumann problem Theo- rem 1.5. Theorem 2.3 (Maximal L1-regularity by the Neumann boundary data). Let 1 < p < ∞and −1 + 1/p < s ≤0. There exists a unique solution Theorem 2.3 (Maximal L1-regularity by the Neumann boundary data). Let 1 < p < ∞and −1 + 1/p < s ≤0. There exists a unique solution u ∈˙W 1,1 R+; ˙Bs p,1(Rn +) , u ∈L1 R+; ˙Bs p,1(Rn +) , to (2.8) if and only if h ∈˙F1/2−1/2p 1,1 R+; ˙Bs p,1(Rn−1) ∩L1 R+; ˙Bs+1−1/p p,1 (Rn−1) . (2.9) (2.9) Besides, it holds the estimate: where C is depending only on p, s and n. When p = ∞, the analogous result holds under arranging the function classes as in Theorem 1.6. When p = ∞, the analogous result holds under arranging the function classes as in Theorem 1.6. T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. For the L p-estimate in Theorem 1.6 is reduced into the following: For the L p-estimate in Theorem 1.6 is reduced into the following: Theorem 2.4 (L p-estimate by the Neumann boundary data). (1) Let 1 ≤p < ∞. If the boundary data h satisﬁes h ∈˙F1/2−1/2p 1,1 R+; L p(Rn−1) ∩L1 R+; B1−1/p p,1 (Rn−1) , then there exists a unique solution u to (2.8) which fulﬁlls the following estimate: ∥∂tu∥L1(R+;L p(Rn +)) + ∥∇2u∥L1(R+;L p(Rn +)) ≤C ∥h∥˙F1/2−1/2p 1,1 (R+; ˙B0 p,1(Rn−1)) + ∥h∥L1(R+;B1−1/p p,1 (Rn−1)) , where C is depending only on p and n. (2) For p = ∞the corresponding result to (1) holds under the analogous arrangement for the function classes as in Theorem 1.6. where C is depending only on p and n. (2) For p = ∞the corresponding result to (1) holds under the analogous arrangement for the function classes as in Theorem 1.6. Therefore, in order to obtain main results Theorems 1.2–1.3 and Theorems 1.5–1.6, it is enough to show Theorems 2.1–2.2 and Theorems 2.3–2.4, respectively. To show Theorem 2.1, we ﬁrst apply the Laplace transform with respect to t, the partial Fourier transform with respect to x′ and we obtain the solution formula of (2.5) as u(t, x′, xn) = R+ Rn D(t −s, x′ −y′, η)h(s, y′)dy′ds η=xn by using the boundary potential term: D(t, x′, η) = cn−1 2πi Rn−1 eλt+ix′·ξ′λe−√ λ+\|ξ′\|2η dξ′dλ, (2.10) (2.10) where cn−1 = (2π)−(n−1)/2 and is a pass parallel to the imaginary axis. We then extend the boundary data h(t, x′) into t < 0 by the zero extension and it enable us to formulate the above formula by the Fourier transform. The key idea to derive the boundary estimate is applying an almost orthogonal estimate between the boundary potential ψD in (2.10) and the Littlewood-Paley dyadic decompositions of unity in both space x′ and time t variables; {φ j(x′)} j∈Z, {ψk(t)}k∈Z . From (2.10), let η > 0 as a spectral parameter (such as η ≃2−ℓwith ℓ∈Z) and consider D(t, x′, η)η2 = cn−1 2πi Rn−1 eλt+ix′·ξ′λη2e−√ λ+\|ξ′\|2η dξ′dλ and the almost orthogonality between the boundary potential D(t, x′, η)η2 and the Littlewood-Paley dyadic decompositions can be shown. Indeed, by setting and the almost orthogonality between the boundary potential D(t, x′, η)η2 and the Littlewood-Paley dyadic decompositions can be shown. Indeed, by setting D,k, j(t, x′, η) ≡ R Rn−1 D(t −s, x′ −y′, η)ψk(s)φ j(y′)dy′ds (2.11) (2.11) Maximal L1-regularity for parabolic Page 17 of 67 30 Maximal L1-regularity for parabolic Page 17 of 67 30 J. Evol. Equ. Page 17 of 67 30 Page 17 of 67 30 for η > 0, the almost orthogonal property is presented in two-way estimates separated by the time-like estimate and space-like regions. If η ∈[2−ℓ, 2−ℓ+1) for ℓ∈Z, then for η > 0, the almost orthogonal property is presented in two-way estimates separated by the time-like estimate and space-like regions. where C is depending only on p and n. If η ∈[2−ℓ, 2−ℓ+1) for ℓ∈Z, then ∥D,k, j(t, ·, η)η2∥L1(Rn−1 x′ ) ≤ ⎧ ⎨ ⎩ Cn2k−2ℓ 1 + 2(n+2)(k−2ℓ) e−2 1 2 (k−2ℓ)2k⟨2kt⟩−2, k ≥2 j, Cn22 j−2ℓ 1 + 2(n+2)(2 j−2ℓ) e−2 j−ℓ2k⟨2kt⟩−2, k < 2 j, (2.12) (2.12) where ⟨t⟩= (1 + \|t\|2)1/2 which are essentially shown in Lemma 6.1 in Section 6. Then, our main strategy to show maximal L1-regularity for the boundary term (2.5) relies on the above estimate (2.12) and the proof can be complete after exchanging the solution potential D into the Littlewood-Paley dyadic decompositions φ j and ψk. In order to complete such procedure, the above type estimate (2.12) plays a key role. The only difference between the case of Dirichlet boundary condition and the Neumann boundary condition is to the regularity of the boundary data. This is because the solution can be realized by the potential term N(t, x′, η) = −cn−1 2πi Rn−1 eλt+ix′·ξ′ λ λ + \|ξ′\|2 e−√ λ+\|ξ′\|2η dξ′dλ N(t, x′, η) = −cn−1 2πi Rn−1 eλt+ix′·ξ′ λ λ + \|ξ′\|2 e−√ λ+\|ξ′\|2η dξ′dλ nstead of (2.10). instead of (2.10). instead of (2.10). 3. Boundary potentials 3.1. The Dirichlet boundary potential and the compatibility condition In this subsection, we derive the exact solution formula of (2.5). Let h = h(t, x′) be the boundary data extended into t < 0 by the zero extension. We apply the Laplace transform L in time and the Fourier transform in Rn−1-dimensional spatial variables to the equation (2.5). Noting ˆu(0, ξ′, xn) = 0, we obtain that (λ + \|ξ′\|2 −∂2 n) Lu(λ, ξ′, xn) = 0, Lu(λ, ξ′, 0) = Lh(λ, ξ′). Then, it follows that Lu(λ, ξ′, xn) = Lh(λ, ξ′)e−√ λ+\|ξ′\|2xn. Lu(λ, ξ′, xn) = Lh(λ, ξ′)e−√ λ+\|ξ′\|2xn. Hence, the solution is expressed by Hence, the solution is expressed by u(t, x) = cn−1 2πi eλt Rn−1 eix′·ξ′ Lh(λ, ξ′)e−√ λ+\|ξ′\|2xndξ′dλ, (3.1) (3.1) where cn−1 = (2π)−n−1 2 and denotes an integral path in holomorphic domain parallel to the imaginary axis Reλ > 0. The solution (3.1) satisﬁes the heat equation 30 Page 18 of 67 T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu J. Evol. Equ. and the boundary condition u(t, x′, 0) = h(t, x′). From (3.1), the Green function G D(t, x) of the initial-boundary value problem (2.5) is identiﬁed as and the boundary condition u(t, x′, 0) = h(t, x′). From (3.1), the Green function G D(t, x) of the initial-boundary value problem (2.5) is identiﬁed as G D(t, x′, xn) = cn−1 2πi Rn−1 eλteix′·ξ′e−√ λ+\|ξ′\|2xndξ′dλ. (3.2) (3.2) Introducing the Dirichlet boundary potential by D(t, x′, η) = cn−1 2πi Rn−1 eλteix′·ξ′λe−√ λ+\|ξ′\|2ηdξ′dλ, (3.3) (3.3) where we set η = xn, we decompose this boundary potential (3.3) by a combination of two families of the Littlewood-Paley dyadic decomposition of unity. If we put an extra-parameter η ≃2−ℓto D with ℓ∈Z which is a substitution of the boundary parameter xn, then we ﬁnd that Dη2 can be expressed by the time Littlewood-Paley decomposition {ψk(t)}k and the space Littlewood-Paley decomposition {φ j(x′)} j and the relations between the parameters ℓand (k, j) are explicitly estimated. Such es- timates stand for the almost orthogonality between D(t, x′, η)η2 η=2−ℓ ℓ∈Z and {ψk(t), φ j(x′)}k, j∈Z. 3. Boundary potentials Here, we notice that from (3.1)-(3.3), the potential D repre- sents the solution operated by the Laplace operator, namely D(t, x′, η) ≡ G D(t, x′, η). (3.4) (3.4) D(t, x′, η) ≡ G D(t, x′, η). Then, it follows that the potential D represents Then, it follows that the potential D represents u(t, x′, η) = Rn−1 R D(t −s, x′ −y′, η)h(s, y′)dsdy′. If we consider the case when 1 − 1 2p < 1 ρ , then the class F 1−1 2p ρ,p (R+; Lq(Rn−1)) of the boundary data with p > 1 is embedded into the class Cb(I; Lq(Rn−1)) and the data have to satisfy continuity at t = 0 as in Proposition 1.1. On the other hand, if we consider maximal L1 regularity, the class ˙F 1−1 2p 1,1 (R+; ˙Bs p,1(Rn−1)) is not embedded into Cb(R+; ˙Bs p,1(Rn−1)) because 1 −1 2p < 1 under p < ∞, it does not necessarily require continuity at t = 0. This shows that it is not necessary to require the compati- bility condition g(t, x′) t=0 = u0(x′, xn) xn=0 point-wisely when p < ∞. In order to obtain maximal L p regularity for 1 < p < ∞, the boundary data is extended to the zero extension for t < 0. From (3.3), we change the integral path into = γ + ε →ε with ε = Lε ∪Cε, where where Lε = λ = iτ; τ ∈(−∞, ε) ∪(ε, ∞) , Cε = λ = εeiθ; ε > 0, θ : −π 2 →π 2 . (3.5) (3.5) Maximal L1-regularity for parabolic Page 19 of 67 30 Maximal L1-regularity for parabolic Maximal L1-regularity for parabolic Page 19 of 67 30 J. Evol. Equ. Page 19 of 67 30 Maximal L1-regularity for parabolic Page 19 of 6 γ + Γε ε as γ →0 Im Re Figure 1. The integral path γ + ε and ε Figure 1. The integral path γ + ε and ε Then by setting λ ∈γ + ε, Then by setting λ ∈γ + ε, D(t, x′, η) = lim γ →0 cn+1i−1 γ +Lε Rn−1 eλt+ix′·ξ′λe−√ λ+\|ξ′\|2ηdξ′dλ + lim γ →0 cn+1i−1 γ +Cε Rn−1 eλt+ix′·ξ′λe−√ λ+\|ξ′\|2ηdξ′dλ ≡ILε + ICε, (3.6) (3.6) where cn+1 = (2π)−(n+1)/2 (Fig. 1). where cn+1 = (2π)−(n+1)/2 (Fig. 1). 3. Boundary potentials where cn+1 = (2π)−(n+1)/2 (Fig. 1). The ﬁrst term of the right hand side of (3.6) is converging even γ →0 if we observe the real part of the exponential integrant is λe−√ iτ+\|ξ′\|2η = \|τ\| exp −(τ 2 + \|ξ′\|4) 1 4 η cos 1 2 tan−1 τ \|ξ′\|2 (3.7) (3.7) from iτe−√ iτ+\|ξ′\|2η = iτ exp −(τ 2 + \|ξ′\|4) 1 4 ηe i 2 tan−1 τ \|ξ′\|2 = iτ exp −(τ 2 + \|ξ′\|4) 1 4 η cos 1 2 tan−1 τ \|ξ′\|2 + i sin 1 2 tan−1 τ \|ξ′\|2 . We emphasize that for −π 4 < 1 2 tan−1 τ \|ξ′\|2 < π 4 the exponent appeared in (3.7) is negative deﬁnite unless (τ, ξ′) = (0, 0) and hence limε→0 ILε converges. i.e., We emphasize that for −π 4 < 1 2 tan−1 τ \|ξ′\|2 < π 4 the exponent appeared in (3.7) is negative deﬁnite unless (τ, ξ′) = (0, 0) and hence limε→0 ILε converges. i.e., T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. J. Evol. Equ. lim ε→0 ILε = lim ε→0 lim γ →0 cn+1i−1eγ t Lε Rn−1 eiτt+ix′·ξ′(γ + iτ)e−√ γ +iτ+\|ξ′\|2ηdξ′idτ = lim ε→0 cn+1 Lε Rn−1 eiτt+ix′·ξ′iτe−√ iτ+\|ξ′\|2ηdξ′dτ = cn+1 R\{0} Rn−1 eiτt+ix′·ξ′iτe−√ iτ+\|ξ′\|2ηdξ′dτ. (3.8) = cn+1 R\{0} Rn−1 eiτt+ix′·ξ′iτe−√ iτ+\|ξ′\|2ηdξ′dτ. 3. Boundary potentials Page 21 of 67 30 Note that the 0-initial condition for G D(t, x′, η) t=0 = 0 is fulﬁlled by the Cauchy integral theorem on the same complex path (3.5) avoiding the branch cut at the negative real-line and passing the limit γ →0 and ε →0 along the analogous estimates (3.8) and (3.10). Note that the 0-initial condition for G D(t, x′, η) t=0 = 0 is fulﬁlled by the Cauchy integral theorem on the same complex path (3.5) avoiding the branch cut at the negative real-line and passing the limit γ →0 and ε →0 along the analogous estimates (3.8) and (3.10). 3.2. The Neumann boundary potential and the compatibility condition 3. Boundary potentials (3.8) (3.8) For the convergence of the second term ICε of the right hand side of (3.6), we see by noting tan−1 ε sin θ ε cos θ + \|ξ′\|2 ≤ tan−1 sin θ cos θ = \|θ\|, −π 2 ≤θ ≤π 2 (3.9) tan−1 ε sin θ ε cos θ + \|ξ′\|2 ≤ tan−1 sin θ cos θ = \|θ\|, −π 2 ≤θ ≤π 2 (3.9) that for η > 0 (3.9) that for η > 0 that for η > 0 lim ε→0 \|ICε\| ≤lim ε→0 cn+1i−1 Cε Rn−1 eλt+ix′·ξ′λe−√ λ+\|ξ′\|2ηdξ′dλ ≤lim ε→0 cn+1 Rn−1 π 2 −π 2 eεteiθ εeiθe−√ εeiθ +\|ξ′\|2ηεieiθdθ dξ′ = lim ε→0 cn+1 Rn−1 π 2 −π 2 eεt cos θε2η2 × exp −η (ε cos θ + \|ξ′\|2)2 + ε2 sin2 θ 1/4 cos 1 2 tan−1 ε sin θ ε cos θ + \|ξ′\|2 dθdξ′ lim ε→0 \|ICε\| ≤lim ε→0 cn+1i−1 ≤lim ε→0 cn+1 Rn−1 π 2 −π 2 eεt cos θε2 exp −η (ε cos θ + \|ξ′\|2)2 1/4 cos θ 2 dθdξ′ ≤lim ε→0 cn+1ε2 Rn−1 exp − √ 2 2 η\|ξ′\| dξ′ π 2 −π 2 eεt cos θdθ ≤cn+1 lim ε→0 ε2 Rn−1 exp − √ 2 2 η\|ξ′\| dξ′ 1 −1 eεt\|ζ\| 1 −ζ 2 dζ = 0. (3.10) ≤cn+1 lim ε→0 ε2 Rn−1 exp − √ 2 2 η\|ξ′\| dξ′ 1 −1 eεt\|ζ\| 1 −ζ 2 dζ = 0. (3.10) (3.10) Therefore, by passing ε →0 in (3.6), we obtain the following formula from (3.8) and (3.10): D(t, x′, η) = cn+1 R Rn−1 eitτ+ix′·ξ′iτe−√ iτ+\|ξ′\|2ηdξ′dτ. (3.11) (3.11) Remark. We note that from (3.1) and similar way in (3.3)-(3.10), the Green’s function (3.2) is expressed by the Fourier inverse transform as Remark. We note that from (3.1) and similar way in (3.3)-(3.10), the Green’s function (3.2) is expressed by the Fourier inverse transform as G D(t, x′, η) = cn+1 R Rn−1 eitτ+ix′·ξ′e−√ iτ+\|ξ′\|2ηdξ′dτ. Therefore, the potential function D given by (3.11) is understood as the Green’s function operated by the Laplacian; D(t, x′, η) = (x′,η)G D(t, x′, η). Maximal L1-regularity for parabolic J. Evol. Equ. 3.2. The Neumann boundary potential and the compatibility condition 3.2. The Neumann boundary potential and the compatibility condition Following the method of the case of the Dirichlet boundary condition, we consider the initial-boundary value problem (2.8). Then, we deduce the boundary potential (the Green’s function) that yields the solution to (2.8). We apply the Laplace transform in time and the Fourier transform in Rn−1-dimensional spatial variables and noting Lu(0, ξ′, xn) = 0 to have (λ + \|ξ′\|2 −∂2 n) Lu = 0, ∂n Lu(λ, ξ′, 0) = Lh(λ, ξ′). ∂n Lu(λ, ξ′, 0) = Lh(λ, ξ′). Then, we write it explicitly, Then, we write it explicitly, Then, we write it explicitly, u(t, x) = −cn−1 2πi eλt Rn−1 eix′·ξ′ Lh(λ, ξ′) λ + \|ξ′\|2 e−√ λ+\|ξ′\|2xndξ′dλ, (3.12) where cn−1 = (2π)−n−1 2 and is a proper path on the analytic region. From (3.12), the Green’s function G N(t, x) of the initial-boundary value problem (2.8) is given by G N(t, x′, xn) = −cn−1 2πi Rn−1 eλteix′·ξ′ 1 λ + \|ξ′\|2 e−√ λ+\|ξ′\|2xndξ′dλ. (3.13) One can choose as the parallel line to the imaginary axis in Reλ > 0. Let N(t, x′, η) = −cn−1 2πi Rn−1 eλteix′·ξ′ λ λ + \|ξ′\|2 e−√ λ+\|ξ′\|2ηdξ′dλ, (3.14) where is a proper integral path basically parallel to the imaginary axis. For the case of maximal L1-regularity, the boundary regularity is ˙F 1 2 −1 2p 1,p I; L p(Rn−1) and the continuity in time direction does not hold; hence, the compatibility condition is redundant. Along the Dirichlet boundary case before, the boundary condition can be prolonged to t < 0 by zero extension and hence the solution is understood by zero extension. From (3.14), we pass the integral path (3.5) into = γ + ε →ε with ε = Lε ∪Cε by γ →0. Then, T. Ogawa and S. Shimizu J. Evol. E T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. N(t, x′, η) = lim γ →0 cn+1i γ +Lε Rn−1 eλt+ix′·ξ′ λ λ + \|ξ′\|2 e−√ λ+\|ξ′\|2ηdξ′dλ + lim γ →0 cn+1i γ +Cε Rn−1 eλt+ix′·ξ′ λ λ + \|ξ′\|2 e−√ λ+\|ξ′\|2ηdξ′dλ ≡I ILε + I ICε. 3.2. The Neumann boundary potential and the compatibility condition (3.15) (3.15) The ﬁrst term of the right hand side of (3.15) is converging even γ →0 if we observe the real part of the exponential integrant is for the case γ = 0 that The ﬁrst term of the right hand side of (3.15) is converging even γ →0 if we observe the real part of the exponential integrant is for the case γ = 0 that iτ iτ + \|ξ′\|2 = iτe i 2 tan−1 −τ \|ξ′\|2 (τ 2 + \|ξ′\|4) 1 4 , (3.16) (3.16) and (3.3). Then, we obtain that and (3.3). Then, we obtain that iτ iτ + \|ξ′\|2 e−√ iτ+\|ξ′\|2η = \|τ\| (τ 2 + \|ξ′\|4) 1 4 exp −(τ 2 + \|ξ′\|4) 1 4 η cos 1 2 tan−1 τ \|ξ′\|2 . (3.17 (3.17) Under the condition on the argument −π 4 < 1 2 tan−1 τ \|ξ′\|2 < π 4 the real part of the exponential function in (3.16) is negative unless (τ, ξ) = (0, 0). We observe that for −π 4 < 1 2 tan−1 τ \|ξ′\|2 < π 4 (3.17) is integrable unless (τ, ξ′) = (0, 0) and hence I ILε converges. i.e., lim ε→0 I ILε = −cn+1 R\{0} Rn−1 eiτt+ix′·ξ′ iτ iτ + \|ξ′\|2 e−√ iτ+\|ξ′\|2ηdξ′dτ. (3.18) (3.18) For the convergence of the second term ICε of the right hand side of (3.15), we see by noting (3.9) and η > 0 that For the convergence of the second term ICε of the right hand side of (3.15), we see by noting (3.9) and η > 0 that lim ε→0 I ICε ≤lim ε→0 cn+1 Rn−1 π 2 −π 2 eεeiθ εeiθ εeiθ + \|ξ′\|2 e−√ εeiθ +\|ξ′\|2ηεieiθdθ dξ′ ≤lim ε→0 cn+1 Rn−1 π 2 −π 2 eε cos θt ε2 (ε cos θ + \|ξ′\|2)2 + ε2 sin2 θ 1/4 × exp −η (ε cos θ + \|ξ′\|2)2 + ε2 sin2 θ 1/4 cos 1 2 tan−1 ε sin θ ε cos θ + \|ξ′\|2 dθdξ′ ≤lim ε→0 cn+1 Rn−1 π 2 −π 2 eε cos θt ε2 ε cos θ + \|ξ′\|2 1/2 exp −η\|ξ′\| cos θ 2 dθdξ ≤cn+1 lim ε→0 ε 3 2 Rn−1 exp − √ 2 2 η\|ξ′\| dξ′ 1 −1 eεt\|ζ\| √\|ζ\| 1 −ζ 2 dζ = 0. 3.2. The Neumann boundary potential and the compatibility condition (3.1 Then, it follows from (3.18) and (3.19) that Then, it follows from (3.18) and (3.19) that Then, it follows from (3.18) and (3.19) that N(t, x′, η) = −cn+1 R Rn−1 eitτ+ix′·ξ′ iτ iτ + \|ξ′\|2 e−√ iτ+\|ξ′\|2ηdξ′dτ. (3.20) Maximal L1-regularity for parabolic Page 23 of 67 30 Maximal L1-regularity for parabolic Page 23 of 67 30 J. Evol. Equ. J. Evol. Equ. We notice that the potential of the solution operated by the Laplace operator is given by by N(t, x′, η) ≡ G N(t, x′, η), (3.21) (3.21) where where G N(t, x′, xn) = −cn+1 R Rn−1 eitτ+ix′·ξ′ 1 iτ + \|ξ′\|2 e−√ λ+\|ξ′\|2ηdξ′dτ. (3.22) We then regard this boundary potential as a role of Littlewood-Paley dyadic decom- position of unity and the main argument for the proof consists on exchanging the boundary potential into the standard Littlewood-Paley decomposition {ψk}k∈Z and {φ j} j∈Z. 4.1. The Besov spaces on the half-spaces 4.1. The Besov spaces on the half-spaces 4.1. The Besov spaces on the half-spaces First, we recall the summary for the Besov spaces over the half-space on the Eu- clidean space Rn +. Deﬁnition. Let 1 ≤p < ∞and 1 ≤σ < ∞with s ≥0. Let ◦ Bs p,σ(Rn +) = C∞ 0 (Rn +) ˙Bs p,σ (Rn +), ⊙ Bs p,σ(Rn +) = { f ∈˙Bsp,σ(Rn); supp f ⊂Rn +} ˙Bsp,σ (Rn) , by the Besov norm ˙Bs p,σ(Rn +) (see Triebel [46] Section 2.9.3). It is shown that the above-deﬁned space coincides the space ˙Bs p,σ(Rn +) deﬁned by the restriction in (1.5). Namely, the following proposition is shown by Triebel [46] and Danchin–Mucha [12]. by the Besov norm ˙Bs p,σ(Rn +) (see Triebel [46] Section 2.9.3). It is shown that the above-deﬁned space coincides the space ˙Bs p,σ(Rn +) deﬁned by the restriction in (1.5). Namely, the following proposition is shown by Triebel [46] and Danchin–Mucha [12]. Proposition 4.1 ([12,46]). Let 1 < p < ∞. Proposition 4.1 ([12,46]). Let 1 < p < ∞. Let 1 ≤p < ∞, 1 ≤σ < ∞and −1 + 1 p < s < 1 p. It holds that R0 : ˙Bs p,σ(Rn) →˙Bs p,σ(Rn +), (4.1) E0 : ˙Bs p,σ(Rn +) →˙Bs p,σ(Rn), (4.2) (4.1) are linear bounded operators. Besides, it holds that Proposition 4.1 ([12,46]). Let 1 < p < ∞. Proposition 4.1 ([12,46]). Let 1 < p < ∞. (1) For 0 < s, 1 ≤σ < ∞, (1) For 0 < s, 1 ≤σ < ∞, ˙B−s p′,σ ′(Rn +) ≃ ◦ Bs p,σ(Rn +) ∗. (2) For −∞< s ≤1 p and for 1 < σ < ∞, (2) For −∞< s ≤1 p and for 1 < σ < ∞, ◦ Bs p,σ(Rn +) ≃˙Bs p,σ(Rn +). (3) For −∞< s < 1 p and σ = 1, ◦ Bs p,1(Rn +) ≃˙Bs p,1(Rn +). 30 Page 24 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 24 of 67 T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. (4) For −1 + 1 p < s < 1 p and 1 ≤σ < ∞, ⊙ Bs p,σ(Rn +) ≃˙Bs p,σ(Rn +). (4) For −1 + 1 p < s < 1 p and 1 ≤σ < ∞, (4) For −1 + 1 p < s < 1 p and 1 ≤σ < ∞, ⊙ Bs p,σ(Rn +) ≃˙Bs p,σ(Rn +). ⊙ Bs p,σ(Rn +) ≃˙Bs p,σ(Rn +). We consider the restriction operator R0 by multiplying a cutoff function χRn +(x) = 1, in Rn +, 0, in Rn −. i.e., for f ∈˙Bs p,σ(Rn), set R0 f = χRn +(x) f (x) in ˙Bs p,σ(Rn) if s > 0 and it is understood as a distribution sense. Let the extension operator E0 from ⊙ Bs p,σ(Rn +) given by the zero-extention, i.e., for any f ∈ ⊙ Bs p,σ(Rn +), set E0 f = f (x), in Rn +, 0, in Rn −. E0 f = f (x), in Rn +, 0, in Rn −. One can ﬁnd that those operators are basic tool to recognize the homogeneous Besov spaces. Using Proposition 4.1, the following statement is a variant introduced by Triebel [46, p.228] Proposition 4.2. Let 1 ≤p < ∞, 1 ≤σ < ∞and −1 + 1 p < s < 1 p. It holds that R0 : ˙Bs p,σ(Rn) →˙Bs p,σ(Rn +), (4.1) E0 : ˙Bs (Rn +) →˙Bs (Rn) (4 2) roposition 4.2. Let 1 ≤p < ∞, 1 ≤σ < ∞and −1 + 1 p < s < 1 p. It holds that Proposition 4.2. are linear bounded operators. Besides, it holds that R0E0 = Id : ˙Bs p,σ(Rn +) →˙Bs p,σ(Rn +), (4.3) (4.3) where Id denotes the identity operator. Namely, E0 and R0 are a retraction and a co-retraction, respectively. where Id denotes the identity operator. Namely, E0 and R0 are a retraction and a co-retraction, respectively. The proof of Proposition 4.2 is along the same line of the proof in [46]. Note that the spaces are homogeneous Besov spaces and then the arrangement appears in Proposition 3 in Danchin–Mucha [12] is required. Proof of Proposition 4.2. It is clear that both Proof of Proposition 4.2. It is clear that both R0 : ˙Bs p,σ(Rn) →˙Bs p,σ(Rn +), E0 : ˙Bs p,σ(Rn +) →˙Bs p,σ(Rn) (4.4) (4.4) are linear operators and are linear operators and R0E0 = Id : ˙Bs p,σ(Rn +) →˙Bs p,σ(Rn +), Maximal L1-regularity for parabolic Page 25 of 67 30 J. Evol. Equ. Maximal L1-regularity for parabolic Page 25 of 67 30 J. Evol. Equ. Page 25 of 67 30 then it sufﬁcient to show that both operators are bounded. Then, they are retraction and co-retraction, respectively. ˙ then it sufﬁcient to show that both operators are bounded. Then, they are retraction and co-retraction, respectively. To see the ﬁrst operator (4.1) is bounded, let f ∈˙Bs p,σ(Rn) and we show that ∥R0 f ∥˙Bsp,σ (Rn +) = inf ∥ R0 f ∥˙Bsp,σ (Rn) ≤∥χRn + f ∥˙Bsp,σ (Rn) ≤∥f ∥˙Bsp,σ (Rn) is valid under the restriction s > 0. Let −1/p′ < s < 0 and f ∈˙Bs′ p,σ(Rn). Suppose that φ ∈C∞ 0 (Rn), Then ∥R0 f ∥˙Bsp,σ (Rn +) = inf ∥ R0 f ∥˙Bsp,σ (Rn) ≤∥χRn + f ∥˙Bsp,σ (Rn) ≤∥f ∥˙Bsp,σ (Rn) lid d h i i 0 L 1/ ′ 0 d f ˙ s′ (Rn) S ∥R0 f ∥˙Bsp,σ (Rn +) = inf ∥R0 f ∥˙Bsp,σ (Rn) ≤∥χRn + f ∥˙Bsp,σ (Rn) ≤∥f ∥˙Bsp,σ (Rn) is valid under the restriction s > 0. Let −1/p′ < s < 0 and f ∈˙Bs′ p,σ(Rn). Suppose that φ ∈C∞ 0 (Rn), Then ⟨χRn + f, φ⟩ = ⟨f, χRn +φ⟩ ≤∥f ∥˙Bsp,σ (Rn)∥χRn +φ∥˙B−s p′,σ′(Rn) ≤∥f ∥˙Bsp,σ (Rn)∥φ∥˙B−s p′,σ′(Rn), (4.5) (4.5) since 0 < −s < 1/p′ and the last inequality follow from the pointwise sense. are linear bounded operators. Besides, it holds that Thus from the deﬁnition of the norm in ˙Bs p,σ(Rn +), it holds similarly to the above that since 0 < −s < 1/p′ and the last inequality follow from the pointwise sense. Thus from the deﬁnition of the norm in ˙Bs p,σ(Rn +), it holds similarly to the above that ∥R0 f ∥˙Bsp,σ (Rn +) ≤∥χRn + f ∥˙Bsp,σ (Rn) = sup φ∈˙B−s p′,σ′(Rn)\{0} ⟨χRn + f, φ⟩ ∥φ∥˙B−s p′,σ′(Rn) ≤∥f ∥˙Bsp,σ (Rn). To see the second bound (4.2), we introduce a quotient space ˙Bs p,σ(Rn)/ ∼, where we identify all f ∈˙Bs p,σ(Rn) that coincides on Rn +. Then, the restriction operator R0 is one to one mapping from ˙Bs p,σ(Rn)/ ∼onto ˙Bs p,σ(Rn +), and then the extension operator E0 is an inverse operator of R0. Thus, the open mapping theorem implies the required boundedness directly. □ In what follows, we restrict ourselves to the regularity range of the Besov spaces ˙Bs p,σ(Rn +) in −1 + 1/p < s < 1/p for 1 < p < ∞. Hence, we freely use the above-mentioned results. As a consequence, any component in ˙Bs p,σ(Rn +) under such restriction on s and p can be extended into the one over whole space Rn and conversely the component f ∈˙Bs p,σ(Rn) is restricted into the one over the half-space Rn +. We frequently use those facts without noticing for every case below. 4.2. The L-P decomposition with separation of variables In order to split the variables x′ ∈Rn−1 and xn ∈R+, we introduce an x′-parallel decomposition and an xn-parallel decomposition by Littlewood-Paley type. In what follows, η ∈R+ denotes a parameter for xn-axis in Rn +. We introduce {m}m∈Z as a Littlewood-Paley dyadic frequency decomposition of unity in separated variables (ξ′, ξn) ∈Rn−1 × R. Deﬁnition (The Littlewood-Paley decomposition of separated variables). For m ∈Z, let let let ζm(ξn) = ⎧ ⎨ ⎩ 1, 0 ≤\|ξn\| ≤2m, smooth, 2m < \|ξn\| < 2m+1, 0, 2m+1 ≤\|ξn\|, ζm(ξn) = ζm−1(ξn) + φm(ξn) (4.6) (4.6) 30 Page 26 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. f 67 Figure 2. The support of Littlewood-Paley decomposition {m}m∈Z Figure 2. are linear bounded operators. Besides, it holds that The support of Littlewood-Paley decomposition {m}m∈Z (one can choose ζm(r) = ℓ≤m−1 φℓ(r) + φ−∞(r) with a correction distribution φ−∞(r) at r = 0) and set (one can choose ζm(r) = ℓ≤m−1 φℓ(r) + φ−∞(r) with a correction distribution φ−∞(r) at r = 0) and set (one can choose ζm(r) = ℓ≤m−1 φℓ(r) + φ−∞(r) with a correction distribution φ−∞(r) at r = 0) and set m(ξ) ≡ φm(\|ξ′\|) ⊗ ζm−1(ξn) + ζm(\|ξ′\|) ⊗ φm(ξn). (4.7) m(ξ) ≡ φm(\|ξ′\|) ⊗ ζm−1(ξn) + ζm(\|ξ′\|) ⊗ φm(ξn). (4.7) (4.7) Then, it is obvious from Fig. 2 (restricted on the upper half region in Rn) that Then, it is obvious from Fig. 2 (restricted on the upper half region in Rn) that m∈Z m(ξ) ≡1, ξ = (ξ′, ξn) ∈Rn \ {0}. (4.8) d d f 6 d m∈Z m(ξ) ≡1, ξ = (ξ′, ξn) ∈Rn \ {0}. (4.8) (4.8) Indeed, from (4.6) and (4.7), Indeed, from (4.6) and (4.7), Indeed, from (4.6) and (4.7), m∈Z m(ξ) = m∈Z φm(\|ξ′\|) ⊗ −∞≤ℓ≤m−1 φℓ(ξn) + m∈Z −∞≤ℓ≤m φℓ(\|ξ′\|) ⊗ φm(ξn) = m∈Z φm(\|ξ′\|) ⊗ −∞≤ℓ≤m−1 φℓ(ξn) + m∈Z ℓ≤m φℓ(\|ξ′\|) ⊗ φm(ξn) + m∈Z φ−∞(\|ξ′\|) ⊗ φm(ξn) = m∈Z φm(\|ξ′\|) ⊗ −∞≤ℓ≤m−1 φℓ(ξn) + ℓ∈Z φℓ(\|ξ′\|) ⊗ m≥ℓ φm(ξn) + m∈Z φ−∞(\|ξ′\|) ⊗ φm(ξn) = m∈Z φm(\|ξ′\|) ⊗ −∞≤ℓ≤m−1 φℓ(ξn) + ℓ≥m φℓ(ξn) + φ−∞(\|ξ′\|) ⊗ m∈Z φm(ξn) = m∈Z φm(\|ξ′\|) ⊗ ℓ∈Z∪{−∞} φℓ(ξn) + φ−∞(\|ξ′\|) ⊗ ℓ∈Z∪{−∞} φm(ξn) − φ−∞(\|ξ′\|) ⊗ φ−∞(ξn) m∈Z J. Evol. Equ. Maximal L1-regularity for parabolic Page 27 of 67 30 Maximal L1-regularity for parabolic Page 27 of 67 30 Maximal L1-regularity for parabolic Page 27 of 67 30 Page 27 of 67 30 J. Evol. Equ. J. Evol. Equ. = m∈Z∪{−∞} φm(\|ξ′\|) ⊗1 − φ−∞(\|ξ′\|) ⊗ φ−∞(ξn) = 1 − φ−∞(\|ξ′\|) ⊗ φ−∞(ξn). Deﬁnition (VarietiesoftheLittlewood-Paleydyadicdecompositions).Let(τ, ξ′, ξn) ∈ R×Rn−1 ×R be Fourier adjoint variables corresponding to (t, x′, η) ∈R+ ×Rn−1 × R+. Deﬁnition (VarietiesoftheLittlewood-Paleydyadicdecompositions).Let(τ, ξ′, ξn) ∈ R×Rn−1 ×R be Fourier adjoint variables corresponding to (t, x′, η) ∈R+ ×Rn−1 × R+. • {m(x)}m∈Z: the standard (annulus type) Littlewood-Paley dyadic decomposi- tion by x = (x′, η) ∈Rn +. x = (x′, η) ∈Rn +. are linear bounded operators. Besides, it holds that • {m(x)}m∈Z: the Littlewood-Paley dyadic decomposition over x = (x′, η) ∈ Rn + given by (4.7). • {ψk(˜t)}k∈Z: the Littlewood-Paley dyadic decompositions in ˜t ∈R. • {φ j(x′)} j∈Z and{φ j(˜η)} j∈Z:thestandard(annulustype)Littlewood-Paleydyadic decompositions in x′ ∈Rn−1 and ˜η ∈R, respectively. • {ζm(x′)}m∈Z and {ζm(˜η)}m∈Z: the lower frequency smooth cutoff given by (4.6), respectively. • Let φ j = φ j−1 + φ j + φ j+1 be the Littlewood-Paley dyadic decompositions with its j-neighborhood. • Since all the above-deﬁned decompositions are even functions, we identify ˜t ∈R and ˜η ∈R with \|˜t\| = t > 0 and \|˜η\| = η > 0, respectively. Then, it is easy to see that the norm of the Besov spaces on Rn + deﬁned by {m}m is equivalent to the one from the Littlewood-Paley decomposition of direct sum type, {m}m: ∥ u(t)∥˙Bs p,1(Rn +) = m∈Z 2sm∥m ∗ (x) \|m−k\|≤1 k ∗ (x) u(t)∥L p(Rn +) ≤3C m∈Z 2sm∥m ∗ (x) u(t)∥L p(Rn +). (4.9) ∥ u(t)∥˙Bs p,1(Rn +) = m∈Z 2sm∥m ∗ (x) \|m−k\|≤1 k ∗ (x) u(t)∥L p(Rn +) ≤3C m∈Z 2sm∥m ∗ (x) u(t)∥L p(Rn +). (4.9) 4.3. Separation on the Dirichlet potential (4.9) 4.3. Separation on the Dirichlet potential 4.3. Separation on the Dirichlet potential 4.3. Separation on the Dirichlet potential In order to show the sufﬁciency part of Theorem 2.1, we ﬁrst decompose the solution potential D deﬁned in (3.11). From the solution formula with the Green function (3.4) and (3.11), we see that m ∗ (x′,η) D(t, x′, η) = φm(\|x′\|) ⊗ζm−1(η) ∗ (x′,η) D(t, x′, η) + ζm(\|x′\|) ⊗φm(η) ∗ (x′,η) D(t, x′, η) = ζm−1(η) ∗ (η) φm(\|x′\|) ∗ (x′) D(t, x′, η) + φm(η) ∗ (η) ζm(\|x′\|) ∗ (x′) D(t, x′, η) = φm(\|x′\|) ⊗ζm−1(η) ∗ (x′,η) D(t, x′, η) + ζm(\|x′\|) ⊗φm(η) ∗ (x′,η) D(t, x′, η) = ζm−1(η) ∗ (η) φm(\|x′\|) ∗ (x′) D(t, x′, η) + φm(η) ∗ (η) ζm(\|x′\|) ∗ (x′) D(t, x′, η) T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. = ζm−1(η) ∗ (η) cn+1 cn−1 R Rn−1 eitτ+ix′·ξ′ φm(\|ξ′\|) iτ e−√ iτ+\|ξ′\|2ηdξ′dτ + φm(η) ∗ (η) cn+1 cn−1 R Rn−1 eitτ+ix′·ξ′ ζm(\|ξ′\|) iτ e−√ iτ+\|ξ′\|2ηdξ′dτ. (4.10) (4.10) To estimate the solution by the Besov-norm, we involve the Littlewood-Paley dyadic decomposition of the direct sum type (4.7). Besides, ζm(η)∗can be treated by the Hausdorff–Young inequality of η-variable, i.e., the L p(Rn +) norm of the ﬁrst term of the right hand side of (4.10) is estimated as follows: ζm−1 ∗ (η) φm(x′) ∗ (x′) D(t, x′, η) L p(R+,η;L p(Rn−1 x′ )) ≤∥ζm−1∥L1(R+,η) φm(x′) ∗ (x′) D(t, x′, η) L p(R+,η;L p(Rn−1 x′ )) = C φm(x′) ∗ (x′) D(t, x′, η) L p(R+,η;L p(Rn−1 x′ )). 4.3. Separation on the Dirichlet potential Applying the Hausdorff–Young inequality to the ﬁrst term of the right hand side of (4.10) and restrict the range of the summation of j by ζm(x′) ∗ (x′), we have from (4.9) that Applying the Hausdorff–Young inequality to the ﬁrst term of the right hand side of (4.10) and restrict the range of the summation of j by ζm(x′) ∗ (x′), we have from (4.9) that ∞ 0 ∥ u(t)∥˙Bs p,1(Rn +)dt ≤C m∈Z 2smm ∗ (x′,η) D(t, x′, η) ∗ (t,x′) h(t, x′) L p(Rn +) L1t (R+) ≤C m∈Z 2sm R+ D(t, x′, η) ∗ (t,x′) φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) + C m∈Z 2sm R+ φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) 0 ∥ u(t)∥Bs p,1(Rn +)dt ≤C m∈Z 2smm ∗ (x′,η) D(t, x′, η) ∗ (t,x′) h(t, x′) L p(Rn +) L1t (R+) ≤C m∈Z 2sm R+ D(t, x′, η) ∗ (t,x′) φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) ′ × k∈Z j≤m+1 ψk(t) ∗ (t) φ j(y′) ∗ (y′) h(t, y′) p L p(Rn−1 x′ )dη 1/p L1t (R+) (4.11) ≡∥P D 1 ∥L1t (R+) + ∥P D 2 ∥L1t (R+), ≡∥P D 1 ∥L1t (R+) + ∥P D 2 ∥L1t (R+), (4.11) where the ﬁrst term of the right hand side of (4.11) includes φm(x′), once the outer decomposition m∈Z is ﬁxed then the inner decomposition {φ j(x′)∗} j∈Z is restricted into only \| j −m\| ≤1 and the summation for j disappears. 4.4. Time-space splitting argument 4.4. Time-space splitting argument We separate the estimate of (4.11) into two regions ; one is time-dominated area and the other is space-dominated area (Fig. 3). • The relation between each variables: • The relation between each variables: • The relation between each variables: In order to prove Theorem 2.1, it is enough to prove Lemma 4.3. emma 4.3. Let 1 < p ≤∞. The term P D 1 deﬁned in (4.11) is estimated as follows: Lemma 4.3. Let 1 < p ≤∞. The term P D 1 deﬁned in (4.11) is estimated as follows: Lemma 4.3. Let 1 < p ≤∞. 4.3. Separation on the Dirichlet potential The term P D 1 deﬁned in (4.11) is estimated as follows: ∥P D 1 ∥L1t (R+) ≤C h ˙F1−1/2p 1 1 (R+; ˙Bs p 1(Rn−1)) + h L1(R+; ˙Bs+2−1/p p 1 (Rn−1)) . (4.12) ∥P D 1 ∥L1t (R+) ≤C h ˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + h L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) . (4.12) Page 29 of 67 30 Maximal L1-regularity for parabolic J. Evol. Equ. 0 \|ξ 2j, j ∈Z τ 2k, k ∈Z Figure 3. The time-space splitting Figure 3. The time-space splitting Simultaneously the term P D 2 deﬁned in (4.11) is estimated as follows: ∥P D 2 ∥L1t (R+) ≤C h ˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + h L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) . (4.13) (4.13) Proof of Lemma 4.3. We split the data h into the time-dominated region and the space- dominated region. h(t, x′) = k∈Z j∈Z ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) = k∈Z 2 j≤k ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′)+ k∈Z 2 j>k ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′). (4.14) Z ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) k ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′)+ k∈Z 2 j>k ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′). (4.14) Letting hm(t, x′) ≡ φm ∗ (x′) h(t, x′) (m ∈Z), we proceed Letting hm(t, x′) ≡ φm ∗ (x′) h(t, x′) (m ∈Z), we proceed P D 1 = C m∈Z 2sm R+ k∈Z \| j−m\|≤1,2 j<k D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) × φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p + C m∈Z 2sm R+ j∈Z \| j−m\|≤1,k<2 j D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) × φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p ≤C m∈Z 2sm R+ k≥2m D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) hm(t, x′) p L p x′ dη 1/p + C m∈Z 2sm R+ k<2m D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) hm(t, x′) p L p x′ dη 1/p ≡L1 + L2. 4.3. Separation on the Dirichlet potential (4.15) P D 1 = C m∈Z 2sm R+ k∈Z \| j−m\|≤1,2 j<k D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) × φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p + C m∈Z 2sm R+ j∈Z \| j−m\|≤1,k<2 j D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) × φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p ≤C m∈Z 2sm R+ k≥2m D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) hm(t, x′) p L p x′ dη 1/p + C m∈Z 2sm R+ k<2m D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) hm(t, x′) p L p x′ dη 1/p ≡L1 + L2. (4.15) k D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) (4.15) T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. Setting D,k,m(t, x′, η) ≡ R Rn−1 D(t −s, x′ −y′, η)ψk(s)φm(y′)dy′ds, (4.16) (4.16) we see that L1 is the time-dominated region and applying the Minkowski and the Hausdorff–Young inequality with using (4.16), we have we see that L1 is the time-dominated region and applying the Minkowski and the Hausdorff–Young inequality with using (4.16), we have L1 ≤C m∈Z 2sm R+ k≥2m R+ Rn−1 D,k,m(t −s, x′ −y′, η) × ψk(s) ∗ (s) hm(s, y′) dy′ L p x′ ds p dη 1/p L1 ≤C m∈Z 2sm R+ k≥2m R+ Rn−1 D,k,m(t −s, x′ −y′, η) × ψk(s) ∗ (s) hm(s, y′) dy′ L p x′ ds p dη 1/p ≤C m∈Z 2sm R+ k≥2m R+ D,k,m(t −s, ·, η) L1 x′ ψk(s) ∗ (s) hm(s, ·) L p x′ ds p dη 1/p . (4.17) ≤C m∈Z 2sm R+ k≥2m R+ D,k,m(t −s, ·, η) L1 x′ ψk(s) ∗ (s) hm(s, ·) L p x′ ds p dη 1/p . 4.3. Separation on the Dirichlet potential (4.17) (4.17) Then by the almost orthogonal estimate between the boundary potential D and the Littlewood-Paley decomposition ψk in time, namely we invoke Lemma 6.1, for any t, η ∈R+, D,k,m(t, ·, η) L1 x′ ≤ ⎧ ⎪⎪⎪⎨ ⎪⎪⎪⎩ C2k exp −2 k 2 −1η 2k ⟨2kt⟩2 , k ≥2m, C2k exp −2m−1η 2k ⟨2kt⟩2 , k < 2m. (4.18) (4.18) Noting the restriction k ≥2m on the time-dominated region, we apply (4.18) to (4.17) and obtain that Noting the restriction k ≥2m on the time-dominated region, we apply (4.18) to (4.17) and obtain that L1 L1t (R+) t ≤C m∈Z 2sm R+ k≥2m 2k exp(−2 k 2 −1η) R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds p dη 1/p L1t (R+) = C m∈Z 2sm k≥2m 2k R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds R+ exp(−p2 k 2 −1η)dη 1/p L1t (R+) ≤C m∈Z 2sm k∈Z 2(1−1 2p )k R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k m∈Z 2sm R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k m∈Z 2sm ψk ∗ (s) hm(s, ·) L p x′ L1t (R+) = C k∈Z 2(1−1 2p )k m∈Z 2smψk ∗ (s) hm(s, ·) L p x′ L1t (R+) = C h ˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1 x′ )). 1Up to this level, there is no restriction on p nor s. 4.3. Separation on the Dirichlet potential (4.19) ≤C m∈Z 2sm R+ k≥2m 2k exp(−2 k 2 −1η) R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds p dη 1/p L1t (R+) = C m∈Z 2sm k≥2m 2k R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds R+ exp(−p2 k 2 −1η)dη 1/p L1t (R+) ≤C m∈Z 2sm k∈Z 2(1−1 2p )k R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k m∈Z 2sm R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) hm(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k m∈Z 2sm ψk ∗ (s) hm(s, ·) L p x′ L1t (R+) = C k∈Z 2(1−1 2p )k m∈Z 2smψk ∗ (s) hm(s, ·) L p x′ L1t (R+) = C h ˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1 x′ )). (4.19 (4.19) Meanwhile,thesecondterm L2 isthespace-dominatedregionandlettinghm(t, x′) ≡ φm ∗h(t, x′), we apply again the Minkowski inequality, the Hausdorff–Young inequal- ity and the estimate (4.16), Maximal L1-regularity for parabolic Maximal L1-regularity for parabolic Page 31 of 67 30 Maximal L1-regularity for parabolic Page 31 of 67 30 Page 31 of 67 30 J. Evol. Equ. L2 L1t (R+) L2 L1t (R+) t ( +) ≤C m∈Z 2sm η∈R+ k<2m 2ke−2m−1η R 2k ⟨2k(t −s)⟩2 hm(s, ·) L p x′ ds p dη 1/p L1t (R+) = C m∈Z 2sm k<2m 2k R 2k ⟨2k(t −s)⟩2 hm(s, ·) L p x′ ds η∈R+ exp(−p2m−1η)dη 1/p L1t (R+) ≤C m∈Z 2sm+2m−m p k<2m 2k−2m R 2k ⟨2k(t −s)⟩2 hm(s, ·) L p x′ ds L1t (R+) ≤C m∈Z 2sm+2m−m p k<2m 2k−2m R 2k ⟨2k(t −s)⟩2 hm(s, ·) L p x′ ds L1t (R+) ≤C m∈Z 2(s+2−1 p )m k<2m 2k−2m hm(t, ·) L p x′ L1t (R+) ≤C h L1(R+; ˙Bs+2−1/p p,1 (Rn−1 x′ )). (4.20) (4.20) From (4.15), (4.19) and (4.20), the estimate (4.12) is shown.1 We then prove (4.13). 4.3. Separation on the Dirichlet potential By (4.14), we split P D 2 into the time-like region and the space-like region; P D 2 = C m∈Z 2sm R+ φm(η) ∗ (η) D(t −s, x′, η) ∗ (t,x′) × k∈Z j≤m+1 ψk(t) ∗ (t) φ j(y′) ∗ (y′) h(t, y′) p L p(Rn−1 x′ )dη 1/p ≤C m∈Z 2sm k>2m j≤m+1 φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) + k≤2m 2 j≤k φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ L p(R+η ) + C m∈Z 2sm k≤2m k<2 j≤2m+2 φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) × ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ L p(R+η ) ≤C m∈Z 2sm k∈Z 2 j≤min(k,2m+2) φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) × ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ L p(R+η ) + C m∈Z 2sm k∈Z 2 j≤2m+2 φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) × ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ L p(R+η ) ≡M1 + M2. (4.21) ≤C m∈Z 2sm k>2m j≤m+1 φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) k φm(η) ∗ (η) D(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ L p(R+η ) (4.21) The ﬁrst term M1 of the right hand side is the time-dominated part, letting h j ≡ φ j ∗h, we apply the Minkowski inequality and the Hausdorff–Young inequality with The ﬁrst term M1 of the right hand side is the time-dominated part, letting h j ≡ φ j ∗h, we apply the Minkowski inequality and the Hausdorff–Young inequality with T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu J. Evol. Equ. 4.3. Separation on the Dirichlet potential (4.16) to see we use the almost orthogonal estimate (4.18) between ψm and D,k,m (Lemma 6.2): For any N ∈N, (4.16) to see we use the almost orthogonal estimate (4.18) between ψm and D,k,m (Lemma 6.2): For any N ∈N, φm(η) ∗ (η) D,k, j(t, ·, η) L1(Rn−1 x′ ) ≤ ⎧ ⎪⎪⎪⎨ ⎪⎪⎪⎩ CN2k2−\| k 2 −m\| ⟨2min( k 2 ,m)η⟩N 2k ⟨2kt⟩2 , k ≥2 j, CN2k2−\| j−m\| ⟨2 jη⟩N 2k ⟨2kt⟩2 , k < 2 j for some CN > 0. Then setting 2mη = ˜η and shifting (m, k, j) →(m′, k, j) by m −k 2 = m′, the ﬁrst term of the right hand side of (4.21) can be estimated as follows: for some CN > 0. 4.3. Separation on the Dirichlet potential Exchanging the order of the summation of j and k and setting h j(t) ≡φ j ∗ (x′) h(t), it follows by changing (m, k, j) →(m, k, j) with m −j = m′ and (4.21) that M2 L1 ≤ m∈Z 2sm R+ j∈Z k<2 j R+ φm(η) ∗ (η) D,k, j(t −s, x′, η) L1 x′ × h j(s, ·) L p x′ds p dη 1/p L1t (R+) Lt (R+) ≤C m∈Z 2sm R+ j∈Z k<2 j 2k CN2−\| j−m\| ⟨2 j\|η\|⟩N × R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm R+ j∈Z k<2 j 2k2−\| j−m\| × R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds R+ 1 ⟨2 j\|η\|⟩pN dη 1/p L1t (R+) (changing the variable 2 jη = ˜η) C m∈Z 2sm R+ j∈Z k<2 j 2k CN2−\| j−m\| ⟨2 j\|η\|⟩N × R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds p dη 1/p ≤C m∈Z 2sm R+ j∈Z k<2 j 2k2−\| j−m\| × R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds R+ 1 ⟨2 j\|η\|⟩pN dη 1/p L1t (R+) (changing the variable 2 jη = ˜η) ≤C m∈Z 2sm j∈Z 2−\| j−m\|2−j p 22 j k<2 j 2k−2 j R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds × R+ 1 ⟨\|˜η\|⟩pN d ˜η 1/p L1t (R+) ≤C j∈Z m′∈Z 2 2 2 2 p 2 × k<2 j 2k−2 j R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds L1t (R+) ≤C j∈Z 2(s+2−1 p ) j k<2 j 2k−2 j R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds L1t (R+) ≤C j∈Z 2(s+2−1 p ) j k<2 j 2k−2 jh j(s, ·) L p x′ L1t (R+) ≤C j∈Z 2(s+2−1 p ) jh j(t, ·) L p x′ L1t (R+) = h L1(R+; ˙Bs+2−1/p p,1 (Rn−1 x′ )). 4.3. Separation on the Dirichlet potential Maximal L1-regularity for parabolic Page 33 of 67 30 Maximal L1-regularity for parabolic Page 33 of 67 30 Maximal L1-regularity for parabolic Page 33 of 67 30 Page 33 of 67 30 J. Evol. Equ. On the other hand for the estimate M2, we proceed a similar way to treat M1. 4.3. Separation on the Dirichlet potential Then setting 2mη = ˜η and shifting (m, k, j) →(m′, k, j) by m −k 2 = m′, the ﬁrst term of the right hand side of (4.21) can be estimated as follows: M1 L1t (R+) Lt (R+) ≤ m∈Z 2sm R+ k∈Z 2 j≤min(k,2m+2) R+ φm(η) ∗ (η) D,k, j(t −s, x′, ζ) L1 x′ × ψk ∗ (s) h j(s, x′) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm R+ k∈Z 2−\| k 2 −m\| ⟨2min( k 2 ,m)\|η\|⟩N 2k R 2k ⟨2k(t −s)⟩2 × 2 j≤k ψk ∗ (s) h j(s, ·) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm R+ k∈Z 2−\| k 2 −m\|2k R 2k ⟨2k(t −s)⟩2 × 2 j≤k ψk ∗ (s) h j(s, ·) L p x′ ds 1 ⟨\|˜η\|⟩N p 2−min( k 2 ,m)d ˜η 1/p L1t (R ≤C m∈Z 2sm k∈Z 2−\| k 2 −m\|2k2−1 p min( k 2 ,m) ≤C m∈Z 2sm k∈Z 2−\| k 2 −m\|2k2−1 p min( k 2 ,m) × R 2k ⟨2k(t −s)⟩2 2 j≤k ψk ∗ (s) h j(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k m′∈Z 2−\|m′\|+sm′+max(0,−m′ p ) 2 s 2 k × R 2k ⟨2k(t −s)⟩2 2 j≤k ψk ∗ (s) h j(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k 2 j≤k 2sjψk ∗ (s) h j(s, ·) L p x′ L1t (R+) ≤C k∈Z 2(1−1 2p )kψk ∗ (s) h(s, ·) ˙Bs p,1(Rn−1 x′ ) L1t (R+) = C h ˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1 x′ )), × R 2k ⟨2k(t −s)⟩2 2 j≤k ψk ∗ (s) h j(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k m′∈Z 2−\|m′\|+sm′+max(0,−m′ p ) 2 s 2 k × R 2k ⟨2k(t −s)⟩2 2 j≤k ψk ∗ (s) h j(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k m′∈Z 2−\|m′\|+sm′+max(0,−m′ p ) 2 s 2 k × R 2k ⟨2k(t −s)⟩2 2 j≤k ψk ∗ (s) h j(s, ·) L p x′ ds L1t (R+) ≤C k∈Z 2(1−1 2p )k 2 j≤k 2sjψk ∗ (s) h j(s, ·) L p x′ L1t (R+) ≤C k∈Z 2(1−1 2p )kψk ∗ (s) h(s, ·) ˙Bs p,1(Rn−1 x′ ) L1t (R+) = C h ˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1 x′ )), ( (4.22) where we used the fact that \|s −1/p\| < 1, i.e., −1 + 1/p < s < 1 + 1/p and s ≤0 at the second line from the bottom. 4.5. Proof of Theorem 2.2 For the proof of Theorem 2.2, it is simpler than the proof of Theorem 2.1 since it does not involve the Littlewood-Paley decomposition in (x′, η)-variable. We recall I−ℓ= [2−ℓ, 2−ℓ+1) for all ℓ∈Z. We again split the case for the time-dominated part and space-dominated part and proceed the estimate: ∞ 0 ∥ u(t)∥L p(Rn +)dt = R+ ∞ 0 Rn−1 x′ D(t −s, x′ −y′, η)h(s, y′)dsdy′ p dη 1/p L p(Rn−1 x′ ) L1t (R+) = ℓ∈Z ℓ∈I−ℓ ∞ 0 Rn−1 x′ D(t −s, x′ −y′, η)h(s, y′)dsdy′ p dη 1/p L p(Rn−1 x′ ) L1t (R+) ≤C ℓ∈Z 2−ℓD(t, x′, η) η≃2−ℓ ∗ (t,x′) h(t, x′) p L p(Rn−1 x′ ) 1/p L1t (R+), (4.24) 0 = R+ ∞ 0 Rn−1 x′ D(t −s, x′ −y′, η)h(s, y′)dsdy′ p dη 1/p L p(Rn−1 x′ ) L1t (R+) where we split the last term of (4.24) into two parts: 4.3. Separation on the Dirichlet potential (4.2 (4.23) Here, we used the fact that \|s\| < 1 for convergence of the summation on m′ In the last estimate, we exchange the order of the integration in time and the summation of T. Ogawa and S. Shimizu 30 Page 34 of 67 J. Evol. Equ. ℓ′ and use the Hausdorff–Young inequality to remove the convolution with the time potential term and then recovers the time integration outside. From (4.21), (4.22) and (4.23), the estimate (4.13) is shown. This completes the proof of Lemma 4.3. □ ℓ′ and use the Hausdorff–Young inequality to remove the convolution with the time potential term and then recovers the time integration outside. From (4.21), (4.22) and (4.23), the estimate (4.13) is shown. This completes the proof of Lemma 4.3. □ 4.5. Proof of Theorem 2.2 4.5. Proof of Theorem 2.2 where we split the last term of (4.24) into two parts: D(t, x′, η) η≃2−ℓ ∗ (t,x′) h(t, x′) L p(Rn−1 x′ ) ≤ D(t, x′, η) ∗ (t,x′) k∈Z 0≤j≤k/2 ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ + D(t, x′, η) ∗ (t,x′) j≥0 k<2 j ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ ≤ k∈Z Rn−1 R 0≤j≤k/2 D,k, j(t −r, x′ −y′) ψk ∗ (t) φ j(y′) ∗ (y′) h(r, y′) drdy′ L p x′ + j≥˜0 Rn−1 R k<2 j D,k, j(t −s, x′ −z′) φ j ∗ (x′) ψk ∗ (s) h(s, z′) dsdz′ L p x′ ≤ k∈Z R sup j∈Z D,k, j(t −r, ·) L1 x′ 0≤j≤k/2 φ j ∗ (x′) ψk ∗ (r) h(r, ·) L p x′ dr + j≥˜0 R k<2 j D,k, j(t −s, ·) L1 x′ φ j ∗ (·) h(s, ·) L p x′ ds ≡D1 ℓ+ D2 ℓ. (4.25 D(t, x′, η) η≃2−ℓ ∗ (t,x′) h(t, x′) L p(Rn−1 x′ ) ≤ D(t, x′, η) ∗ (t,x′) k∈Z 0≤j≤k/2 ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ + D(t, x′, η) ∗ (t,x′) j≥0 k<2 j ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p x′ ≡D1 ℓ+ D2 ℓ. (4.25) (4.25) Here, the notation j = 0 for the Littlewood-Paley decomposition is given in (1.4). Thus from (4.24), (4.25), we estimate the following two terms: J. Evol. Equ. Maximal L1-regularity for parabolic Page 35 of 67 30 Maximal L1-regularity for parabolic Page 35 of 67 30 Maximal L1-regularity for parabolic Page 35 of 67 30 Page 35 of 67 30 J. Evol. Equ. J. Evol. Equ. ∞ 0 ∥ u(t)∥L p(Rn)dt ≤C ℓ∈Z 2−ℓD(t, x′, η) ∗ (t,x′) h(t, x′) p L p(Rn−1) η∈I−ℓ 1/p L1t (R+) ≤C ∥{D1 ℓ}ℓ∈Z∥ℓ−1/p p L1t (R+) + C ∥{D2 ℓ}ℓ∈Z∥ℓ−1/p p L1t (R+). ∞ 0 ∥ u(t)∥L p(Rn)dt ∞ 0 ∥ u(t)∥L p(Rn)dt 0 ≤C ℓ∈Z 2−ℓD(t, x′, η) ∗ (t,x′) h(t, x′) p L p(Rn−1) η∈I−ℓ 1/p L1t (R+) ≤C ∥{D1 ℓ}ℓ∈Z∥ℓ−1/p p L1t (R+) + C ∥{D2 ℓ}ℓ∈Z∥ℓ−1/p p L1t (R+). where we split the last term of (4.24) into two parts: We now see that the following proposition is enough to ensure that Theorem 2.2 hold We now see that the following proposition is enough to ensure that Theorem 2.2 hold. Proposition 4.4. Let 1 ≤p ≤∞. Let D1 ℓand D2 ℓbe deﬁned in (4.25). Then, there exist constants C > 0 such that the following estimate holds: ∥{D1 ℓ}ℓ∈Z∥ℓ−1/p p L1t (R+) ≤C h ˙F1−1/2p 1,1 (R+;L p(Rn−1)), (4.26) ∥{D2 ℓ}ℓ∈Z∥ℓ−1/p p L1t (R+) ≤C h L1(R+;B2−1/p p,1 (Rn−1)). (4.27) (4.26) (4.27) Proof of Proposition 4.4. The estimates (4.26) and (4.27) are proved in the similar manner as (4.12) and (4.13) in the case 1 ≤p < ∞. Setting k −2ℓ= −2ℓ′ to change (ℓ, k, j) →(ℓ′, k, j), ! ℓ∈Z \|2−ℓ p )D1 ℓ\|p " 1 p L1t (R+) ≤ ℓ∈Z 2(−1 p )pℓ k∈Z R sup j∈Z χℓ,k, j(t −r, ·) L1 x′ ≤ ℓ∈Z 2(−1 p )pℓ k∈Z R sup j∈Z χℓ,k, j(t −r, ·) L1 x′ × 0≤j≤k/2 ψk ∗ (t) φ j ∗ (x′) h(r, ·) L p x′ dr p 1 p L1t (R+) ≤C ℓ∈Z k∈Z 2(2−1 p )ℓ 2k−2ℓe−2 1 2 (k−2ℓ) × R 2k ⟨2k(t −r)⟩2 ψk ∗h(r, ·) L p x′ dr p1/p L1t (R+) ≤C ℓ∈Z k∈Z 2k−ℓ p e−2 1 2 (k−2ℓ) × R 2k ⟨2k(t −r)⟩2 ψk ∗ (r) h(r, ·) L p x′ dr p1/p L1t (R+) = C ℓ′∈Z/2 k∈Z 2(1−1 2p )k−1 p ℓ′e−2−ℓ′ / × R 2k ⟨2k(t −r)⟩2 ψk ∗ (r) h(r, ·) L p x′ dr p1/p L1t (R+) = C ℓ′∈Z/2 2−1 p ℓ′e−2−ℓ′ p k∈Z 2(1−1 2p )k × R 2k ⟨2k(t −r)⟩2 ψk ∗ (r) h(r, ·) L p x′ dr p1/p L1t (R+) × R 2k ⟨2k(t −r)⟩2 ψk ∗ (r) h(r, ·) L p x′ dr p1/p L1t (R+) × R 2k ⟨2k(t −r)⟩2 ψk ∗ (r) h(r, ·) L p x′ dr p1/p L1t (R+) 30 Page 36 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 36 of 67 T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu T. where we split the last term of (4.24) into two parts: Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. ≤C k∈Z 2(1−1 2p )k R 2k ⟨2k(t −r)⟩2 ψk ∗ (r) h(r, ·) L p x′ dr L1t (R+) ≤C k∈Z 2(1−1 2p )k ψk ∗ (t) h(t, ·) L p x′ L1t (R+) ≤C k∈Z 2(1−1 2p )kψk ∗ (t) h(t, ·) L p x′ L1t (R+) ≤C h ˙F1−1/(2p) 1,1 (R+;L p x′ ). ≤C k∈Z 2(1−1 2p )k R 2k ⟨2k(t −r)⟩2 ψk ∗ (r) h(r, ·) L p x′ dr L1t (R+) ≤C k∈Z 2(1−1 2p )k ψk ∗ (t) h(t, ·) L p x′ L1t (R+) ≤C k∈Z 2(1−1 2p )kψk ∗ (t) h(t, ·) L p x′ L1t (R+) ≤C h ˙F1−1/(2p) 1,1 (R+;L p x′ ). (4.28) (4.28) ≤C h ˙F1−1/(2p) 1,1 (R+;L p x′ ). (4.28) ≤C h ˙F1−1/(2p) 1,1 (R+;L p x′ ). For the case p = ∞in (4.26), setting k −2ℓ= −2ℓ′ and changing ℓ→ℓ′, we have For the case p = ∞in (4.26), setting k −2ℓ= −2ℓ′ and changing ℓ→ℓ′, we have have sup ℓ∈Z \|D1 ℓ\| L1t (R+) ≤C sup ℓ∈Z k∈Z 22ℓ 2k−2ℓe−2 1 2 (k−2ℓ) sup ℓ∈Z \|D1 ℓ\| L1t (R+) ≤C sup ℓ∈Z k∈Z 22ℓ 2k−2ℓe−2 1 2 (k−2ℓ) × R 2k ⟨2k(t −r)⟩2 0≤j≤k/2 φ j ∗ (·) hk(r, ·) L∞ x′ dr L1t (R+) = C sup ℓ′∈Z k∈Z 2k+2ℓ′ 2−2ℓ′e−2−ℓ′ × R 2k ⟨2k(t −r)⟩2 0≤j≤k/2 φ j ∗ (·) hk(r, ·) L∞ x′ dr L1t (R+) = C sup ℓ′∈Z e−2−ℓ′ k∈Z 2k × R 2k ⟨2k(t −r)⟩2 0≤j≤k/2 φ j ∗ (·) hk(r, ·) L∞ x′ dr L1t (R+) = C k∈Z 2k R 2k ⟨2k(t −r)⟩2 hk(r, ·) L∞ x′ dr L1t (R+) ≤C h ˙F1 1,1(R+;L∞ x′ ). Page 37 of 67 30 × j≥0 2(2−1 p ) j k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩ h j(r, ·) L p x′ dr p1/p L1t (R+) ≤C j≥0 2(2−1 p ) j k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩ h j(r, ·) L p x′ dr L1t (R+) ≤C j≥0 2(2−1 p ) j k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩ h j(r, ·) L p x′ dr L1t (R+) ≤C j≥0 2(2−1 p ) jh j(t, ·) L p x′ L1t (R+) = C h L1(R+;B2−1/p p,1 (Rn−1 x′ )). = C h L1(R+;B2−1/p p,1 (Rn−1 x′ )). For the case p = ∞in (4.27), by setting j −ℓ= −ℓ′ to change (ℓ, j) →(ℓ′, j), we see that For the case p = ∞in (4.27), by setting j −ℓ= −ℓ′ to change (ℓ, j) →(ℓ′, j), we see that see that see that sup ℓ∈Z \|D2 ℓ\| L1t (R+) ≤C sup ℓ∈Z 22ℓ R j≥0 k<2 j 2k−2ℓe−2 j−ℓ 2k ⟨2k(t −r)⟩2 h j(r, ·) L∞ x′ dr L1t (R+) ≤C sup ℓ∈Z j≥0 22ℓ 22( j−ℓ)e−2( j−ℓ) × k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩2 h j(r, ·) L∞ x′ dr L1t (R+) = C sup ℓ′∈Z j≥0 22ℓ′+2 j2−2ℓ′e−2−ℓ′ × k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩2 h j(r, ·) L∞ x′ dr L1t (R+) = C sup ℓ′∈Z e−2−ℓ′ j≥0 22 j k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩2 h j(r, ·) L∞ x′ dr L1t (R+) ≤C j≥˜0 22 j k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩2 h j(r, ·) L∞ x′ dr L1t (R+) ≤C j≥0 22 j h j(t, ·) L∞ x′ L1t (R+) = C h L1(R+;B2 ∞,1(Rn−1 x′ )). This concludes the proof of Proposition 4.4. □ □ This concludes the proof of Proposition 4.4. □ This concludes the proof of Proposition 4.4. This concludes the proof of Proposition 4.4. where we split the last term of (4.24) into two parts: ≤C sup ℓ∈Z k∈Z 22ℓ 2k−2ℓe−2 1 2 (k−2ℓ) × R 2k ⟨2k(t −r)⟩2 0≤j≤k/2 φ j ∗ (·) hk(r, ·) L∞ x′ dr L1t (R+) = C sup ℓ′∈Z k∈Z 2k+2ℓ′ 2−2ℓ′e−2−ℓ′ × R 2k ⟨2k(t −r)⟩2 0≤j≤k/2 φ j ∗ (·) hk(r, ·) L∞ x′ dr L1t (R+) = C sup ℓ′∈Z e−2−ℓ′ k∈Z 2k × R 2k ⟨2k(t −r)⟩2 0≤j≤k/2 φ j ∗ (·) hk(r, ·) L∞ x′ dr L1t (R+) × R 2k ⟨2k(t −r)⟩2 0≤j≤k/2 φ j ∗ (·) hk(r, ·) L∞ x′ dr L1t (R+) = C k∈Z 2k R 2k ⟨2k(t −r)⟩2 hk(r, ·) L∞ x′ dr L1t (R+) ≤C h ˙F1 1,1(R+;L∞ x′ ). = C k∈Z 2k R 2k ⟨2k(t −r)⟩2 hk(r, ·) L∞ x′ dr L1t (R+) ≤C h ˙F1 1,1(R+;L∞ x′ ). To show the estimate (4.27) for 1 ≤p < ∞, we derive it by setting g j(s, y′) ≡ φ j ∗g(s, y′) and then j −ℓ= −ℓ′ and (ℓ, k, j) →(ℓ′, k, j) that To show the estimate (4.27) for 1 ≤p < ∞, we derive it by setting g j(s, y′) ≡ φ j ∗g(s, y′) and then j −ℓ= −ℓ′ and (ℓ, k, j) →(ℓ′, k, j) that ! ℓ∈Z 2(−1 p )pℓ\|D2 ℓ\|p " 1 p L1t (R+) ≤ ⎧ ⎨ ⎩ ℓ∈Z 2(−1 p )pℓ R j≥0 k<2 j ψD,k, j(t −r, ·) L1 x′ h j(r, ·) L p x′ dr p ⎫ ⎬ ⎭ 1/p L1t (R+) ≤C ℓ∈Z 2(2−1 p )pℓ j≥0 22( j−ℓ)e−2( j−ℓ) × k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩ h j(r, ·) L p x′ dr p1/p L1t (R+) = C ℓ′∈Z 2−ℓ′ p e−2−ℓ′ p × k<2 j 2k−2 j R 2k ⟨2k(t −r)⟩ h j(r, ·) L p x′ dr p1/p L1t (R+) J. Evol. Equ. Maximal L1-regularity for parabolic Page 37 of 67 30 Maximal L1-regularity for parabolic Page 37 of 67 30 Maximal L1-regularity for parabolic Page 37 of 67 30 J. Evol. Equ. Page 37 of 67 30 5. The Neumann boundary condition 5.1. The Neumann boundary condition and the boundary trace For the Neumann boundary condition, the boundary potential associated to the Green’s function is slightly different from the one for the Dirichlet case. However as T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. we see the explicit form of the potential function, it simply adjusts the order of the derivative and the estimate is very similar to the Dirichlet case. From (3.21), (3.20) and (4.9), we see the explicit form of the potential function, it simply adjusts the order of the derivative and the estimate is very similar to the Dirichlet case. From (3.21), (3.20) and (4.9), m ∗ (x′,η) η−1N(t, x′, η) = −c2 nζm−1(η) ∗ (η) Rn−1 R eitτ+ix′·ξ′ φm(\|ξ′\|) τ iτ + \|ξ′\|2 e−√ iτ+\|ξ′\|2ηdξ′dτ −c2 nφm(η) ∗ (η) R Rn−1 R eitτ+ix′·ξ′ ζm(\|ξ′\|) τ iτ + \|ξ′\|2 e−√ iτ+\|ξ′\|2ηdξ′dτ. (5.1) (5.1) To estimate the Besov-norm of the solution, we use the Littlewood-Paley decompo- sition for direct sum type (4.7). The estimates for the right hand side of (5.1) follow very similar manner to the case of the Dirichlet boundary condition. To estimate the Besov-norm of the solution, we use the Littlewood-Paley decompo- sition for direct sum type (4.7). The estimates for the right hand side of (5.1) follow very similar manner to the case of the Dirichlet boundary condition. ∞ 0 ∥ u(t)∥˙Bs p,1(Rn)dt ≤C m∈Z 2sm R+ N(t, x′, η) ∗ (t,x′) φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) + C m∈Z 2sm R+ φm(η) ∗ (η) N(t, x′, η) ∗ (t,x′) × k∈Z j≤m ψk(t) ∗ (t) φ j(x′) ∗ (y′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) ≡∥P N 1 ∥L1t (R+) + ∥P N 2 ∥L1t (R+). 5. The Neumann boundary condition (5.2) 5 2 Proof of Theorem 2 3 ∞ 0 ∥ u(t)∥˙Bs p,1(Rn)dt ≤C m∈Z 2sm R+ N(t, x′, η) ∗ (t,x′) φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) + C m∈Z 2sm R+ φm(η) ∗ (η) N(t, x′, η) ∗ (t,x′) ∞ 0 ∥ u(t)∥˙Bs p,1(Rn)dt ≤C m∈Z 2sm R+ N(t, x′, η) ∗ (t,x′) φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) 0 p, ≤C m∈Z 2sm R+ N(t, x′, η) ∗ (t,x′) φm(x′) ∗ (x′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) + C m∈Z 2sm R+ φm(η) ∗ (η) N(t, x′, η) ∗ (t,x′) m∈ × k∈Z j≤m ψk(t) ∗ (t) φ j(x′) ∗ (y′) h(t, x′) p L p(Rn−1 x′ )dη 1/p L1t (R+) ≡∥P N 1 ∥L1t (R+) + ∥P N 2 ∥L1t (R+). (5.2) 5.2. Proof of Theorem 2.3 ≡∥P N 1 ∥L1t (R+) + ∥P N 2 ∥L1t (R+). (5.2) 5.2. Proof of Theorem 2.3 5.2. Proof of Theorem 2.3 5.2. Proof of Theorem 2.3 We prove the sufﬁciency part of Theorem 2.3, which is reduced to prove Lemma 5.1. We assume that h satisﬁes (2.9). For the Neumann case, we use the time-space splitting argument same as the Dirichlet case. We split each of two terms in the right hand side of (5.2) into time-dominated area and space-dominated area. Lemma 5.1. Let 1 < p < ∞and −1 + 1 p < s ≤0. There exists a constant C > 0 such that P N 1 deﬁned in (5.2) satisﬁes Lemma 5.1. Let 1 < p < ∞and −1 + 1 p < s ≤0. There exists a constant C > 0 such that P N 1 deﬁned in (5.2) satisﬁes ∥P N 1 ∥L1t (R+) ≤C h ˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + h L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) . (5.3) Similarly P N 2 deﬁned in (5.2) satisﬁes the following estimate: Similarly P N 2 deﬁned in (5.2) satisﬁes the following estimate: ∥P N 2 ∥L1t (R+) ≤C h ˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + h L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) . (5.4) ∥P N 2 ∥L1t (R+) ≤C h ˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + h L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) . 5. The Neumann boundary condition (5.4) Maximal L1-regularity for parabolic Page 39 of 67 30 Maximal L1-regularity for parabolic Maximal L1-regularity for parabolic Page 39 of 67 30 Page 39 of 67 30 J. Evol. Equ. J. Evol. Equ. Proof of Lemma 5.1. Introducing Proof of Lemma 5.1. Introducing Proof of Lemma 5.1. Introducing N,k, j(t, x′, η) ≡ R Rn−1 N(t −s, x′ −y′, η)ψk(s)φ j(y′)dy′ds, we apply (4.14) to P N 1 and set hm ≡ φm ∗ (x′) h. Similar to the estimates (4.15) and (4.17) in the Dirichlet boundary case, we obtain that we apply (4.14) to P N 1 and set hm ≡ φm ∗ (x′) h. Similar to the estimates (4.15) and (4.17) in the Dirichlet boundary case, we obtain that ( ) (4.17) in the Dirichlet boundary case, we obtain that P N 1 P N 1 ≤C m∈Z 2sm R+ k∈Z 2 j≤k N(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) h j(t, x′) p L p(Rn−1 x′ )dη 1/p + C m∈Z 2sm R+ k∈Z 2 j>k N(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) h j(t, x′) p L p(Rn−1 x′ )dη 1/p ≤C m∈Z 2sm R+ k∈Z N(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) hm(t, x′) p L p x′ dη 1/p + C m∈Z 2sm R+ k∈Z N(t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φm(x′) ∗ (x′) hm(t, x′) p L p x′ dη 1/p ≤C m∈Z 2sm R+ k≥2m R+ N,k,m(t −s, x′, η) L1 x′ ψk ∗ (s) hm(s, x′) L p x′ ds p dη 1/p + C m∈Z 2sm R+ k≤2m R+ N,k,m(t −s, x′, η) L1 x′ hm(s, x′) L p x′ ds p dη 1/p ≡L + L (5 5) m∈Z k≤2m ≡L1 + L2. (5.5) ≡L1 + L2. (5.5) (5.5) ≡L1 + L2. Then similar to the Dirichlet boundary condition, we invoke Lemma 6.5 in Sect. 6 and it implies Then similar to the Dirichlet boundary condition, we invoke Lemma 6.5 in Sect. 6 and it implies N,k,m(t, ·) L1 x′ ≤ ⎧ ⎪⎪⎪⎨ ⎪⎪⎪⎩ Cn2 k 2 e−2 k 2 η 2k ⟨2kt⟩2 , k ≥2m, Cn2 k 2 e−2mη 2k ⟨2kt⟩2 , k < 2m. 5. The Neumann boundary condition (5.6) (5.6) Applying (5.6) to (5.5), by the similar manner as the proof of (4.19) in Lemma 4.3, we have for the ﬁrst term of the right hand side that Applying (5.6) to (5.5), by the similar manner as the proof of (4.19) in Lemma 4.3, we have for the ﬁrst term of the right hand side that L1∥L1t (R+) ≤C m∈Z 2sm ⎛ ⎝ R+ k≥2m 2 k 2 e−2 k 2 η R 2k ⟨2k(t −s)⟩2 ψk ∗ (t) hm(s, ·) L p x′ ds p dη ⎞ ⎠ 1/p L1t (R+) = C m∈Z 2sm k≥2m 2 k 2 R 2k ⟨2k(t −s)⟩2 ψk ∗ (t) hm(s, ·) L p x′ ds R+ e−p2 k 2 ηdη 1/p L1t (R+) = C m∈Z 2sm k∈Z 2 k 2 −k 2p R 2k ⟨2k(t −s)⟩2 ψk ∗ (t) hm(s, ·) L p x′ ds L1t (R+) T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. ≤C m∈Z 2sm k∈Z 2 k 2 −k 2p ψk ∗ (t) hm(s, ·) L p x′ L1t (R+) ≤C h ˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1 x′ )). (5.7) ≤C m∈Z 2sm k∈Z 2 k 2 −k 2p ψk ∗ (t) hm(s, ·) L p x′ L1t (R+) ≤C h (5 7) (5.7) The term L2 stands for the space-dominated part and we set hm ≡ φm ∗h. 5. The Neumann boundary condition Observing φ j ∗φm ≃φm (\| j −m\| ≤1) we apply (5.6) to (5.5) to see that L2∥L1t (R+) t ( +) ≤C m∈Z 2sm R+ k<2m 2 k 2 e−2mη R 2k ⟨2k(t −s)⟩2 hm(s, x′) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm k<2m 2 k 2 R 2k ⟨2k(t −s)⟩2 hm(s, x′) L p x′ ds R+ e−p2mηdη 1/p L1t (R+) ≤C m∈Z 2sm k<2m 2 k 2 2−m p R 2k ⟨2k(t −s)⟩2 hm(s, x′) L p x′ ds L1t (R+) ≤C m∈Z 2(s+1−1 p )m k<2m 2 k 2 −m hm(s, x′) L p x′ L1t (R+) ≤C m∈Z 2(s+1−1 p )mhm(s, x′) L p x′ L1t (R+) = C h L1(R+; ˙Bs+1−1/p p,1 (Rn−1 x′ )). (5.8) (5.8) To see (5.4), we split in a similar way to (4.21) again and obtain that P N 2 ≤C m∈Z 2sm k∈Z 2 j≤k φm(η) ∗ (η) N (t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p(Rn−1 x′ ) L p(R+η ) + C m∈Z 2sm k∈Z 2 j>k φm(η) ∗ (η) N (t, x′, η) ∗ (t,x′) ψk(t) ∗ (t) φ j(x′) ∗ (x′) h(t, x′) L p(Rn−1 x′ ) L p(R+η ) ≤C m∈Z 2sm ⎛ ⎝ R+ ⎧ ⎨ ⎩ k∈Z 2 j≤k R+ φm(η) ∗ (η) N,k, j(t −s, x′, η) L1 x′ ψk ∗ (s) h j(s, x′) L p x′ ds ⎫ ⎬ ⎭ p dη ⎞ ⎠ 1/p + C m∈Z 2sm ⎛ ⎝ R+ j∈Z k<2 j R+ φm(η) ∗ (η) N,k, j(t−s, x′, η) L1 x′ h j(s) L p x′ ds p dη ⎞ ⎠ 1/p ≡M1 + M2. (5.9) (5.9) (5.9) ≡M1 + M2. We use the almost orthogonal estimate (5.6) between ψm and N,k, j. From (5.9), We use the almost orthogonal estimate (5.6) between ψm and N,k, j. 5. The Neumann boundary condition From (5.10) and (5.11), we conclude the proof of (5.4). □ 5. The Neumann boundary condition Setting h j(t) ≡φ j ∗ (x′) h(t) in (4.21) and changing the order of summation between j and k, we have similar to the estimate for M1 by the Minkowski and the Hausdorff–Young inequalities with (4.16) that M2 L1t (R+) M2 L1t (R+) ≤C m∈Z 2sm R+ j∈Z k≤2 j CN 2−\| j−m\| ⟨2 j\|η\|⟩N 2 k 2 Lt (R+) ≤C m∈Z 2sm R+ j∈Z k≤2 j CN 2−\| j−m\| ⟨2 j\|η\|⟩N 2 k 2 × R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm j∈Z k≤2 j 2−\| j−m\|2 k 2 R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds × 2−j p R+ 1 ⟨\|˜η\|⟩pN d ˜η 1/p L1t (R+) ≤C j∈Z 2sj+ j−j p k≤2 j 2 k 2 −j m′∈Z 2−\|m′\|2sm′ h j(s, ·) L p x′ L1t (R+) = C∥h∥L1(R+; ˙Bs+1−1/p p,1 (Rn−1 x′ )) m∈Z j∈Z k≤2 j × R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm j∈Z k≤2 j 2−\| j−m\|2 k 2 R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds × 2−j p R+ 1 ⟨\|˜η\|⟩pN d ˜η 1/p L1t (R+) ≤C j∈Z 2sj+ j−j p k≤2 j 2 k 2 −j m′∈Z 2−\|m′\|2sm′ h j(s, ·) L p x′ L1t (R+) × R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm j∈Z k≤2 j 2−\| j−m\|2 k 2 R 2k ⟨2k(t −s)⟩2 h j(s, ·) L p x′ ds × 2−j p R+ 1 ⟨\|˜η\|⟩pN d ˜η 1/p L1t (R+) ≤C j∈Z 2sj+ j−j p k≤2 j 2 k 2 −j m′∈Z 2−\|m′\|2sm′ h j(s, ·) L p x′ L1t (R+) ≤C j∈Z 2sj+ j−j p k≤2 j 2 k 2 −j m′∈Z 2−\|m′\|2sm′ h j(s, ·) L p x′ L1t (R+) = C∥h∥L1(R+; ˙Bs+1−1/p p,1 (Rn−1 x′ )) (5.11) (5.11) = C∥h∥L1(R+; ˙Bs+1−1/p p,1 (Rn−1 x′ )) (5.11) under the condition \|s\| < 1. From (5.10) and (5.11), we conclude the proof of (5.4). □ under the condition \|s\| < 1. 5. The Neumann boundary condition From (5.9) M1 L1t (R+) M1 L1t (R+) k ≤C m∈Z 2sm R+ k∈Z 2 j≤k CN 2−\| k 2 −m\| ⟨2min( k 2 ,m)η\|⟩N × 2 k 2 R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) h j(s, ·) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm k∈Z 2 j≤k 2−\| k 2 −m\|2 k 2 R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) h j(s, ·) L p x′ ds ≤C m∈Z 2sm R+ k∈Z 2 j≤k CN 2−\| k 2 −m\| ⟨2min( k 2 ,m)η\|⟩N j × 2 k 2 R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) h j(s, ·) L p x′ ds p dη 1/p L1t (R+) ≤C m∈Z 2sm k∈Z 2 j≤k 2−\| k 2 −m\|2 k 2 R 2k ⟨2k(t −s)⟩2 ψk ∗ (s) h j(s, ·) L p x′ ds J. Evol. Equ. Maximal L1-regularity for parabolic Page 41 of 67 30 J. Evol. Equ. Maximal L1-regularity for parabolic Page 41 of 67 30 Maximal L1-regularity for parabolic Page 41 of 67 30 Page 41 of 67 30 Page 41 of 67 30 J. Evol. Equ. × 2−1 p min( k 2 ,m) R+ 1 ⟨\|˜η\|⟩pN d ˜η 1/p L1t (R+) ≤C k∈Z 2 k 2 −k 2p 2 j≤k 2 s 2 k m′∈Z 2−\|m′\|2max(0,−m′ p )2sm′ × ψk ∗ (s) h j(s, ·) L p x′ L1t (R+) (5.10) = C∥h∥˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1 x′ )) = C∥h∥˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1 x′ )) (5.10) under the condition −1 + 1/p < s ≤0. 5.3. Proof of Theorem 2.4 (5.13 ≡N 1 ℓ+ N 2 ℓ. (5.13) From (5.12) and (5.13), we reduce the estimate for the following: ∞ 0 ∥ u(t)∥L p(Rn +)dt = ℓ∈Z 2−ℓN(t, x′, η) ∗ (t,x′) h(t, x′) p L p(Rn−1) η∈I−ℓ 1/p L1t (R+) ≤ ∥{N 1 ℓ}ℓ∥ℓ−1/p p L1t (R+) + ∥{N 2 ℓ}ℓ∥ℓ−1/p p L1t (R+). = ℓ∈Z 2−ℓN(t, x′, η) ∗ (t,x′) h(t, x′) p L p(Rn−1) η∈I−ℓ 1/p L1t (R+) ≤ ∥{N 1 ℓ}ℓ∥ℓ−1/p p L1t (R+) + ∥{N 2 ℓ}ℓ∥ℓ−1/p p L1t (R+). The proof of Theorem 2.4 is now reduced to show the following proposition. The proof of Theorem 2.4 is now reduced to show the following proposition. Proposition 5.2. Let 1 ≤p ≤∞. For N 1 ℓand N 2 ℓgiven by (5.13), the following estimates hold. ∥{N 1 ℓ}ℓ∥ℓ−1/p p L1t (R+) ≤C h ˙F1/2−1/2p 1,1 (R+;L p(Rn−1)), ∥{N 2 ℓ}ℓ∥ℓ−1/p p L1t (R+) ≤C h L1(R+;B1−1/p p,1 (Rn−1)). The proof of Proposition 5.2 can be shown in a similar way to Proposition 4.4. 5.3. Proof of Theorem 2.4 5.3. Proof of Theorem 2.4 The proof of Theorem 2.4 can be done by a similar way to the case of the Dirichlet boundary condition. Indeed, we obtain a similar estimate to (5.3): ∞ 0 ∥ u(t)∥L p(Rn +)dt = R+ ∞ 0 Rn−1 x′ N(t −s, x′ −y′, η)h(s, y′)dsdy′ p dη 1/p L p(Rn−1 x′ ) L1t (R+) = ℓ∈Z 2−ℓN(t, x′, η) η≃2−ℓ ∗ (t,x′) h(t, x′) p L p(Rn−1 x′ ) 1/p L p(Rn−1 x′ ) L1t (R+), (5.12) 0 = R+ ∞ 0 Rn−1 x′ N(t −s, x′ −y′, η)h(s, y′)dsdy′ p dη 1/p L p(Rn−1 x′ ) L1t (R+) = ℓ∈Z 2−ℓN(t, x′, η) η≃2−ℓ ∗ (t,x′) h(t, x′) p L p(Rn−1 x′ ) 1/p L p(Rn−1 x′ ) L1t (R+), (5.12) (5.12) then we split the last term into two terms: then we split the last term into two terms: 30 Page 42 of 67 30 Page 42 of 67 T. Ogawa and S. Shimizu 30 Page 42 of 67 J. Evol. Equ. N(t, x′, η) ∗ (t,x′) h(t, x′) L p(Rn−1 x′ ) ≤ N(t, x′, η) ∗ (t,x′) k∈Z 0≤j≤k/2 ψk(t) ∗φ j(x′) ∗h(t, x′) L p(Rn−1) + N(t, x′, η) ∗ (t,x′) j≥0 k<2 j ψk(t) ∗φ j(x′) ∗h(t, x′) L p(Rn−1) ≤ k∈Z R sup j∈Z N,k, j(t −r, ·, η) L1 x′ ψk ∗ (t) 0≤j≤k/2 φ j ∗h(r, ·) L p x′ + j≥0 R sup k∈Z N,k, j(t −r, ·, η) L1 x′ k<2 j ψk ∗h(s, ·) L p x′ ds ( , ) ≤ N(t, x′, η) ∗ (t,x′) k∈Z 0≤j≤k/2 ψk(t) ∗φ j(x′) ∗h(t, x′) L p(Rn−1) + N(t, x′, η) ∗ (t,x′) j≥0 k<2 j ψk(t) ∗φ j(x′) ∗h(t, x′) L p(Rn−1) ≤ k∈Z R sup j∈Z N,k, j(t −r, ·, η) L1 x′ ψk ∗ (t) 0≤j≤k/2 φ j ∗h(r, ·) L p x′ + j≥0 R sup k∈Z N,k, j(t −r, ·, η) L1 x′ k<2 j ψk ∗h(s, ·) L p x′ ds (5.13) ≡N 1 ℓ+ N 2 ℓ. ≡N 1 ℓ+ N 2 ℓ. ≡N 1 ℓ+ N 2 ℓ. 6. Almost orthogonality In this section, we prove the almost orthogonality (2.12) (or (4.18) in Sect. 4) between the boundary potential D for the Dirichlet boundary case and N for the Neumann boundary case with the time and space Littlewood-Paley decomposition {ψk}k∈Z and {φ j} j∈Z. The difﬁculty is that the fundamental solution D(t, x′, η) and N(t, x′, η) made by heat kernel is time and space convolution. Maximal L1-regularity for parabolic Page 43 of 67 30 J. Evol. Equ. Maximal L1-regularity for parabolic Page 43 of 67 30 J. Evol. Equ. Page 43 of 67 30 Maximal L1-regularity for parabolic Page 43 of 67 30 Page 43 of 67 30 6.1. The Dirichlet potential case 6.1. The Dirichlet potential case Lemma 6.1 (Crucial potential orthogonality). For k, j, ℓ∈Z let {ψk(t)}k∈Z and {φ j(x)} j∈Z be the time and the space Littlewood-Paley dyadic decomposition and let D(t, x′, η) be the boundary potential deﬁned in (3.11). Set D,k, j(t, x′, η) ≡ R Rn−1 D(t −s, x′ −y′, η)ψk(s)φ j(y′)dy′ds, (6.1) (6.1) for η > 0. Then, there exists a constant Cn > 0 depending only on the dimension n satisfying ∥D,k, j(t, ·, η)∥L1 x′ ≤ ⎧ ⎪⎪⎪⎨ ⎪⎪⎪⎩ Cn2k 1 + (2 k 2 η)n+2 e−2 k 2 −1η 2k ⟨2kt⟩2 , k ≥2 j, Cn2k 1 + (2 jη)n+2 e−2 j−1η 2k ⟨2kt⟩2 , k < 2 j. (6.2) (6.2) Proof of Lemma 6.1. We consider a time-like estimate k > 2 j. Taking ζ ′-space cutoff, and using the change of variables τ = 2kσ, ξ′ = 2 jζ ′ we have Proof of Lemma 6.1. We consider a time-like estimate k > 2 j. 6. Almost orthogonality Taking ζ ′-space cutoff, and using the change of variables τ = 2kσ, ξ′ = 2 jζ ′ we have D,k, j(t, ·, η) L1 x′ x = cn+1 R Rn−1 eitτ+ix′·ξ′τ × exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ L1 x′ = cn+1 R Rn−1 ei2ktσ+i2 j x′·ζ ′σ2k × exp − * 2kiσ + 22 j\|ζ ′\|2η ψ(σ)φ(ζ ′)2(n−1) jdζ ′ · 2kdσ = cn+1 R Rn−1 eitτ+ix′·ξ′τ × exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ L1 x′ = cn+1 R Rn−1 ei2ktσ+i2 j x′·ζ ′σ2k × exp − * 2kiσ + 22 j\|ζ ′\|2η ψ(σ)φ(ζ ′)2(n−1) jdζ ′ · 2kdσ L1 x′ = cn+12k R Rn−1 ei2ktσ+i2 j x′·ζ ′σ × exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ψ(σ)φ(ζ ′)2(n−1) jdζ ′ · 2kdσ L1 x′ = cn+12k R Rn−1 ei2ktσ+iy′·ζ ′ × exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 σ ψ(σ)φ(ζ ′)dζ ′ · 2kdσ L1 y′ (6. = cn+12k R Rn−1 ei2ktσ+i2 j x′·ζ ′σ × exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ψ(σ)φ(ζ ′)2(n−1) jdζ ′ · 2kdσ L1 x′ (6.3) by setting x′ = 2−j y′ in the last equality. Using the identity ei(2ktσ+y′·ζ ′) = −1 i2kt 1 \|y′\|2 ζ ′∂σei(2ktσ+y′·ζ ′), (6.4) (6.4) and integrating by parts with respect to σ and ζ ′, setting p(σ, ζ ′; 22 j−k) ≡ iσ + 22 j−k\|ζ ′\|2, we proceed and integrating by parts with respect to σ and ζ ′, setting p(σ, ζ ′; 22 j−k) ≡ iσ + 22 j−k\|ζ ′\|2, we proceed T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 44 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. ∥D,k, j(t, ·)∥L1 x′(Bc 2 j ) = cn+12k R Rn−1 1 i2kt −1 \|y′\|2 ei(2ktσ+y′·ζ ′) × ζ ′ ∂ ∂σ + exp −2 k 2 ηp(σ, ζ ′; 22 j−k) ψ(σ)φ(ζ ′)2kσ , dζ ′dσ L1 y′(Bc 1). 6. Almost orthogonality y Here, Here, Here, ∂ ∂σ + exp −2 k 2 η p(σ, ζ ′; 22 j−k) ψ(σ)2kσφ(ζ ′) , = exp −2 k 2 η p(σ, ζ ′; 22 j−k) −2 k 2 ηi 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′). 6. Almost orthogonality Evol. Equ. + 2 k 2 η−3(22 j−k)2i\|ζ ′\|2 + 22 j−kip(σ, ζ ′; 22 j−k)2 2p(σ, ζ ′; 22 j−k)5 ψ(σ)2kσφ(ζ ′) + 2 k 2 η 22 j−ki\|ζ ′\| 2p(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ′(ζ ′) + (n −2) exp −2 k 2 ηp(σ, ζ ′; 22 j−k) × −2 k 2 η 22 j−k p(σ, ζ ′; 22 j−k) φ(ζ ′) + φ′(ζ ′) \|ζ ′\| −2 k 2 ηi 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 k 2 η 22 j−ki 2p(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ(ζ ′) . Summarizing the results, all terms have Summarizing the results, all terms have exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 = exp −2 k 2 η p(σ, ζ ′; 22 j−k) namely ψ(σ), φ(ζ ′) or their derivatives and the order of derivatives are the order of partial derivatives with respect to σ and ζ ′. And 2 k 2 η arises the same order of partial derivatives at most and arises one time at least. The function in denominator p(σ, ζ ′; 22 j−k) = iσ + 22 j−k\|ζ ′\|2 is estimated as 2−1/2 ≤ p(σ, ζ ′, 22 j−k) = (σ 2 + 22(2 j−k)\|ζ ′\|4)1/4 ≤201/4, (6.5) (6.5) thanks to the cutoff functions ψ(σ) and φ(ζ ′) or its derivative. Therefore, if we dif- ferentiate it n + 2 times, then the terms involving p(σ, ζ ′; 22 j−k) are estimated from below by 2n+2. Moreover p(σ, ζ ′; 22 j−k) which arises by each derivative contains the surplus scale parameter 22 j−k and it is estimated from above by 1 because of the restriction k ≥2 j. 6. Almost orthogonality In order to take the second derivative for space, setting r = \|ζ ′\| and using the relation ζ ′ = ∂2 r + n−2 r ∂r, we have by φ(ζ ′) = φ(\|ζ ′\|) that In order to take the second derivative for space, setting r = \|ζ ′\| and using the relation ζ ′ = ∂2 r + n−2 r ∂r, we have by φ(ζ ′) = φ(\|ζ ′\|) that ζ r r ζ ′ ∂ ∂σ exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ψ(σ)2kσφ(ζ ′) = ζ ′ exp −2 k 2 η p(σ, ζ ′; 22 j−k) × −2 k 2 ηi 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′) = ∂r + n −2 r exp −2 k 2 η p(σ, ζ ′; 22 j−k) × −2 k 2 η 22 j−k\|ζ ′\| p(σ, ζ ′; 22 j−k) φ(ζ ′) + φ′(ζ ′) −2 k 2 ηi 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 k 2 η 22 j−ki\|ζ ′\| 2p(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ(ζ ′) = exp −2 k 2 ηp(σ, ζ ′; 2k−2 j) × −2 k 2 η 22 j−k\|ζ ′\| p(σ, ζ ′; 22 j−k) −2 k 2 η 22 j−k\|ζ ′\| p(σ, ζ ′; 22 j−k) φ(ζ ′) + φ′(ζ ′) × −2 k 2 ηi 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 k 2 η 22 j−ki\|ζ ′\| 2p(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ(ζ ′) + −2 k 2 η 22 j−kiσ p(σ, ζ ′; 22 j−k)3 φ(ζ ′) −2 k 2 η 22 j−k\|ζ ′\| p(σ, ζ ′; 22 j−k) φ′(ζ ′) + φ′′(ζ ′) × − −2 k 2 ηi 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + −2 k 2 η 22 j−k\|ζ ′\| p(σ, ζ ′; 22 j−k) φ(ζ ′) + φ′(ζ ′) 2 k 2 η 22 j−ki\|ζ ′\| 2p(σ, ζ ′; 22 j−k)3 ψ(σ)2kσ ζ ′ ∂ ∂σ exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ψ(σ)2kσφ(ζ ′) = ζ ′ exp −2 k 2 η p(σ, ζ ′; 22 j−k) × −2 k 2 η 22 j−k\|ζ ′\| p(σ, ζ ′; 22 j−k) φ(ζ ′) + φ′(ζ ′) −2 k 2 ηi 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 k 2 η 22 j−ki\|ζ ′\| 2p(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ(ζ ′) Maximal L1-regularity for parabolic Page 45 of 67 30 Maximal L1-regularity for parabolic Page 45 of 67 30 Maximal L1-regularity for parabolic Maximal L1-regularity for parabolic Page 45 of 67 30 Page 45 of 67 30 Page 45 of 67 30 J. 6. Almost orthogonality × K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ = cn+12k exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × 1 \| ′\|2 n 2 ei(2ktσ+y′·ζ ′)e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k j (σ, ζ ′, η, 22 j−k)dζ ′dσ × K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ = cn+12k exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ = cn+12k exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ ≤C2k exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × R Rn−1 e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k, j (σ, ζ ′, η, 22 j−k) dζ ′dσ ≤C2k exp(−2 k 2 −1η) 1 + C(2 k 2 η)n+2 2k ⟨2kt⟩2 , × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ k k 1 2k × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ ≤C2k exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × R Rn−1 e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k, j (σ, ζ ′, η, 22 j−k) dζ ′dσ ≤C2k exp(−2 k 2 −1η) 1 + C(2 k 2 η)n+2 2k ⟨2kt⟩2 , ≤C2k exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × R Rn−1 e −2 k 2 η p(σ,ζ ′;22 j−k)−1/2 K n+2 k, j (σ, ζ ′, η, 22 j−k) dζ ′dσ ≤C2k exp(−2 k 2 −1η) 1 + C(2 k 2 η)n+2 2k ⟨2kt⟩2 , arranging the sufﬁxes we obtain (6.2). Next we consider the case k ≤2 j. 6. Almost orthogonality Thus for the functions kk, j(σ, ζ ′, η) ≡exp −2 k 2 η p(σ, ζ ′; 22 j−k) ψ(σ)2kσφ(ζ ′), K n+2 k, j (σ, ζ ′, η, 22 j−k) ≡exp 2 k 2 η p(σ, ζ ′; 22 j−k) Dn+2 σ,ζ ′ kk, j(σ, ζ ′, η), where we set where we set Dn+2 σ,ζ ′ ≡(1 − ζ ′) n 2 1 −∂2 ∂σ 2 , under the condition k ≥2 j namely 22 j−k ≤1, we obtain the following estimate: For η > 0 ∥D,k, j(t, ·, η)∥L1 x′ = cn+12k 1 ⟨y′⟩2 n 2 2k ⟨2kt⟩2 R Rn−1 ei(2ktσ+y′·ζ ′) exp −2 k 2 ηp(σ, ζ ′; 22 j−k) 30 Page 46 of 67 T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 46 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu J. Evol. Equ. 6. Almost orthogonality Setting η = 2−ℓand using the change of variables τ = 2kσ, ξ′ = 2 jζ ′ τ = 2kσ, ξ′ = 2 jζ ′ ∥D,k, j(t, ·)∥L1 x′ ∥D,k, j(t, ·)∥L1 x′ = cn+1 R Rn−1 eitτ+ix′·ξ′τ exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ L1 x′ = cn+1 R Rn−1 ei2ktσ+i2 j x′·ζ ′σ2k × exp − * 2kiσ + 22 j\|ζ ′\|2η ψ(σ)φ(ζ ′)2(n−1) jdζ ′ · 2kdσ L1 x′ = 2kcn+1 R Rn−1 ei2ktσ+iy′·ζ ′ exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 σ ψ(σ)φ(ζ ′)dζ ′ · 2kdσ L1 y′ , where we set setting x′ = 2−j y′ in the last equality. Noting the identity (6.4) and integrating by parts with respect to σ and ζ ′, where we set setting x′ = 2−j y′ in the last equality. Noting the identity (6.4) and integrating by parts with respect to σ and ζ ′, ∥D,k, j(t, ·)∥L1 x′(Bc 2 j ) = 2kcn+1 R Rn−1 1 i2kt −1 \|y′\|2 ei(2ktσ+y′·ζ ′) × ζ ′ ∂ ∂σ exp −2 jη * 2k−2 jσi + \|ζ ′\|2 ψ(σ)φ(ζ ′)2kσ dζ ′dσ L1 y′(Bc 1). ∥D,k, j(t, ·)∥L1 x′(Bc 2 j ) Set q(σ, ζ ′; 2k−2 j) = 2k−2 jiσ + \|ζ ′\|2. Then, we have ∂ ∂σ exp −2 jη q(σ, ζ ′; 2k−2 j) ψ(σ)2kσφ(ζ ′) = exp −2 jη q(σ, ζ ′; 2k−2 j) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′). Maximal L1-regularity for parabolic Page 47 of 67 30 Maximal L1-regularity for parabolic Page 47 of 67 30 Page 47 of 67 30 J. Evol. Equ. 6. Almost orthogonality Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu J. Evol. Equ. + 2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ(ζ ′) . + 2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ(ζ ′) . As the case for k > 2 j before, we may summarize that all terms are including As the case for k > 2 j before, we may summarize that all terms are including 6. Almost orthogonality In order to take second derivative with respect to space, setting r = \|ζ ′\| and using the relation ζ ′ = ∂2 r + n−2 r ∂r, we have In order to take second derivative with respect to space, setting r = \|ζ ′\| and using the relation ζ ′ = ∂2 r + n−2 r ∂r, we have order to take second derivative with respect to space, setting r = \|ζ \| and using ation ζ ′ = ∂2 r + n−2 r ∂r, we have ζ ′ ∂ ∂σ exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 ψ(σ)2kσφ(ζ ′) = ζ ′ exp −2 jηq(σ, ζ ′; 2k−2 j) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′) = ∂r + n −2 r exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 jη 2k−2ℓi\|ζ ′\| 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ(ζ ′) = exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) × −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 jη 2k−2ℓi\|ζ ′\| 2q(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ(ζ ′) + −2 jη 2k−2 jiσ q(σ, ζ ′; 2k−2 j)3 φ(ζ ′) − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ′(ζ ′) + φ′′(ζ ′) × −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) 2 jη2k−2 ji\|ζ ′\| 2q(σ, ζ ′; 22 j−k)3 ψ(σ)2kσ + 2 jη−3\|ζ ′\|2 + q(σ, ζ ′; 2k−2 j)2 2q(σ, ζ ′; 2k−2 j)5 ψ(σ)2kσφ(ζ ′) + 2 jη 2k−2 ji\|ζ ′\| 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ′(ζ ′) + (n −2) exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) \|ζ ′\| −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k ζ r r r, ζ ′ ∂ ∂σ exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 ψ(σ)2kσφ(ζ ′) = ζ ′ exp −2 jηq(σ, ζ ′; 2k−2 j) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′) = ∂r + n −2 r exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 jη 2k−2ℓi\|ζ ′\| 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ(ζ ′) = exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) × −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 jη 2k−2ℓi\|ζ ′\| 2q(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ(ζ ′) + −2 jη 2k−2 jiσ q(σ, ζ ′; 2k−2 j)3 φ(ζ ′) − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ′(ζ ′) + φ′′(ζ ′) × −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) 2 jη2k−2 ji\|ζ ′\| 2q(σ, ζ ′; 22 j−k)3 ψ(σ)2kσ + 2 jη−3\|ζ ′\|2 + q(σ, ζ ′; 2k−2 j)2 2q(σ, ζ ′; 2k−2 j)5 ψ(σ)2kσφ(ζ ′) + 2 jη 2k−2 ji\|ζ ′\| 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ′(ζ ′) + (n −2) exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) \|ζ ′\| −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k ζ ′ ∂ ∂σ exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 ψ(σ)2kσφ(ζ ′) = ζ ′ exp −2 jηq(σ, ζ ′; 2k−2 j) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′) = ∂r + n −2 r exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 jη 2k−2ℓi\|ζ ′\| 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ(ζ ′) exp 2 jη q(σ ζ ′; 2k−2 j) ζ ′ ∂ ∂σ exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 ψ(σ)2kσφ(ζ ′) = ζ ′ exp −2 jηq(σ, ζ ′; 2k−2 j) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′) 2 = ∂r + n −2 r exp −2 jη q(σ, ζ ′; 2k−2 j) × − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 jη 2k−2ℓi\|ζ ′\| 2q(σ, ζ ′; 2k−2 j)3 ψ(σ)2kσφ(ζ ′) j ′ k 2 j × − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) − 2 jη\|ζ ′\| q(σ, ζ ′; 2k−2 j) φ(ζ ′) + φ′(ζ ′) × −2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 jη 2k−2ℓi\|ζ ′\| 2q(σ, ζ ′; 22 j−k)3 ψ(σ)2kσφ(ζ ′) 30 Page 48 of 67 T. As the case for k > 2 j before, we may summarize that all terms are including exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 = exp −2 jη q(σ, ζ ′; 2k−2 j) ψ(σ), φ(ζ ′) or their derivatives, and the order of derivatives is the same as the order of partial derivatives of σ and ζ ′. And 2 jη is multiplied by the same order of the partial derivatives at most, and one time at least. In the denominator, the function q(σ, ζ ′; 2k−2 j) = 2k−2 jiσ + \|ζ ′\|2 is estimated from below by 2−1 thanks to the cutoff function φ(ζ ′) or its derivative when j ≥1. Thus, if we derive it n + 2 times, the terms involving q(σ, ζ ′; 2k−2 j) are estimated from below by 2n+2. Moreover, q(σ, ζ ′; 2k−2 j) which arises by derivative contains surplus scale parameter 2k−2 j which is estimated from above by 1 by k < 2 j. Setting hk, j(σ, ζ ′, η) ≡exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 2kσ ψ(σ)φ(ζ ′) Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j) ≡exp 2 jη * 2k−2 jiσ + \|ζ ′\|2 Dn+2 σ,ζ ′ hℓ,k, j(σ, ζ ′, η). As the case for k > 2 j before, we may summarize that all terms are including Evol. Equ. J. Evol. Equ. Page 49 of 67 30 ≤C2k exp(−2 j−1η) 2k ⟨2kt⟩2 R Rn−1 e −2 j η q(σ,ζ ′;2k−2 j )−1/2 Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j) dζ ′dσ ≤C2k exp(−2 j−1η) 1 + C(2 jη))n+2 2k ⟨2kt⟩2 , which completes the proof of Lemma 6.1. As the case for k > 2 j before, we may summarize that all terms are including and and Dn+2 σ,ζ ′ ≡(1 − ζ ′) n 2 1 −∂2 ∂σ 2 , we obtain ∥D,k, j(t, ·)∥L1 x′ = cn+12k 1 ⟨y′⟩2 n 2 2k ⟨2kt⟩2 R Rn−1 ei(2ktσ+y′·ζ ′)(1 − ζ ′) n 2 1 −∂2 ∂σ 2 hk, j(σ, ζ ′, η)dζ ′dσ L1 y′ = cn+12k 1 \|y′\|2 n 2 2k ⟨2kt⟩2 R Rn−1 ei(2ktσ+y′·ζ ′) exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 × Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j)dζ ′dσ L1 y′ = cn+12kη2 exp(−2 j−1η) 2k ⟨2kt⟩2 × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 j η q(σ,ζ ′;2k−2 j )−1/2 Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j)dζ ′dσ L1 y′ ∥D,k, j(t, ·)∥L1 x′ = cn+12k 1 ⟨y′⟩2 n 2 2k ⟨2kt⟩2 R Rn−1 ei(2ktσ+y′·ζ ′)(1 − ζ ′) n 2 1 −∂2 ∂σ 2 hk, j(σ, ζ ′, η)dζ ′dσ L1 y′ = cn+12k 1 \|y′\|2 n 2 2k ⟨2kt⟩2 R Rn−1 ei(2ktσ+y′·ζ ′) exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 × Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j)dζ ′dσ L1 y′ = cn+12kη2 exp(−2 j−1η) 2k ⟨2kt⟩2 ∥D,k, j(t, ·)∥L1 x′ = cn+12k 1 ⟨y′⟩2 n 2 2k ⟨2kt⟩2 R Rn−1 ei(2ktσ+y′·ζ ′)(1 − ζ ′) n 2 1 −∂2 ∂σ 2 hk, j(σ, ζ ′, η)dζ ′dσ L1 y′ = cn+12k 1 \|y′\|2 n 2 2k ⟨2kt⟩2 R Rn−1 ei(2ktσ+y′·ζ ′) exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 × Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j)dζ ′dσ L1 y′ = cn+12kη2 exp(−2 j−1η) 2k ⟨2kt⟩2 ⟨2 t⟩ × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 j η q(σ,ζ ′;2k−2 j )−1/2 Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j)dζ ′dσ L1 y′ × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 j η q(σ,ζ ′;2k−2 j )−1/2 Hn+2 ℓ,k, j(σ, ζ ′, η, 2k−2 j)dζ ′dσ L1 y′ Maximal L1-regularity for parabolic Page 49 of 67 30 Maximal L1-regularity for parabolic Maximal L1-regularity for parabolic Page 49 of 67 30 Page 49 of 67 30 Page 49 of 67 30 J. which completes the proof of Lemma 6.1. □ 6.2. The second orthogonal estimate 6.2. The second orthogonal estimate For the estimating the term P D 2 in (4.21), it involving an η-convolution between the potential D and φm(η). We show the following second orthogonal estimate: Lemma 6.2. (Potential orthogonality 2) Let k, j, m ∈Z and assume j ≤m + 1. Let D(t, x′, η) be the potential of the solution for the Dirichlet data deﬁned by (3.11) and let {ψk(t)}k∈Z and {φ j(x)} j∈Z be a spatial and time Littlewood-Paley decomposition. Let D,k, j(t, x′, η) be deﬁned by (6.1). Then for any N ∈N, there exists a constant CN > 0 such that ∥(φm ∗ (η) D,k, j)(t, ·, η)∥L1 x′ ≤ ⎧ ⎪⎪⎪⎨ ⎪⎪⎪⎩ CN2k 2−\| k 2 −m\| ⟨2min( k 2 ,m)η⟩N 2k ⟨2kt⟩2 , k ≥2 j, CN2k 2−\| j−m\| ⟨2 jη⟩N 2k ⟨2kt⟩2 , k < 2 j. (6.6) (6.6) Proof of Lemma 6.2. Assuming 2 j < k, we show the t(time) dominated estimate. Changing τ = 2kσ, ξ′ = 2 jζ ′, as we have seen in (6.3), it also follows that Proof of Lemma 6.2. Assuming 2 j < k, we show the t(time) dominated estimate. Changing τ = 2kσ, ξ′ = 2 jζ ′, as we have seen in (6.3), it also follows that ∥φm ∗ (η) D,k, j(t, ·, η)∥L1 x′ ∥φm ∗ (η) D,k, j(t, ·, η)∥L1 x′ (η) = cn+1 R Rn−1 eitτ+ix′·ξ′iτ φm ∗ (η) exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ L1 x′ = cn+1 R Rn−1 ei2ktσ+i2 j x′·ζ ′i2kσ × φm ∗ (η) exp − * 2kiσ + 22 j\|ζ ′\|2η ψ(σ)φ(ζ ′)2(n−1) jdζ ′ · 2kdσ L1 x′ = cn+1 R Rn−1 ei2ktσ+i2 j x′·ζ ′iσ2k × φm ∗ (η) exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ψ(σ)φ(ζ ′)2(n−1) jdζ ′ · 2kdσ L1 x′ = cn+12k R Rn−1 ei2ktσ+iy′·ζ ′ φm ∗ (η) exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 σ ψ(σ)φ(ζ ′)dζ ′ · 2kdσ L1 y′ . = cn+1 R Rn−1 eitτ+ix′·ξ′iτ φm ∗ (η) exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ L1 x′ T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. which completes the proof of Lemma 6.1. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 50 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. Then using ei(2ktσ+y′·ζ ′) = −1 i2kt 1 \|y′\|2 ζ ′∂σei(2ktσ+y′·ζ ′) we apply the integration by parts by σ, ζ ′ as before. Then, the following lemma is a crucial step. we apply the integration by parts by σ, ζ ′ as before. Then, the following lemma is a crucial step. Lemma 6.3. Let k, j, m ∈Z and assume j < m + 1 and 2−1 ≤σ, \|ζ ′\| ≤2. For α = α1 + α2, 0 ≤α1 ≤1, 0 ≤α2 ≤n, the following estimates hold: φm ∗ (η) \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ≤ CN2−\| k 2 −m\| ⟨2min( k 2 ,m)η⟩N , k ≥2 j, (6.7) φm ∗ Cn α∂α1 σ ∂α2 ζ ′ exp −2 jη * i2k−2 jσ + \|ζ ′\|2 ≤CN2−\| j−m\| j N , k < 2 j. φm ∗ (η) \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ≤ CN2−\| k 2 −m\| ⟨2min( k 2 ,m)η⟩N , k ≥2 j, (6.7) φm ∗ (η) \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 jη * i2k−2 jσ + \|ζ ′\|2 ≤CN2−\| j−m\| ⟨2 jη⟩N , k < 2 j. (6.8) φm ∗ (η) \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ≤ CN2−\| k 2 −m\| ⟨2min( k 2 ,m)η⟩N , k ≥2 j, (6 7) (6.7) φm ∗ (η) \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 jη * i2k−2 jσ + \|ζ ′\|2 ≤CN2−\| j−m\| ⟨2 jη⟩N , k < 2 j. (6.8) (6.8) Proof of Lemma 6.3. Weﬁrstshowthelemmaforthecaseofk > 2 j.For p(σ, ζ ′, 22 j−k) ≡ iσ + 22 j−k\|ζ ′\|2, let f of Lemma 6.3. Weﬁrstshowthelemmaforthecaseofk > 2 j.For p(σ, ζ ′, 22 j−k) iσ + 22 j−k\|ζ ′\|2, let Pα(σ, ζ ′, 2k/2η) = exp 2 k 2 η * iσ + 22 j−k\|ζ ′\|2 \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 . which completes the proof of Lemma 6.1. Pα(σ, ζ ′, 2k/2η) = exp 2 k 2 η * iσ + 22 j−k\|ζ ′\|2 \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 . −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 . (6.9) (6.9) Then, Pα is a polynomial of 2 k 2 η and \|p(σ, ζ ′, 22 j−k)\|−1 = \| iσ + 22 j−k\|ζ ′\|2\|−1 and by (6.5), it holds that Then, Pα is a polynomial of 2 k 2 η and \|p(σ, ζ ′, 22 j−k)\|−1 = \| iσ + 22 j−k\|ζ ′\|2\|−1 and by (6.5), it holds that Then, Pα is a polynomial of 2 k 2 η and \|p(σ, ζ ′, 22 j−k)\|−1 = \| iσ + 22 j−k\|ζ ′\|2\|−1 and by (6.5), it holds that \|Pα(σ, ζ ′, 2k/2η)\| ≤C 1 + (2 k 2 η)n+2 . (6.10) We also set ⎧ ⎪⎨ ⎪⎩ ¯θ = 2 k 2 * iσ + 22 j−k\|ζ ′\|2θ = 2 k 2 pθ, ¯η = 2 k 2 * iσ + 22 j−k\|ζ ′\|2η = 2 k 2 pη. (6.11) (6.11) ⟨Step 1⟩: The case k ≥2m ≥2 j. Using (6.5), ⟨Step 1⟩: The case k ≥2m ≥2 j. Using (6.5), ⟨Step 1⟩: The case k ≥2m ≥2 j. Using (6.5), R+ φm(η −θ) Pα(σ, ζ ′, 2k/2θ) exp −2 k 2 θ * iσ + 22 j−k\|ζ ′\|2 dθ = R+ 2mφ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 (¯η −¯θ) ¯Pα( ¯θ) exp −¯θ 2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 d ¯θ ≤C2m−k 2 R φ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 (¯η −¯θ) ¯Pα( ¯θ) exp −\| ¯θ\| d ¯θ R+ φm(η −θ) Pα(σ, ζ ′, 2k/2θ) exp −2 k 2 θ * iσ + 22 j−k\|ζ ′\|2 dθ J. Evol. Equ. Maximal L1-regularity for parabolic Page 51 of 67 30 Maximal L1-regularity for parabolic Page 51 of 67 30 Maximal L1-regularity for parabolic Page 51 of 67 30 Page 51 of 67 30 J. Evol. Equ. J. Evol. Equ. which completes the proof of Lemma 6.1. ≤C2m−k 2 \| ¯θ\|> 1 2 \|¯η\| CN - 2m2−k 2 iσ + 22 j−k\|ζ ′\|2 −1(¯η −¯θ) .N ¯Pα( ¯θ) exp −\| ¯θ\| d ¯θ + C2m−k 2 \| ¯θ\|≤1 2 \|¯η\| CN - 2m2−k 2 iσ + 22 j−k\|ζ ′\|2 −1(¯η −¯θ) .N ˜Pα( ¯θ) exp −\| ¯θ\| d ¯θ ≡I + I I, (6.12 \| \| 2 \|η\| - i + \|ζ \| (η θ) . + C2m−k 2 \| ¯θ\|≤1 2 \|¯η\| CN - 2m2−k 2 iσ + 22 j−k\|ζ ′\|2 −1(¯η −¯θ) .N ˜Pα( ¯θ) exp −\| ¯θ\| d ¯θ ≡I + I I, (6.12) 2 - \|ζ \| (η ) . ≡I + I I, (6.12) (6.12) where we set ¯Pα( ¯θ) as a polynomial of 2 k 2 ¯θ and has the following estimate form (6.10); ¯Pα( ¯θ) ≤Cα 1 + \|2 k 2 ¯θ\| n+2 (6.13) 1 ≤σ < 2, 2−1 < \|ζ ′\| ≤2 and k ≥2 j. ¯Pα( ¯θ) ≤Cα 1 + \|2 k 2 ¯θ\| n+2 (6.13) ¯Pα( ¯θ) ≤Cα 1 + \|2 k 2 ¯θ\| n+2 (6.13) under the restrictions 2−1 ≤σ < 2, 2−1 < \|ζ ′\| ≤2 and k ≥2 j. (6.13) nder the restrictions 2−1 ≤σ < 2, 2−1 < \|ζ ′\| ≤2 and k ≥2 j. Then, the ﬁrst term of the right hand side of (6.12) is estimated by using (6.13) as follows: I ≤C2m−k 2 exp −1 4\|¯η\| \| ¯θ\|> 1 2 \|¯η\| CN - 2m2−k 2 iσ + 22 j−k\|ζ ′\|2 −1(¯η −¯θ) .N ¯Pα( ¯θ) exp −1 2\| ¯θ\| d ¯θ ≤C2m−k 2 CN ⟨2 k 2 iσ + 22 j−k\|ζ ′\|2 η⟩N \| ¯θ\|> 1 2 \|¯η\| ¯Pα( ¯θ) exp −1 2\| ¯θ\| d ¯θ ≤CN2m−k 2 - 2 k 2 η .N ≤CN2m−k 2 - 2mη .N (6.14) (6.14) by k ≥2m. For the second term in (6.12), we note that \| ¯θ\| < 1 2\|¯η\| implies \|¯η −¯θ\| \|¯η\| −\| ¯θ\| ≥\|¯η\| −1 2\|¯η\| = 1 2\|¯η\|, and it follows that by k ≥2m. which completes the proof of Lemma 6.1. and R+ φm(θ) Pα(τ, ξ′, 2k/2(η −θ)) exp −2 k 2 (η −θ) * iσ + 22 j−k\|ζ ′\|2 dθ = R+ φm(θ) Pα(τ, ξ′, 2k/2(η −θ)) exp −2 k 2 (η −θ) * iσ + 22 j−k\|ζ ′\|2 −Pα(τ, ξ′, 2k/2η) exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 dθ = R+ φm(θ) 1 0 d dν Pα(τ, ξ′, 2k/2(η −νθ)) × exp −2 k 2 (η −νθ) * iσ + 22 j−k\|ζ ′\|2 dν dθ = 1 0 R+ φm(θ)2 k 2 θ * iσ + 22 j−k\|ζ ′\|2 × Pα(τ, ξ′, 2k/2(η −νθ)) −∂μPα(τ, ξ′, μ) μ=2 k 2 (η−νθ) × exp −2 k 2 (η −νθ) * iσ + 22 j−k\|ζ ′\|2 dθdν = 1 0 R 2mφ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 ¯θ ¯θ × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −(¯η −ν ¯θ) 2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 d ¯θdν ≤ 1 0 \| ¯θ\|> 1 2 \|¯η\| φ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 ¯θ 2m−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 \| ¯θ\| × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν + 1 0 \| ¯θ\|≤1 2 \|¯η\| φ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 ¯θ 2m−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 \| ¯θ\| × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν ≡I I I + I V, (6.17) and R+ φm(θ) Pα(τ, ξ′, 2k/2(η −θ)) exp −2 k 2 (η −θ) * iσ + 22 j−k\|ζ ′\|2 dθ = R+ φm(θ) Pα(τ, ξ′, 2k/2(η −θ)) exp −2 k 2 (η −θ) * iσ + 22 j−k\|ζ ′\|2 −Pα(τ, ξ′, 2k/2η) exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 dθ = R+ φm(θ) 1 0 d dν Pα(τ, ξ′, 2k/2(η −νθ)) R+ φm(θ) Pα(τ, ξ′, 2k/2(η −θ)) exp −2 k 2 (η −θ) * iσ + 22 j−k\|ζ ′\|2 dθ = R+ φm(θ) Pα(τ, ξ′, 2k/2(η −θ)) exp −2 k 2 (η −θ) * iσ + 22 j−k\|ζ ′\|2 −Pα(τ, ξ′, 2k/2η) exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 dθ R+ = R+ φm(θ) Pα(τ, ξ′, 2k/2(η −θ)) exp −2 k 2 (η −θ) * iσ + 22 j−k\|ζ ′\|2 P (τ ξ′ 2k/2η) exp 2 k 2 η * iσ + 22 j−k\|ζ ′\|2 dθ = R+ φm(θ) 1 0 d dν Pα(τ, ξ′, 2k/2(η −νθ)) × exp −2 k 2 (η −νθ) * iσ + 22 j−k\|ζ ′\|2 dν dθ p (η ) * + \|ζ \| = 1 0 R+ φm(θ)2 k 2 θ * iσ + 22 j−k\|ζ ′\|2 × Pα(τ, ξ′, 2k/2(η −νθ)) −∂μPα(τ, ξ′, μ) μ=2 k 2 (η−νθ) × exp −2 k 2 (η −νθ) * iσ + 22 j−k\|ζ ′\|2 dθdν = 1 0 R 2mφ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 ¯θ ¯θ × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −(¯η −ν ¯θ) 2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 d ¯θdν ≤ 1 0 \| ¯θ\|> 1 2 \|¯η\| φ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 ¯θ 2m−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν + 1 0 \| ¯θ\|≤1 2 \|¯η\| φ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 ¯θ 2m−k 2 * iσ + 22 j−k\|ζ ′\|2 − × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν = 1 0 R+ φm(θ)2 k 2 θ * iσ + 22 j−k\|ζ ′\|2 × Pα(τ, ξ′, 2k/2(η −νθ)) −∂μPα(τ, ξ′, μ) μ=2 k 2 (η−νθ) × exp −2 k 2 (η −νθ) * iσ + 22 j−k\|ζ ′\|2 dθdν = 1 0 R 2mφ 2m2−k 2 * iσ + 22 j−k\|ζ ′\|2 −1 ¯θ ¯θ (6.17) (6.17) where we use the variables deﬁned in (6.11) and set where we use the variables deﬁned in (6.11) and set where we use the variables deﬁned in (6.11) and set Pα(τ, ζ ′, (¯η −ν ¯θ)) ≡Pα(τ, ξ′, 2k/2(η −νθ)) −∂μPα(τ, ξ′, μ) μ=2 k 2 (η−νθ) =Pα(τ, ξ′, \|p\|−1(¯η −ν ¯θ)) −∂μPα(τ, ξ′, μ) μ=\|p\|−1(¯η−ν ¯θ). = 1 0 R+ which completes the proof of Lemma 6.1. For the second term in (6.12), we note that \| ¯θ\| < 1 2\|¯η\| implies \|¯η −¯θ\| ≥ \|¯η\| −\| ¯θ\| ≥\|¯η\| −1 2\|¯η\| = 1 2\|¯η\|, and it follows that I I ≤C2m−k 2 \| ¯θ\|≤1 2 \|¯η\| CN - 2m2−k 2 iσ + 22 j−k\|ζ ′\|2 −1(¯η −¯θ) .N ¯Pα( ¯θ) exp −\| ¯θ\| d ¯θ ≤ CN2m−k 2 - 2m−12−k 2 iσ + 22 j−k\|ζ ′\|2 −1\|¯η\| .N R ¯Pα( ¯θ) exp −\| ¯θ\| d ¯θ ≤CN2m−k 2 - 2m−1η .N . (6.15) I I ≤C2m−k 2 \| ¯θ\|≤1 2 \|¯η\| CN - 2m2−k 2 iσ + 22 j−k\|ζ ′\|2 −1(¯η −¯θ) .N ¯Pα( ¯θ) exp −\| ¯θ\| d ¯θ ≤ CN2m−k 2 - 2m−12−k 2 iσ + 22 j−k\|ζ ′\|2 −1\|¯η\| .N R ¯Pα( ¯θ) exp −\| ¯θ\| d ¯θ k ≤CN2m−k 2 - 2m−1η .N . (6.15) (6.15) ≤CN2 2 - 2m−1η .N . (6.15) Hence by (6.12), (6.14) and (6.15), we obtain Hence by (6.12), (6.14) and (6.15), we obtain R+ φm(η−θ) Pα(τ, ξ′, 2k/2θ) exp −2 k 2 θ * iσ +22 j−k\|ζ ′\|2 dθ ≤CN2m−k 2 - 2m−1η .N . (6.16) Step 2⟩: When 2 j ≤k ≤2m, we note that ⟨Step 2⟩: When 2 j ≤k ≤2m, we note that R φm(θ)dθ = 0 T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. which completes the proof of Lemma 6.1. Since φ is rapidly decreasing function, for any N ∈N, there exists a constant CN > 0 and the ﬁrst term of the right hand side of (6.17) is J. Evol. Equ. Maximal L1-regularity for parabolic Page 53 of 67 30 Maximal L1-regularity for parabolic Page 53 of 67 30 Maximal L1-regularity for parabolic Page 53 of 67 30 Page 53 of 67 30 J. Evol. Equ. I I I ≤ 1 0 \| ¯θ\|> 1 2 \|¯η\| CN\|2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯θ\| ⟨2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯θ⟩2N Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν ≤ CN2−m2 k 2 ⟨2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯η⟩N 1 0 \| ¯θ\|> 1 2 \|¯η\| (2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1)2\|\| ¯θ\| ⟨2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯θ⟩N × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν ≤CN2−m+ k 2 ⟨2mη⟩N 1 0 R x ⟨x⟩N dxdν ≤CN2−m+ k 2 ⟨2mη⟩N . (6.18) I I I ≤ 1 0 \| ¯θ\|> 1 2 \|¯η\| CN\|2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯θ\| ⟨2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯θ⟩2N Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν ≤ CN2−m2 k 2 ⟨2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯η⟩N 1 0 \| ¯θ\|> 1 2 \|¯η\| (2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1)2\|\| ¯θ\| ⟨2m2−k 2 iσ + 22 j−k\|ζ ′\|2−1 ¯θ⟩N × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν ≤CN2−m+ k 2 ⟨2mη⟩N 1 0 R x ⟨x⟩N dxdν ≤CN2−m+ k 2 ⟨2mη⟩N . (6.18) (6.18) For the second term in (6.17), we prepare the following simple estimate: For the second term in (6.17), we prepare the following simple estimate: For N = 2, 3, . . . and a > 0, Lemma 6.4. For N = 2, 3, . . . and a > 0, \|x\|≤a dx (1 + \|x\|2)N/2 ≤ √ 2πa (1 + \|a\|2)1/2 .. (6.19) (6.19) Proof of Lemma 6.4. Applying the integration by parts, Proof of Lemma 6.4. which completes the proof of Lemma 6.1. Applying the integration by parts, \|x\|≤a dx (1 + \|x\|2)N/2 ≤ \|x\|≤a dx = 2a ≤ 2 √ 2a (1 + \|a\|2)1/2 for a ≤1, \|x\|≤a dx (1 + \|x\|2)N/2 ≤ \|x\|≤a dx = 2a ≤ 2 √ 2a (1 + \|a\|2)1/2 for a ≤1, while while \|x\|≤a dx (1 + \|x\|2)N/2 ≤ \|x\|≤a dx 1 + \|x\|2 = 2 tan−1 a ≤ √ 2πa (1 + \|a\|2)1/2 for a ≥1. This shows the estimate. □ □ This shows the estimate. which completes the proof of Lemma 6.1. (6.20) ≤CN2−m+ k 2 exp −2−22−m+ k 2 \|p\|\|¯¯η\| \| ¯¯θ\|< 1 2 \|¯¯η\| 1 ⟨¯¯θ⟩N d ¯¯θ ≤CN2−m+ k 2 exp −2−22−m+ k 2 \|p\|\|¯¯η\| \|¯¯η\| (1 + \|¯¯η\|2)1/2 k ≤CN2−m+ 2 ⟨2 k 2 η⟩N for N ≥2. (6.20) (6.20) Hence, we obtain from (6.18) and (6.20) that Hence, we obtain from (6.18) and (6.20) that R+ φm(η −θ) Pα(τ, ξ′, 2k/2θ) exp −2 k 2 θ * iσ + 22 j−k\|ζ ′\|2 dθ ≤CN2−m+ k 2 ⟨2 k 2 η⟩N . (6.21) (6.21) The estimates (6.16) and (6.21) yield (6.7). ⟨Step 3⟩The case k < 2 j: p ⟩ j show (6.8), we use q(σ, ζ ′, 22 j−k) ≡ i2k−2 jσ + \|ζ ′\|2 instead of (6.9) and let Qα(σ, ζ ′, η, 2 j) = exp 2 jη * i2k−2 jσ + \|ζ ′\|2 \|α\|≤n+2 Cn,α∂α1 σ ∂α2 ζ ′ exp −2 jη * i2k−2 jσ + \|ζ ′\|2 . Since \|Qα\| is a polynomial of η and \|q(σ, ζ ′, 22 j−k)\|−1, it follows from the assumption that Since \|Qα\| is a polynomial of η and \|q(σ, ζ ′, 22 j−k)\|−1, it follows from the assumption that 2−1 ≤ q(σ, ζ ′, 2k−2 j) = 22(k−2 j)σ 2 + \|ζ ′\|41/4 ≤201/4. Hence as in the previous step, Hence as in the previous step, \|Qα(τ, ξ′, 2 jη)\| ≤C 1 + (2 jη)n+2 . All the other estimate is very similar to the case of ⟨Step 2⟩and all the terms involving 2 k 2 are arranged into 2 j. The estimate corresponding to the case ⟨Step 1⟩is redundant by the assumption j < m. □ Proof of Lemma 6.2, continued. The proof of Lemma 6.2 goes in a similar way to the case of Lemma 6.1. After integrating by parts, the integrable factors ⟨t⟩−2 and ⟨y′⟩−n appear and then estimate the integrant to obtain the desired estimate in Lemma 6.3 for the case 2 j < k. The other case 2 j ≥k is also obtained from (6.8) in Lemma 6.3. This complete the proof of Lemma 6.2. □ 6.3. The Neumann potential case which completes the proof of Lemma 6.1. Using \|¯η−ν ¯θ\| ≥\|¯η\|−\| ¯θ\| ≥1 2\|¯η\| under the restriction \| ¯θ\| ≤1 2\|¯η\| and the estimates (6.5), (6.10) and (6.19), Using \|¯η−ν ¯θ\| ≥\|¯η\|−\| ¯θ\| ≥1 2\|¯η\| under the restriction \| ¯θ\| ≤1 2\|¯η\| and the estimates (6.5), (6.10) and (6.19), I V ≤ 1 0 \| ¯θ\|≤1 2 \|¯η\| CN\|2m2−k 2 \| iσ + 22 j−k\|ζ ′\|2\|−1 ¯θ\| ⟨2m2−k 2 \| iσ + 22 j−k\|ζ ′\|2\|−1 ¯θ⟩N+1 × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp − × Pα(τ, ζ , (η νθ)) exp \|η νθ\| dθdν ≤CN2−m+ k 2 1 0 \| ¯θ\|≤1 2 \|¯η\| (2m2−k 2 \| iσ + 22 j−k\|ζ ′\|2\|−1)2\| ¯θ\| ⟨2m2−k 2 \| iσ + 22 j−k\|ζ ′\|2\|−1 ¯θ⟩N+1 × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν (by changing ¯¯θ ≡2m2−k 2 \|p\|−1 ¯θ, ¯¯η ≡2m2−k 2 \|p\|−1 ¯η) ≤CN2−m+ k 2 1 0 \| ¯¯θ\|< 1 2 \|¯¯η\| \| ¯¯θ\| ⟨¯¯θ⟩N+1 exp −2−12−m+ k 2 \|p\|\|¯¯η −ν ¯¯θ\| d ¯¯θdν ≤CN2−m+ k 2 1 0 \| ¯¯θ\|< 1 2 \|¯¯η\| 1 ⟨¯¯θ⟩N exp −2−12−m+ k 2 \|p\|\|¯¯η −ν ¯¯θ\| d ¯¯θdν ≤CN2−m+ k 2 1 0 \| ¯θ\|≤1 2 \|¯η\| (2m2−k 2 \| iσ + 22 j−k\|ζ ′\|2\|−1)2\| ¯θ\| ⟨2m2−k 2 \| iσ + 22 j−k\|ζ ′\|2\|−1 ¯θ⟩N+1 × Pα(τ, ζ ′, (¯η −ν ¯θ)) exp −\|¯η −ν ¯θ\| d ¯θdν (by changing ¯¯θ ≡2m2−k 2 \|p\|−1 ¯θ, ¯¯η ≡2m2−k 2 \|p\|−1 ¯η) (by changing ¯¯θ ≡2m2−k 2 \|p\|−1 ¯θ, ¯¯η ≡2m2−k 2 \|p\|−1 ¯η) ≤CN2−m+ k 2 1 0 \| ¯¯θ\|< 1 2 \|¯¯η\| \| ¯¯θ\| ⟨¯¯θ⟩N+1 exp −2−12−m+ k 2 \|p\|\|¯¯η −ν ¯¯θ\| d ¯¯θdν ≤CN2−m+ k 2 1 0 \| ¯¯θ\|< 1 2 \|¯¯η\| 1 ⟨¯¯θ⟩N exp −2−12−m+ k 2 \|p\|\|¯¯η −ν ¯¯θ\| d ¯¯θdν T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 54 of 67 T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. ≤CN2−m+ k 2 exp −2−22−m+ k 2 \|p\|\|¯¯η\| \| ¯¯θ\|< 1 2 \|¯¯η\| 1 ⟨¯¯θ⟩N d ¯¯θ ≤CN2−m+ k 2 exp −2−22−m+ k 2 \|p\|\|¯¯η\| \|¯¯η\| (1 + \|¯¯η\|2)1/2 ≤CN2−m+ k 2 exp −2−2\|¯η\| ≤CN2−m+ k 2 ⟨2 k 2 η⟩N for N ≥2. 6.3. The Neumann potential case The almost orthogonal estimate for the Neumann boundary potential is very similar to the case of Dirichlet potential case except the order of the derivative. The following lemma shows the estimates (5.6) hold valid. Page 55 of 67 30 Page 55 of 67 30 Maximal L1-regularity for parabolic J. Evol. Equ. J. Evol. Equ. Lemma 6.5 (A crucial potential orthogonality). For k, j, ℓ∈Z let {ψk(t)}k∈Z and {φ j(x)} j∈Z be the time and the space Littlewood-Paley dyadic decomposition and let N(t, x′, η) be the boundary potential deﬁned in (3.20). Set N,k, j(t, x′, η) ≡ R Rn−1 N(t −s, x′ −y′, η)ψk(s)φ j(y′)dy′ds for η = xn ∈Iℓ= [2−ℓ, 2−ℓ+1). Then, there exists a constant Cn > 0 depending only on the dimension n satisfying for η = xn ∈Iℓ= [2−ℓ, 2−ℓ+1). Then, there exists a constant Cn > 0 depending only on the dimension n satisfying ∥N,k, j(t, ·, η)∥L1 x′ ≤ ⎧ ⎪⎪⎪⎨ ⎪⎪⎪⎩ Cn2 k 2 1 + (2 k 2 η)n+2 e−2 k 2 −1η 2k ⟨2kt⟩2 , k ≥2 j, Cn2 k 2 1 + (2 jη)n+2 e−2 j−1η 2k ⟨2kt⟩2 , k < 2 j, (6.22) (6.22) and ∥φm ∗ (η) N,k, j(t, ·, η)∥L1 x′ ≤ ⎧ ⎪⎪⎪⎨ ⎪⎪⎪⎩ Cn2 k 2 2−\| k 2 −m\| ⟨2min( k 2 ,m)η⟩N 2k ⟨2kt⟩2 , k ≥2 j, Cn2 k 2 2−\| j−m\| ⟨2 jη⟩N 2k ⟨2kt⟩2 , k < 2 j. (6.23) (6.23) The proof for Lemma 6.5 is shown in a parallel way to the proof of Lemma 6.1. The only difference stems from the difference of two potentials D and N of boundary data and the difference from (6.3) reﬂects the order of spatial derivatives appearing the Fourier image of the following expression: N,k, j(t, x′, η) = −cn+1 R Rn−1 eitτ+ix′·ξ′ iτ iτ + \|ξ′\|2 exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ. = −cn+1 R Rn−1 eitτ+ix′·ξ′ iτ iτ + \|ξ′\|2 exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ. Proof of Lemma 6.5. We only show the out-lined proof of (6.22). The other esti- mate (6.23) follows very much in a similar way to the proof of the Dirichlet case in Lemma 6.2. Proof of Lemma 6.5. We only show the out-lined proof of (6.22). 6.3. The Neumann potential case The other esti- mate (6.23) follows very much in a similar way to the proof of the Dirichlet case in Lemma 6.2. We consider a time-like estimate k ≥2 j as is in the Dirichlet case. Taking ζ ′-space cutoff, we have by using the change of variables τ = 2kσ, ξ′ = 2 jζ ′ that ∥N,k, j(t, ·, η)∥L1 x′ x = cn+1 R Rn−1 eitτ+ix′·ξ′ iτ iτ + \|ξ′\|2 × exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ L1 x′ = 2 k 2 cn+1 R Rn−1 ei2ktσ+iy′·ζ ′ σ iσ + 22 j−k\|ζ ′\|2 × exp −2 k 2 η * iσ + 22 j−k\|ζ ′\|2 ψ(σ)φ(ζ ′)dζ T. Ogawa and S. Shimizu J. Evol. Equ. T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. where we set x′ = 2−j y′ in the last line. Using (6.4) and integrating by parts, we see by setting p(σ, ζ ′; 22 j−k) ≡ iσ + 22 j−k\|ζ ′\|2 that here we set x′ = 2−j y′ in the last line. Using (6.4) and integrating by parts, we see y setting p(σ, ζ ′; 22 j−k) ≡ iσ + 22 j−k\|ζ ′\|2 that where we set x′ = 2−j y′ in the last line. Using (6.4) and integrating by parts, we see by setting p(σ, ζ ′; 22 j−k) ≡ iσ + 22 j−k\|ζ ′\|2 that ∥N,k, j(t, ·, η)∥L1 x′(Bc 2 j ) ∥N,k, j(t, ·, η)∥L1 x′(Bc 2 j ) = 2 k 2 cn+1 R Rn−1 1 i2kt −1 \|y′\|2 ei(2ktσ+y′·ζ ′) × ζ ′ ∂ ∂σ / 1 p(σ, ζ ′; 22 j−k) exp −2 1 2 kη p(σ, ζ ′; 22 j−k) ψ(σ)φ(ζ ′)2kσ 0 dζ ′dσ L1 y′(Bc 1) . Since Since Since ∂ ∂σ / 1 p(σ, ζ ′; 22 j−k) exp −2 k 2 η p(σ, ζ ′; 22 j−k) ψ(σ)2kσφ(ζ ′) 0 = 1 p(σ, ζ ′; 22 j−k) exp −2 k 2 η p(σ, ζ ′; 22 j−k) × − 2 k 2 η i 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ − i 2p(σ, ζ ′; 22 j−k)2 ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k φ(ζ ′). 6.3. The Neumann potential case Under the condition k ≥2 j, i.e., the time-like condition 22 j−k ≤1, we obtain the following estimate: Under the condition k ≥2 j, i.e., the time-like condition 22 j−k ≤1, we obtain the following estimate: ∥N,k, j(t, ·, η)∥L1 x′ x′ = 2 k 2 1 ⟨y′⟩2 n 2 2k ⟨2kt⟩2 cn+1 R Rn−1 ei(2ktσ+y′·ζ ′)(1 − ζ ′) n 2 1 −∂2 ∂σ 2 kk, j(σ, ζ ′, η)dζ ′dσ L1 y′ = C 2 k 2 exp(−2 k 2 −1η) 2k ⟨2kt⟩2 x′ = 2 k 2 1 ⟨y′⟩2 n 2 2k ⟨2kt⟩2 cn+1 R Rn−1 ei(2ktσ+y′·ζ ′)(1 − ζ ′) n 2 1 −∂2 ∂σ 2 kk, j(σ, ζ ′, η)dζ ′dσ L1 y′ = C 2 k 2 exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 k 2 η p(σ,ζ ′;22 j−k)−1 2 K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ ≤C 2 k 2 exp(−2 k 2 −1η) 2k ⟨2kt⟩2 R Rn−1 e −2 k 2 η Re√ iσ+22 j−k\|ζ ′\|2−1 2 K n+2 k, j (σ, ζ ′, η, 22 j−k) dζ ′dσ ≤C2 k 2 exp(−2 k 2 −1η) 1 + (2 k 2 η)n+2 2k ⟨2kt⟩2 . = C 2 k 2 exp(−2 k 2 −1η) 2k ⟨2kt⟩2 × 1 \|y′\|2 n 2 R Rn−1 ei(2ktσ+y′·ζ ′)e −2 k 2 η p(σ,ζ ′;22 j−k)−1 2 K n+2 k, j (σ, ζ ′, η, 22 j−k)dζ ′dσ L1 y′ ≤C 2 k 2 exp(−2 k 2 −1η) 2k ⟨2kt⟩2 R Rn−1 e −2 k 2 η Re√ iσ+22 j−k\|ζ ′\|2−1 2 K n+2 k, j (σ, ζ ′, η, 22 j−k) dζ ′dσ ≤C2 k 2 exp(−2 k 2 −1η) 1 + (2 k 2 η)n+2 2k ⟨2kt⟩2 . For the case of the space-like region k < 2 j, we proceed similar way. 6.3. The Neumann potential case we take second derivative by setting r = \|ζ ′\| and ζ ′ = ∂2 r + n−2 r ∂r to have ζ ′ ∂ ∂σ / 1 p(σ, ζ ′; 22 j−k) exp −2 k 2 η p(σ, ζ ′; 22 j−k) ψ(σ)2kσφ(ζ ′) 0 = ∂r + n −2 r exp −2 k 2 η p(σ, ζ ′; 22 j−k) p(σ, ζ ′; 22 j−k) × / −2 k 2 η22 j−k \|ζ ′\| p(σ, ζ ′; 22 j−k) φ(ζ ′) − 22 j−k \|ζ ′\| p(σ, ζ ′; 22 j−k)2 φ(ζ ′) + φ′(ζ ′) × − 2 k 2 η i 2p(σ, ζ ′; 22 j−k) ψ(σ)2kσ − i 2p(σ, ζ ′; 22 j−k)2 ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k + 2 k 2 η 22 j−k\|ζ ′\|i 2p(σ, ζ ′; 22 j−k)3 + 22 j−k\|ζ ′\|i p(σ, ζ ′; 22 j−k)4 ψ(σ)2kσφ(ζ ′) 0 . ( (6.24) As the Dirichlet boundary case, all terms again have exp −2 k 2 ηp(σ, ζ ′; 22 j−k) , with ψ(σ), φ(ζ ′) or their derivatives. The function in denominator p(σ, ζ ′; 22 j−k) = iσ + 22 j−k\|ζ ′\|2 is estimated from below by 2−1 thanks to the cutoff function φ(ζ ′) or its derivative. Therefore, if we derivate n + 2 times, then it is estimated from below by 2n+2. Moreover, p(σ, ζ ′; 22 j−k) which arises by derivative contains surplus scale parameter 22 j−k and it is estimated from above by 1 by k ≥2 j. Thus for the functions J. Evol. Equ. Maximal L1-regularity for parabolic Page 57 of 67 30 J. Evol. Equ. Maximal L1-regularity for parabolic Page 57 of 67 30 Maximal L1-regularity for parabolic Page 57 of 67 30 Maximal L1-regularity for parabolic Page 57 of 67 30 Page 57 of 67 30 J. Evol. Equ. kk, j(σ, ζ ′, η) ≡exp −2 k 2 η p(σ, ζ ′; 22 j−k) p(σ, ζ ′; 22 j−k) ψ(σ)φ(ζ ′)2kσ, K n+2 k, j (σ, ζ ′, η, 22 j−k) ≡exp 2 k 2 η p(σ, ζ ′; 22 j−k) Dn+2 σ,ζ ′ kk, j(σ, ζ ′, η), where we put where we put Dn+2 σ,ζ ′ ≡(1 − ζ ′) n 2 1 −∂2 ∂σ 2 . 6.3. The Neumann potential case Setting p(σ, ζ ′; 22 j−k) = iσ + 22 j−k\|ζ ′\|2, q(σ, ζ ′; 2k−2 j) = 2k−2 jiσ + \|ζ ′\|2, we see that ∂ ∂σ 1 exp −2 jη q(σ, ζ ′; 2k−2 j) p(σ, ζ ′; 22 j−k) ψ(σ)2kσφ(ζ ′) 2 = exp −2 jη q(σ, ζ ′; 2k−2 j) p(σ, ζ ′; 22 j−k) × / − 2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) − i 2p(σ, ζ ′; 22 j−k)2 ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k 0 φ(ζ ′). Setting Setting p(σ, ζ ′; 22 j−k) = iσ + 22 j−k\|ζ ′\|2, q(σ, ζ ′; 2k−2 j) = 2k−2 jiσ + \|ζ ′\|2, we see that ∂ ∂σ 1 exp −2 jη q(σ, ζ ′; 2k−2 j) p(σ, ζ ′; 22 j−k) ψ(σ)2kσφ(ζ ′) 2 = exp −2 jη q(σ, ζ ′; 2k−2 j) p(σ, ζ ′; 22 j−k) × / − 2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) − i 2p(σ, ζ ′; 22 j−k)2 ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k 0 φ(ζ ′). Taking the second derivative in r = \|ζ ′\| Taking the second derivative in r = \|ζ ′\| ζ ′ ∂ ∂σ 1 σ iσ + 22 j−k\|ζ ′\|2 exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 ψ(σ)2kσφ(ζ ′) 2 = ζ ′ exp −2 jη q(σ, ζ ′; 2k−2 j) p(σ, ζ ′; 22 j−k) × / − 2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) − i 2p(σ, ζ ′; 22 j−k)2 ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k 0 φ(ζ ′). Similarly before all terms again have exp −2 k 2 ηq(σ, ζ ′; 2k−2 j) , ψ(σ), φ(ζ ′) or their derivatives. The denominators q(σ, ζ ′; 2k−2 j) and p(σ, ζ ′, 22 j−k) are estimated from below by 2−1 thanks to the cutoff functions ψ(σ) and φ(ζ ′) or its derivative. Moreover, p(σ, ζ ′; 22 j−k) is estimated from above by 1 by k ≤2 j. 6.3. The Neumann potential case By changing τ = 2kσ, ξ′ = 2 jζ ′, τ = 2kσ, ξ′ = 2 jζ ′, ∥N,k, j(t, ·, η)∥L1 x′ ∥N,k, j(t, ·, η)∥L1 x′ = cn+1 R Rn−1 eitτ+ix′·ξ′ τ iτ + \|ξ′\|2 exp − iτ + \|ξ′\|2η ψ(2−kτ)φ(2−jξ′)dξ′dτ L1 x′ = 2 k 2 cn+1 R Rn−1 ei2ktσ+iy′·ζ ′ σ iσ + 22 j−k\|ζ ′\|2 exp −2 jη * 2k−2 jiσ + \|ζ ′\|2 × ψ(σ)φ(ζ ′)dζ ′ · 2kdσ L1 y′ . × ψ(σ)φ(ζ ′)dζ ′ · 2kdσ L1 y′ . Here, we set x′ = 2−j y′ in the last line. Using (6.4) and integrating by parts, y g ( ) g g y p ∥N,k, j(t, ·, η)∥L1 x′(Bc 2 k 2 ) ∥N,k, j(t, ·, η)∥L1 x′(Bc 2 k 2 ) 2 2 = 2 k 2 cn+1 R Rn−1 1 i2kt −1 \|y′\|2 ei(2ktσ+y′·ζ ′) × ζ ′ ∂ ∂σ 1 σ iσ + 22 j−k\|ζ ′\|2 exp −2 jη * 2k−2 jσi + \|ζ ′\|2 ψ(σ)φ(ζ ′)2k 0 dζ ′dσ L T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 58 of 67 T. Ogawa and S. Shimizu Setting p(σ, ζ ′; 22 j−k) = iσ + 22 j−k\|ζ ′\|2, q(σ, ζ ′; 2k−2 j) = 2k−2 jiσ + \|ζ ′\|2, we see that ∂ ∂σ 1 exp −2 jη q(σ, ζ ′; 2k−2 j) p(σ, ζ ′; 22 j−k) ψ(σ)2kσφ(ζ ′) 2 = exp −2 jη q(σ, ζ ′; 2k−2 j) p(σ, ζ ′; 22 j−k) × / − 2 jη 2k−2 ji 2q(σ, ζ ′; 2k−2 j) − i 2p(σ, ζ ′; 22 j−k)2 ψ(σ)2kσ + ψ′(σ)2kσ + ψ(σ)2k 0 φ(ζ ′). 6.3. The Neumann potential case Thus introducing hk, j(σ, ζ ′, η) ≡exp −2 jη 2k−2 jσi + \|ζ ′\|2 σi + 22 j−k\|ζ ′\|2 2kσ ψ(σ)φ(ζ ′), Hn+2 k, j (σ, ζ ′, η, 2k−2 j) ≡exp −2 jη 2k−2 jσi + \|ζ ′\|2 σi + 22 j−k\|ζ ′\|2 Dn+2 σ,ζ ′ hk, j(σ, ζ ′, η), we obtain J. Evol. Equ. Maximal L1-regularity for parabolic Page 59 of 67 30 Maximal L1-regularity for parabolic Page 59 of 67 30 Maximal L1-regularity for parabolic Page 59 of 67 30 Page 59 of 67 30 J. Evol. Equ. ∥N,k, j(t, ·, η)∥L1 x′ = 2 k 2 1 ⟨y′⟩2 n 2 2k ⟨2kt⟩2 cn+1 R Rn−1 ei(2ktσ+y′·ζ ′)(1 − ζ ′) n 2 1 −∂2 ∂σ 2 hℓ,k, j(σ, ζ ′, η)dζ ′dσ L1 y′ ≤C 2 k 2 exp(−2 j−1η) 2k ⟨2kt⟩2 R Rn−1 e −2 j η Re q(σ,ζ ′;2k−2 j )−1 Hn+2 k, j (σ, ζ ′, η, 2k−2 j) dζ ′dσ ≤C 2 k 2 exp(−2 j−1η) 1 + (2 jη)n+2 2k ⟨2kt⟩2 . y ≤C 2 k 2 exp(−2 j−1η) 2k ⟨2kt⟩2 R Rn−1 e −2 j η Re q(σ,ζ ′;2k−2 j )−1 Hn+2 k, j (σ, ζ ′, η, 2k−2 j) dζ ′dσ ≤C 2 k 2 exp(−2 j−1η) 1 + (2 jη)n+2 2k ⟨2kt⟩2 . This completes the proof of (6.22) in Lemma 6.5. This completes the proof of (6.22) in Lemma 6.5. This completes the proof of (6.22) in Lemma 6.5. This completes the proof of (6.22) in Lemma 6.5. □ This completes the proof of (6.22) in Lemma 6.5. □ □ 7. Optimal boundary trace estimates (7 2) (7.2) Using the assumption u(0, x′, η) = 0 almost everywhere, Using the assumption u(0, x′, η) = 0 almost everywhere, Using the assumption u(0, x′, η) = 0 almost everywhere, ψk(t) ∗ (t) u(t, x′, η) = − R+ ∂s ∞ s ψk(t −r)dr u(s, x′, η)ds = − ∞ s ψk(t −r)dr u(s, x′, η) ∞ s=0 + R+ ∞ s ψk(t −r)dr∂su(s, x′, η)ds = ∂−1 t ψk(t) ∗ (t) ∂tu(t, x′, η), + = ∂−1 t ψk(t) ∗ (t) ∂tu(t, x′, η), (7.3) (7.3) where we set ∂−1 t ψk(t −s) ≡ t−s −∞ ψk(r)dr = ∞ s ψk(t −r)dr. (7.4) (7.4) Here, we recall the Littlewood-Paley decomposition of direct sum type deﬁned in (4.7).Since∂−1 t ψk = 2−k(∂−1 t ψ)k isalsorapidlydecreasingsmoothfunction,ζ j(η) ∗ (η) ζ j−1(η) = ζ j−1(η)bydeﬁnition (4.6)andusing(4.8),theHausdorff–Younginequality with (7.3), there exists a constant C > 0 such that Here, we recall the Littlewood-Paley decomposition of direct sum type deﬁned in (4.7).Since∂−1 t ψk = 2−k(∂−1 t ψ)k isalsorapidlydecreasingsmoothfunction,ζ j(η) ∗ (η) ζ j−1(η) = ζ j−1(η)bydeﬁnition (4.6)andusing(4.8),theHausdorff–Younginequality with (7.3), there exists a constant C > 0 such that I = j∈Z k≥2 j 2(1−1/2p)k2sj∂−1 t ψk ∗ (t) φ j ∗ (x′) ∂tu(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2(1−/2p)k\|2−k(∂−1 t ψ)k(t)\| ∗ (t) φ j ∗ (x′) ∂tu(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2−1 2p k R+ 2k ⟨2k(t −s)⟩2 φ j ∗ (x′) ∂tu(t, ·, η) L p(Rn−1 x′ )ds L1t (R+) ≤C j∈Z 2sj k≥2 j 2−1 2p k 2k ⟨2kt⟩2 L1t (R+) m∈Z m ∗ (x′,η) φ j ∗ (x′) ∂tu(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2−1 2p k \|m−j\|≤1 m ∗ (x′,η) φ j ∗ (x′) ζ j(η) ∗ (η) ζ j−1(η) ∗ (η) ∂tu(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2−1 2p k∥ζ j(η)∥L p′ (R+,η) j ∗ (x′,η) ∂tu(t, ·, η) L p(Rn−1 x′ ) L p(R+η ) L1t (R+) ≤C j∈Z 2sj j ∗ (x′,η) ∂tu(t, ·, η) L p(Rn +,(x′,η)) L1t (R+) ≤C∥∂tu∥L1(R+; ˙Bs p,1(Rn +). 7. Optimal boundary trace estimates In this section, we show the optimality for the boundary trace estimate shown in Theorem 2.1 and 2.3. This shows that the condition on the boundary data in those theorems are not only the sufﬁcient condition but also the necessary condition (see for more detailed estimates for the boundary trace [24,32]) Theorem 7.1 (The Dirichlet boundary trace). Let 1 ≤p ≤∞and s ∈R. There exists a constant C > 0 such that for all function u = u(t, x′, η) ∈˙W 1,1(R+; ˙Bs p,1(Rn +)), u(t, x′, η) ∈L1(R+; ˙Bs p,1(Rn +)) with satisfying u(0, x′, η) = 0 almost everywhere in (x′, η) ∈Rn +, the following estimate holds: sup η∈R+ ∥u(·, ·, η)∥˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥u(·, ·, η)∥L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) ≤C ∥∂tu∥L1(R+; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(R+; ˙Bs p,1(Rn +)) . (7.1) (7.1) From Theorem 7.1, the following corollary immediately follows: From Theorem 7.1, the following corollary immediately follows: Corollary 7.2 (Sharp boundary trace for Dirichlet data). For 1 ≤p ≤∞, there exists a constant C > 0 such that the solution u to the initial-boundary value problem (2.5) with u ∈˙W 1,1 R+; ˙Bs p,1(Rn +) , u ∈L1 R+; ˙Bs p,1(Rn +) gives the following estimate on the boundary condition h: gives the following estimate on the boundary condition h: ∥h∥˙F1−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥h∥L1(R+; ˙Bs+2−1/p p,1 (Rn−1)) ≤ ⎧ ⎨ ⎩ C∥∇2u∥L1(R+; ˙Bs p,1(Rn +)), C∥∂tu∥L1(R+; ˙Bs p,1(Rn +)). Proof of Theorem 7.1. For 1 ≤p ≤∞, we assume that u ∈˙W 1,1(R+; ˙Bs p,1(Rn +)), u ∈L1(R+; ˙Bs p,1(Rn +)),. Proof of Theorem 7.1. For 1 ≤p ≤∞, we assume that u ∈˙W 1,1(R+; ˙Bs p,1(Rn +)), u ∈L1(R+; ˙Bs p,1(Rn +)),. T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 60 of 67 T. Ogawa and S. Shimizu T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. u(·, ·, η) ˙F1−1/2p 1,1 R+; ˙Bs p,1(Rn−1 x′ ) ≤ j∈Z k≥2 j 2(1−1/2p)k2sjψk ∗ (t) φ j ∗ (x′) u(t, ·, η) L p(Rn−1 x′ ) L1t (R+) + j∈Z k≤2 j 2(1−1/2p)k2sjψk ∗ (t) φ j ∗ (x′) u(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≡I + I I. J. Evol. Equ. Let u ∈L1(R+; ˙Bs p,1(Rn +)) and η ∈I−ℓ= [2−ℓ, 2−ℓ+1) with ℓ∈Z, we insert the unity m∈Z m(x′, η) ∗ (x′,η) to the estimate to see j∈Z One can apply the similar treatment for the spatial direction, for 1 ≤p ≤∞. Let u ∈L1(R+; ˙Bs p,1(Rn +)) and η ∈I−ℓ= [2−ℓ, 2−ℓ+1) with ℓ∈Z, we insert the unity Z m(x′, η) ∗ (x′,η) to the estimate to see m∈Z u(·, ·, η) L1(R+; ˙Bs+2−1/p p,1 (Rn−1 x′ )) p,1 x ≤C j∈Z 2sj2(2−1/p) jφ j ∗ (x′) ζ j−1(η) ∗ (η) \|m−j\|≤1 m(x′, η) ∗ (x′,η) u(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+2) j 2−jφ j ∗ (x′) ζ j−1(η) ∗ (η) ζ j(η) ∗ (η) j(x′, η) ∗ (x′,η) u(t, ·, η) p L p(Rn−1 x′ ) 1/p L1t (R+) ≤C j∈Z 2(s+2) j 2−jζ j L p′ (R+,η) × φ j ∗ (x′) ζ j−1(η) ∗ (η) j(x′, η) ∗ (x′,η) u(t, ·, η) L p(Rn−1 x′ ) L p(R+,η) p1/p L1t (R+) ≤C j∈Z 2(s+2) j 2−j2 jφ j ∗ (x′) j(x′, η) ∗ (x′,η) u(t, ·, η) p L p(Rn +) 1/p L1t (R+) ≤C j∈Z 2(s+2) j j(x′, η) ∗ (x′,η) u(s, ·, η) L p(Rn−1 x′ ×R+,η) L1t (R+) ≤C j∈Z 2sj j(·, ·) ∗ (x′,η) u(s, ·, ·) L p(Rn−1 x′ ×R+,η) L1t (R+) ≤C R+ ∥ u(t, x)∥˙Bs p,1dt. 7. Optimal boundary trace estimates (7.5) (7.5) Maximal L1-regularity for parabolic Page 61 of 67 30 Maximal L1-regularity for parabolic Page 61 of 67 30 J. Evol. Equ. Maximal L1-regularity for parabolic Page 61 of 67 30 J. Evol. Equ. On the other hand, the second term of the right hand side of (7.2) can be treated by using the Minkowski inequality that I I = j∈Z k≤2 j 2(1−1/2p)k2sjψk ∗ (t) φ j ∗ (x′) u(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≤2 j 2(1−1/2p)k\|ψk\| ∗ (t) φ j ∗ (x′) u(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≤2 j 2(1−1/2p)k m∈Z m ∗ (x′,η) ζ j(η) ∗ (η) ζ j−1(η) ∗ (η) φ j ∗ (x′) u(t, ·, η) L p(Rn−1 x′ ) L1t (R ≤C j∈Z 2sj2(2−1/p) j \|m−j\|≤1 m ∗ (x′,η) ζ j(η) ∗ (η) φ j ∗ (x′) u(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+2−1/p) j∥ζ j∥L p′ (R+,η) j ∗ (x′,η) u(t, ·, η) L p(Rn +) L1t (R+) ≤C j∈Z 2(s+2) j(− )−1 j ∗ (x′,η) u(t, ·, η) L p(Rn +) L1t (R+) ≤C 2sj j ∗ u(t · η) = C u ˙ (7 ≤C j∈Z 2sj j ∗ (x′,η) u(t, ·, η) L p(Rn +) L1t (R+) = C u L1(R+; ˙Bs p,1(Rn +)). (7.6) ≤C j∈Z 2sj j ∗ (x′,η) u(t, ·, η) L p(Rn +) L1t (R+) = C u L1(R+; ˙Bs p,1(Rn +)). (7.6) ≤C j∈Z 2sj j ∗ (x′,η) u(t, ·, η) L p(Rn +) L1t (R+) = C u L1(R+; ˙Bs p,1(Rn +)). (7.6) (7.6) j∈Z One can apply the similar treatment for the spatial direction, for 1 ≤p ≤∞. Let u ∈L1(R+; ˙Bs p,1(Rn +)) and η ∈I−ℓ= [2−ℓ, 2−ℓ+1) with ℓ∈Z, we insert the unity m∈Z m(x′, η) ∗ (x′,η) to the estimate to see j∈Z One can apply the similar treatment for the spatial direction, for 1 ≤p ≤∞. Let u ∈L1(R+; ˙Bs p,1(Rn +)) and η ∈I−ℓ= [2−ℓ, 2−ℓ+1) with ℓ∈Z, we insert the j∈Z One can apply the similar treatment for the spatial direction, for 1 ≤p ≤∞. J. Evol. Equ. (7.7) p, x ≤C j∈Z 2sj2(2−1/p) jφ j ∗ (x′) ζ j−1(η) ∗ (η) \|m−j\|≤1 m(x′, η) ∗ (x′,η) u(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+2) j 2−jφ j ∗ (x′) ζ j−1(η) ∗ (η) ζ j(η) ∗ (η) j(x′, η) ∗ (x′,η) u(t, ·, η) p L p(Rn−1 x′ ) 1/p L1t (R+) ≤C j∈Z 2(s+2) j 2−jζ j L p′ (R+,η) ≤C j∈Z 2(s+2) j 2−jζ j L p′ (R+,η) × φ j ∗ (x′) ζ j−1(η) ∗ (η) j(x′, η) ∗ (x′,η) u(t, ·, η) L p(Rn−1 x′ ) L p(R+,η) p1/p L1t (R+) ≤C j∈Z 2(s+2) j 2−j2 jφ j ∗ (x′) j(x′, η) ∗ (x′,η) u(t, ·, η) p L p(Rn +) 1/p L1t (R+) ≤C j∈Z 2(s+2) j j(x′, η) ∗ (x′,η) u(s, ·, η) L p(Rn−1 x′ ×R+,η) L1t (R+) ≤C j∈Z 2sj j(·, ·) ∗ (x′,η) u(s, ·, ·) L p(Rn−1 x′ ×R+,η) L1t (R+) ≤C R+ ∥ u(t, x)∥˙Bs p,1 (7.7) Combining the estimates (7.2), (7.5), (7.6) and (7.7), we conclude the estimate (7.1 □ Combining the estimates (7.2), (7.5), (7.6) and (7.7), we conclude the estimate (7.1). □ Combining the estimates (7.2), (7.5), (7.6) and (7.7), we conclude the estimate (7.1). □ Theorem 7.3 (Sharp boundary derivative trace). For 1 ≤p ≤∞and s ∈R, there ex- istsaconstantC > 0suchthatforallfunctionu = u(t, x′, η) ∈˙W 1,1(R+; ˙Bs p,1(Rn +)), Theorem 7.3 (Sharp boundary derivative trace). For 1 ≤p ≤∞and s ∈R, there ex- istsaconstantC > 0suchthatforallfunctionu = u(t, x′, η) ∈˙W 1,1(R+; ˙Bs p,1(Rn +)), T. Ogawa and S. Shimizu J. Evol. Equ. u ∈L1(R+; ˙Bs p,1(Rn +)) with ∂ηu(0, x′, η) = 0, it holds sup η∈R+ ∂ηu(·, ·, η) ˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∂ηu(·, ·, η) L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) sup η∈R+ ∂ηu(·, ·, η) ˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∂ηu(·, ·, η) L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) ≤C ∥∂tu∥L1(R+; ˙Bs p,1(Rn +)) + ∥∇2u∥L1(R+; ˙Bs p,1(Rn +)) . (7.8) (7.8) Similar to the boundary trace estimate, Theorem 7.3 immediately implies the opti- mality of the boundary derivative trace estimate for the Neumann boundary condition. Corollary 7.4 (Sharp boundary trace for Neumann data). J. Evol. Equ. (7.12) (7.12) For the second term of (7.9), we use the Minkowski inequality to see For the second term of (7.9), we use the Minkowski inequality to see I I ≤C j∈Z 2sj k≤2 j 2(1/2−1/2p)k\|ψk\| ∗ (t) φ j ∗ (x′) ∂ηu(t, x′, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≤2 j 2(1/2−1/2p)k × m∈Z m ∗ (x′,η) ζ j(η) ∗ (η) ζ j−1(η) ∗ (η) φ j ∗ (x′) ∂ηu(t, x′, η) L p(Rn−1 x′ ) L1t I I ≤C j∈Z 2sj k≤2 j 2(1/2−1/2p)k\|ψk\| ∗ (t) φ j ∗ (x′) ∂ηu(t, x′, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≤2 j 2(1/2−1/2p)k × m∈Z m ∗ (x′,η) ζ j(η) ∗ (η) ζ j−1(η) ∗ (η) φ j ∗ (x′) ∂ηu(t, x′, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+1−1/p) j \|m−j\|≤1 m ∗ (x′,η) ∂ηζ j(η) ∗ (η) ζ j−1(η) ∗ (η) φ j ∗ (x′) u(t, x′, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+1−1/p) j∥∂η ζ j∥L p′ (R+,η) j ∗ (x′,η) u(t, x′, η) L p(Rn +) L1t (R+) ≤C j∈Z 2(s+2) j(− )−1 j ∗ (x′,η) u(t, x′, η) L p(Rn +) L1t (R+) ≤C u L1(R+; ˙Bs p,1(Rn +)). (7.13) ≤C j∈Z 2(s+1−1/p) j∥∂η ζ j∥L p′ (R+,η) j ∗ (x′,η) u(t, x′, η) L p(Rn +) L1t (R+) ≤C j∈Z 2(s+2) j(− )−1 j ∗ (x′,η) u(t, x′, η) L p(Rn +) L1t (R+) ≤C u L1(R+; ˙Bs p,1(Rn +)). (7.13) The estimate for the spatial direction is slightly simpler. For 1 ≤p ≤∞, u ∈ L1(R+; ˙Bs p,1(Rn +)) and noting (4.8), we estimate very similar way as in (7.7) to obtain The estimate for the spatial direction is slightly simpler. J. Evol. Equ. ≤C j∈Z 2sj k≥2 j 2−( 1 2 + 1 2p )k × \|m−j\|≤1 m ∗ (x′,η) φ j ∗ (x′) ∂ηζ j(η) ∗ (η) ζ j−1(η) ∗ (η) ∂tu(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2−( 1 2 + 1 2p )k∥∂ηζ j(η)∥L p′ (R+,η) × j ∗ (x′,η) φ j ∗ (x′) ζ j−1(η) ∗ (η) ∂tu(t, x, η) L p(Rn−1 x′ ) L p(R+η ) L1t (R+) ≤C j∈Z 2sj j ∗ (x′,η) ∂tu(t, x, η) L p(Rn +,(x′,η)) L1t (R+) ≤C∥∂tu∥L1(R+; ˙Bs p,1(Rn +). (7.12) ≤C j∈Z 2sj k≥2 j 2−( 1 2 + 1 2p )k × \|m−j\|≤1 m ∗ (x′,η) φ j ∗ (x′) ∂ηζ j(η) ∗ (η) ζ j−1(η) ∗ (η) ∂tu(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2−( 1 2 + 1 2p )k × \|m−j\|≤1 m ∗ (x′,η) φ j ∗ (x′) ∂ηζ j(η) ∗ (η) ζ j−1(η) ∗ (η) ∂tu(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2−( 1 2 + 1 2p )k∥∂ηζ j(η)∥L p′ (R+,η) × j ∗ (x′,η) φ j ∗ (x′) ζ j−1(η) ∗ (η) ∂tu(t, x, η) L p(Rn−1 x′ ) L p(R+η ) L1t (R+) ≤C j∈Z 2sj j ∗ (x′,η) ∂tu(t, x, η) L p(Rn +,(x′,η)) L1t (R+) ≤C∥∂tu∥L1(R+; ˙Bs p,1(Rn +). (7.12) × j ∗ (x′,η) φ j ∗ (x′) ζ j−1(η) ∗ (η) ∂tu(t, x, η) L p(Rn−1 x′ ) L p(R+η ) L1t (R+) ≤C j∈Z 2sj j ∗ (x′,η) ∂tu(t, x, η) L p(Rn +,(x′,η)) L1t (R+) ≤C∥∂tu∥L1(R+; ˙Bs p,1(Rn +). J. Evol. Equ. For 1 ≤p ≤∞and s ∈R, let u be a solution to the Cauchy problem (2.8) with ˙W 1,1(R+; ˙Bs p,1(Rn +)) ∩L1(R+; ˙Bs+2 p,1 (Rn +)), then there exists a constant C > 0 such that the boundary condition has to satisfy then there exists a constant C > 0 such that the boundary condition has to satisfy ∥h∥˙F1/2−1/2p 1,1 (R+; ˙Bs p,1(Rn−1)) + ∥h∥L1(R+; ˙Bs+1−1/p p,1 (Rn−1)) ≤ ⎧ ⎨ ⎩ C∥∇2u∥L1(R+; ˙Bs p,1(Rn +)), C∥∂tu∥L1(R+; ˙Bs p,1(Rn +)). Proof of Theorem 7.3. The proof is very similar to the proof for Theorem 7.1. For 1 ≤ p ≤∞and s ∈R, assume u ∈˙W 1,1(R+; ˙Bs p,1(Rn +)), u ∈L1(R+; ˙Bs p,1(Rn +)). Then Proof of Theorem 7.3. The proof is very similar to the proof for Theorem 7.1. For 1 ≤ p ≤∞and s ∈R, assume u ∈˙W 1,1(R+; ˙Bs p,1(Rn +)), u ∈L1(R+; ˙Bs p,1(Rn +)). Then ∂ηu(·, ·, η) ˙F1/2−1/2p 1,1 R+; ˙Bs p,1(Rn−1 x′ ) ≤ j∈Z 2sj k≥2 j 2(1/2−1/2p)kψk ∗ (t) φ j ∗ (x′) ∂ηu(t, ·, η) L p(Rn−1 x′ ) L1t (R+) + j∈Z 2sj k≤2 j 2(1/2−1/2p)kψk ∗ (t) φ j ∗ (x′) ∂ηu(t, ·, η) L p(Rn−1 x′ ) L1t (R+) ≡I + I I. (7.9) (7.9) For all j ∈Z, ∂ηu(0, x′, η) = 0, (x′, η) ∈Rn−1 × R+, (7.10) and it follows from (7.4) that and it follows from (7.4) that ψk(t) ∗ (t) ∂ηu(t, x′, η) =∂−1 t ψk(t) ∗ (t) ∂η∂tu(t, x′, η). (7.11) Hence from (7.10) and (7.11) and using the Hausdorff–Young inequality I = j∈Z 2sj k≥2 j 2(1/2−1/2p)k∂−1 t ψk ∗ (t) φ j ∗ (x′) ∂η∂tu(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2sj k≥2 j 2−( 1 2 + 1 2p )k R+ \|(∂−1 t ψ)k(t −s)\| φ j ∗ (x′) ∂η∂tu(s, x, η) L p(Rn−1 x′ )ds L1t (R+) Maximal L1-regularity for parabolic Page 63 of 67 30 Maximal L1-regularity for parabolic Page 63 of 67 30 Maximal L1-regularity for parabolic Page 63 of 67 30 Page 63 of 67 30 J. Evol. Equ. J. Evol. Equ. For 1 ≤p ≤∞, u ∈ L1(R+; ˙Bs p,1(Rn +)) and noting (4.8), we estimate very similar way as in (7.7) to obtain ∂ηu(·, ·, η) L1 R+; ˙Bs+1−1/p p,1 (Rn−1 x′ ) ≤C j∈Z 2(s+1−1/p) jφ j ∗ (x′) m∈Z m(x′, η) ∗ (x′,η) ∂ηu(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+1−1/p) jφ j ∗ (x′) ∂ηζ j(η) ∗ (η) ζ j−1(η) ∗ (η) \|m−j\|≤1 m(x′, η) ∗ (x′,η) u(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+1) j ∥∂ηζ j∥p L p′ (R+,η) × φ j ∗ (x′) ζ j−1(η) ∗ (η) j(x′, η) ∗ (x′,η) u(t, x, η) p L p(Rn x′ ×R+,η)2−j1/p L1t (R+) ≤C j∈Z 2(s+2) j j(x′, η) ∗ (x′,η) u(s, ·, η) L p(Rn−1 x′ ×R+,η) L1t (R+) ≤C R+ ∥ u(t)∥˙Bs p,1(Rn +)dt. (7.14) ∂ηu(·, ·, η) L1 R+; ˙Bs+1−1/p p,1 (Rn−1 x′ ) ≤C j∈Z 2(s+1−1/p) jφ j ∗ (x′) m∈Z m(x′, η) ∗ (x′,η) ∂ηu(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C j∈Z 2(s+1−1/p) jφ j ∗ (x′) ∂ηζ j(η) ∗ (η) ζ j−1(η) ∗ (η) \|m−j\|≤1 m(x′, η) ∗ (x′,η) u(t, x, η) L p(Rn−1 x′ ) L1t (R+) ≤C 2(s+1) j ∥∂ζ j∥p ≤C R+ ∥ u(t)∥˙Bs p,1(Rn +)dt. (7.14) (7.14) Combining the estimates (7.9), (7.12), (7.13) and (7.14), we obtain the result (7.8). This completes the proof. □ Combining the estimates (7.9), (7.12), (7.13) and (7.14), we obtain the result (7.8). This completes the proof. □ T. Ogawa and S. Shimizu J. Evol. Equ. 30 Page 64 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. J. Evol. Equ. 8. Concluding remarks Since we establish maximal L1-regularity for the inhomogeneous boundary data of Dirichlet and Neumann boundary conditions, the estimate can be generalized into other boundary conditions. For instance, one can generalize to the case of the oblique boundary condition: ⎧ ⎪⎨ ⎪⎩ ∂tu − u = f (t, x), t > 0, x ∈Rn +, b · ∇u\|xn=0 = h(t, x′), t > 0, x′ ∈Rn−1, u\|t=0 = u0, x ∈Rn +, (8.1) (8.1) where b = (b′, bn) is a given constant vector in Rn with bn ̸= 0. We also obtain that the integral kernel to the Laplacian of the solution ob is given by where b = (b′, bn) is a given constant vector in Rn with bn ̸= 0. We also obtain that the integral kernel to the Laplacian of the solution ob is given by ob(t, x) = cn+1 R Rn−1 eiτt+ix′·ξ′ iτ ib′ · ξ′ −bn iτ + \|ξ′\|2 e−√ iτ+\|ξ′\|2xndξ′dτ. (8.2) (8.2) Then, the analogous estimate to the boundary potential in Theorem 2.4 can be obtained once we establish the almost orthogonal estimate as in Lemma 6.5. The detailed estimate is shown in the forthcoming paper [37]. Then, the analogous estimate to the boundary potential in Theorem 2.4 can be obtained once we establish the almost orthogonal estimate as in Lemma 6.5. The detailed estimate is shown in the forthcoming paper [37]. For a general domain ⊂Rn, for instance bounded domain with smooth boundary, we may generalize our results. In such a case, by standard decompositions of unity near the boundary ∂, it may be reduced into a problem in the half-space by a smooth diffeomorphism. Then if we establish maximal L1-regularity for the parabolic initial- boundary value problems with lower order spatial derivative terms, we may extend the estimate for general domain cases. Further application to the nonlinear problem is also available. We discuss such an application to the ﬂuid mechanics in a forthcoming paper [36] (cf. [35]). Acknowledgements The authors are grateful to the anonymous referee for valuable suggestions and com- ments that improve the presentation of this paper largely. The ﬁrst author is par- tially supported by JSPS grant-in-aid for Scientiﬁc Research (S) #19H05597, Scien- tiﬁc Research (B) #18H01131 and Challenging Research (Pioneering) #20K20284. The second author is partially supported by JSPS grant-in-aid for Scientiﬁc Research (B) #16H03945 and (B) #21H00992 and Fostering Joint International Research (B) #18KK0072. REFERENCES [1] Abidi, H., Paicu, M. Existence globale pour un ﬂuide inhomogène. Ann. Inst. Fourier (Grenoble) 57 (2007) 883–917. [2] Amann, H., Linear and Quasilinear Parabolic Problems. Vol I Abstract Linear Theory, Monographs in Math. Vol 89, Birkhäuser Verlag, Basel·Boston·Berlin, 1995. [2] Amann, H., Linear and Quasilinear Parabolic Problems. Vol I Abst in Math. Vol 89, Birkhäuser Verlag, Basel·Boston·Berlin, 1995. [3] Amann, H., Linear and Quasilinear Parabolic Problems. Vol II: Math. Vol 106, Birkhäuser Verlag, Basel·Boston·Berlin, 2019. [3] Amann, H., Linear and Quasilinear Parabolic Problems. Vol II: Function Spaces, Monographs in Math. Vol 106, Birkhäuser Verlag, Basel·Boston·Berlin, 2019. [4] Bahouri, H., Chemin, J-Y., Danchin, R., Fourier Analysis and Nonlinear Partial Differential Equa- tions. Grundlehren der mathematische Wisssenshaften 343, Springer-Verlag, Berlin-Heidelberg- Dordrecht-London-New York 2011. [5] Benedek, A., Calderón, A.P., Panzone, R., Convolution operators on Banach space valued functions. Proc. Nat. Acad. Sci. USA 48 (1962) 356–365. [6] Bergh, J., Löfström, J., Interpolation Spaces; an introduction, Springer-Verlag, Berlin, 1976. [7] Chemin, J.-Y., Théorèmes d’unicité pour le systéme de Navier-Stokes tridimensionnel, J. Ana Math., 77 (1999), 27–50. [8] Clément, Ph., Prüss, J., Global existence for a semilinear parabolic Volterra equation, Math. Z., 209 (1992) 17–26. [9] Coulhon, T., Lamberton, D., Régularité L p pour les équations dévolution. In: Séminaire d’analyse fonctionelle 1984-85, Publications mathématiques de l’Université Paris VII, 26 (1987) 141–153. [9] Coulhon, T., Lamberton, D., Régularité L p pour les équations dévolution. In: Séminaire d’analyse f i ll 1984 85 P bli i hé i d l’U i i é P i VII 26 (1987) 141 153 [10] Danchin, R., Density-dependent incompressible viscous ﬂuids in critical spaces, Proc. Roy Soc. Edinburgh 133A (2003), 1311–1334. Edinburgh 133A (2003), 1311–1334. [11] Danchin, R., Well-posedness in critical spaces for barotropic viscous ﬂuids with truly not constant density, Comm. Partial Differential Equations, 32 (2007) 1373–1397. [12] Danchin, R., Mucha, P. B., A critical functional framework for the inhomogeneous Navier-Stokes equations in the half-space, J. Funct. Anal., 256 (2009) 881–927. [13] Da Prato, G., Grisvard, P., Sommes d’opérateurs linéaires et équations différentielles opérationelles, J. Math. Pure Appl. 54 (1975) 305–387. [14] Denk, R., Hieber, M., Prüss, J., R-boundedness, Fourier multipliers and problems of elliptic and parabolic type, Memoirs of AMS, 166, No. 788 (2003). [15] Denk, R., Hieber, M., Prüss, J., Optimal L p-Lq-regularity for parabolic problems with inhomog neous boundary data, Math. Z., 257 (2007) 193–224. Declarations Conﬂict of interest The authors declare that there is no conﬂict of interest. Page 65 of 67 30 Page 65 of 67 30 J. Evol. Equ. J. Evol. Equ. J. Evol. Equ. Maximal L1-regularity for parabolic Open Access. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/ by/4.0/. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. REFERENCES [16] De Simon, L., Un’applicazione della teoria degli integrali allo studio delle equazioni differenzia astratta del primo ordine, Rend. Sem. Mat. Univ. Padova, 34 (1964) 157–162. [17] Dore, G., L p regularity for abstract differential equations. In: Functional Analysis and Relate Topics, H.Komatsu (ed.), Lecture Notes in Math., 1540, Springer (1993). [18] Dore, G., Venni, A., On the closedness of the sum of two closed operators, Math. Z., 196 (1987) 189–201. [19] Duong, X.T., H∞functional calculus of second order elliptic partial differential operators on L p spaces, In: Miniconference on Operators in Analysis, 1989. Proc. Centre Math. Anal. ANU, Canberra, 24 (1990) 91–102. 30 Page 66 of 67 T. Ogawa and S. Shimizu J. Evol. Equ. 67 [20] Giga, Y., Saal, J., L1 maximal regularity for the Laplacian and applications, Discrete Conti. Dyn. Syst. I (2011) 495–504. [21] Giga, Y., Sohr, H, Abstract L p estimates for the Cauchy problem with applications to the Navie Stokes equations in exterior domains, J. Funct. Anal., 102 (1991) 72–94. [22] Hieber, M., Prüss, J., Heat kernels and maximal L p-Lq estimates for parabolic evolution equation Comm. P.D.E., 22 (1997), 1674-1669. [23] Iwabuchi, T., Global solutions for the critical Burgers equation in the Besov spaces and the large time behavior, Ann. I. H. Poincaré, (2015), 687–713. [24] Johnsen, J., Sickel, W., On the trace problem for Lizorkin-Triebel spaces with mixed norms, Math. Nachr. 281 (2008), 669–696. [25] Kalton, N., Lancien, G., A solution to the problem of the L p-maximal regularity, Math. Z., 235 (2000) 559–568. [26] Kalton, N., Weis, L., The H∞-calculus and sums of closed operators, Math. Ann., 321 (2001) 319–345. [27] Krylov, N.V., Parabolic and elliptic equations with V MO coefﬁcients, Comm. Partial Differential Equations, 32 (2007), 453–475. [28] Kunstmann, P.C., Weis, L., Maximal L p-regularity for parabolic equations, Fourier multiplier theorems and H∞-functional calculus, M. Iannelli, R. Nagel and S. Piazzera (ed.) Functional Analytic Methods for Evolution Equations, Lecture Notes in Mathematics 1899, Springer-Verlag Berlin Heidelberg, 2004. [29] Ladyzhenskaya, O.A., Solonnikov, V.A., Ural’tseva, N.N., Linear and quasilinear equations o parabolic type, Amer. Math. Soc. Transl. Math. Monographs, Providence, R.I., 1968. [30] Lizorkin, P. I., Properties of functions of class r p,θ, Trudy Mat. Inst. Steklov, 131 (1974), 158–18 [30] Lizorkin, P. I., Properties of functions of class r p,θ, Trudy Mat. Inst. Steklov, 131 (1974), 158–181. REFERENCES [31] McIntosh, A., Yagi, A., Operators of type ω without a bounded H∞-functional calculus, In: Mini- conference on Operators in Analysis, 1989. Proc. Centre Math. Anal. ANU, Canberra, 24 (1990) 159–172. [32] Meyries, M., Veraar, M. C., Traces and embeddings of anisotropic function spaces, Math. Ann. 360 (2014), 571–606. [33] Ogawa, T., Shimizu, S., End-point maximal regularity and its application to two-dimensional Keller- Segel system, Math. Z., 264 (2010) 601–628. 1 [34] Ogawa,T., Shimizu, S., End-point maximal L1-regularity for a Cauchy problem to parabolic equa- tions with variable coefﬁcient, Math. Ann., 365 (2016) 661–705. [35] Ogawa,T., Shimizu, S., Maximal L1-regularity of the heat equation and application to a free bound- ary problem of the Navier–Stokes near the half-space, J. Elliptic Parabol. Equ., 7 (2021) 509–535. [36] Ogawa,T., Shimizu, S., Maximal L1 regularity and free boundary value problems for the incom- pressible Navier-Stokes equations in critical spaces, preprint (2021). [37] Ogawa,T., Shimizu, S., End-point maximal L1-regularity for a initial boundary value problem of the heat equations under the oblique boundary condition, in preparation. [38] Peetre, J., On spaces of Triebel-Lizorkin type, Ark. Mat. 13 (1975) 123–130. spaces of Triebel-Lizorkin type, Ark. Mat. 13 (1975) [39] Peetre, J., New thoughts on Besov spaces, Duke University Mathematics Series, No.1, Duke Uni- versity, Durham, N., C., 50 1976. [40] Prüss,J.,Simonett,G.,MovingInterfacesandQuasi-linearParabolicDifferentialEquations,Mono- graphs in Math. 105, Birkhäuser, Basel (2016). [41] Shibata, Y., Shimizu, S., On the free boundary problem for the Navier-Stokes equations, Differential Integral Equations 20 no. 3 (2007), 241–276. [42] Shibata, Y., Shimizu, S., On the L p-Lq maximal regularity of the Neumann problem for the Stokes equations in a bounded domain, J. reine angew. Math. 615 (2008), 157–209. [43] Sobolevskii, P.E., Fractional powers of coercively positive sums of operators, Dokl. Akad Izv.11 (1977) 1323–1358. [44] Solonnikov, V. A., A priori estimates for a solutions of second-order equations of parabolic type, Trudy Mat. Inst. Stekelov., 70 (1964), 133-212, (Enlish traslation; Transl Ameri. Math. Soc. 65 (1967), 51–137.) [45] Triebel, H., Spaces of distributions of Besov type in Euclidean n-space, Duality, interpolation, Ark. Mat. 11 (1973), 13–64. Maximal L1-regularity for parabolic J. Evol. Equ. Page 67 of 67 30 Page 67 of 67 30 [46] Triebel, H., Interpolation Theory, Function spaces, Differential Operators, North-Holland, Am terdam - New York - Oxford, 1978. [47] Triebel, H., Theory of Function Spaces, Birkhäuser, Basel, 1983. y f [48] Weidemaier, P., Reﬁnement of an L p-estimate of Solonnikov for a parabolic equation of the second order with conormal boundary condition, Math. Z., 199 (1988) 589–604. [49] Weidemaier, P., On the trace theory for functions in Sobolev spaces with mixed L p-norm, Czech. Math. J. 44 (1994), 7–20. [50] Weidemaier, P., Maximal regularity for parabolic equations with inhomogeneous boundary condi- tions in Sobolev spaces with mixed L p-norm, Electron. Res. Announc. Amer. Math. Soc. 8 (2002) 47–51. [51] Weidemaier, P., Vector-valued Lizorkin-Triebel spaces and sharp trace theory for functions in Sobolev spaces with mixed L p-norm for parabolic problem, Sbornik: Math. 196 (2005), 777–790. [52] Weidemaier, P., On L p-estimate of optimal type for the parabolic oblique derivative problem with VMO-coefﬁcients– A reﬁned version, Progress in Nonlinear Differential Equations and Their Ap- plications vol. 64, 529–536, Birkhäuser Verlag Basel, 2005. [53] Weis, L., Operator-valued Fourier multiplier theorems and maximal L p-regularity, Math. Ann., 319 (2001) 735–758. Takayoshi Ogawa Mathematical Institute/Research Alliance Center of Mathematical Tohoku University Sendai 980-8578 Japan E-mail: takayoshi.ogawa.c8@tohoku.ac.jp Senjo Shimizu Graduate School of Human and Environmental Studies Kyoto University Kyoto 606-8501 Japan E-mail: shimizu.senjo.5s@kyoto-u.ac.jp Accepted: 19 October 2021 Senjo Shimizu Graduate School of Human and Environmental Kyoto University Kyoto 606-8501 Japan E-mail: shimizu.senjo.5s@kyoto-u.ac.jp Accepted: 19 October 2021
https://openalex.org/W2771971360	https://link.springer.com/content/pdf/10.1007/s11432-017-9297-1.pdf	English	null	Almost sure stabilization of hybrid systems by feedback control based on discrete-time observations of mode and state	Science China. Information sciences/Science China. Information Sciences	2,018	cc-by	11,217	SCIENCE CHINA Information Sciences SCIENCE CHINA Information Sciences . RESEARCH PAPER . July 2018, Vol. 61 070213:1–070213:16 https://doi.org/10.1007/s11432-017-9297-1 Special Focus on Modeling, Analysis and Control of Stochastic Systems Almost sure stabilization of hybrid systems by feedback control based on discrete-time observations of mode and state Gongfei SONG1, Zhenyu LU2, Bo-Chao ZHENG1 & Xuerong MAO3* 1Jiangsu Collaborative Innovation Center on Atmospheric Environment and Equipment Technology, School of Information and Control, Nanjing University of Information Science and Technology, Nanjing 210044, China; j g , ; ool of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, China; 3 j g , ; 2School of Electronic and Information Engineering, Nanjing University of Information Science and Techn Nanjing 210044, China; 3Department of Mathematics and Statistics, University of Strathclyde, Glasgow G1 1XH, UK Received 22 August 2017/Accepted 27 October 2017/Published online 13 June 2018 Abstract Although the mean square stabilization of hybrid systems by feedback control based on discrete- time observations of state and mode has been studied by several authors since 2013, the corresponding almost sure stabilization problem has received little attention. Recently, Mao was the ﬁrst to study the almost sure stabilization of a given unstable system ˙x(t) = f(x(t)) by a linear discrete-time stochastic feedback control Ax([t/τ]τ)dB(t) (namely the stochastically controlled system has the form dx(t) = f(x(t))dt + Ax([t/τ]τ)dB(t)), where B(t) is a scalar Brownian, τ > 0, and [t/τ] is the integer part of t/τ. In this paper, we consider a much more general problem. That is, we study the almost sure stabilization of a given unstable hybrid system ˙x(t) = f(x(t), r(t)) by nonlinear discrete-time stochastic feedback control u(x([t/τ]τ), r([t/τ]τ))dB(t) (so the stochastically controlled system is a hybrid stochastic system of the form dx(t) = f(x(t), r(t))dt + u(x([t/τ]τ), r([t/τ]τ))dB(t)), where B(t) is a multi-dimensional Brownian motion and r(t) is a Markov chain. Keywords Brownian motion, Markov chain, generalized Itˆo formula, almost sure exponential stability, stochastic feedback control Citation Song G F, Lu Z Y, Zheng B-C, et al. Almost sure stabilization of hybrid systems by feedback control based on discrete-time observations of mode and state. Sci China Inf Sci, 2018, 61(7): 070213, https://doi.org/10. 1007/s11432-017-9297-1 * Corresponding author (email: x.mao@strath.ac.uk) c⃝The Author(s) 2018. This article is published with open access at link.springer.com info.scichina.com link.springer.com Citation Song G F, Lu Z Y, Zheng B-C, et al. Almost sure stabilization of hybrid systems by feedback control based on discrete-time observations of mode and state. Sci China Inf Sci, 2018, 61(7): 070213, https://doi.org/10. 1007/s11432-017-9297-1 Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:2 Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:2 As is well known, a given unstable system can be stabilized by noise or noise can be used to make a system more stable when it is already stable. Arnold et al. [21] pointed out that a linear system can be stabilized by zero mean stationary parameter noise. In [22], a linear hybrid stochastic system was stabilized by Gaussian type noise. In addition, Khasminskii [23] proposed that a system was stabilized by using two types of white noise. It was shown in [24] that an unstable nonlinear system can be stabilized by Brownian motion provided the growth condition is linear. Mao [25] showed that any nonlinear system ˙x(t) = f(x(t), t) whose coeﬃcient satisﬁed the condition \|f(x, t)\| ⩽K\|x\|, K > 0, it was possible to use the Brownian motions to stabilize the system. It is worth noting that Appleby et al. [26] presented a general theory on the problem of stochastic stabilization for a nonlinear functional diﬀerential equation by noise. Mao et al. [27] developed an unstable Markovian jump system ˙x(t) = f(x(t), r(t), t) that can be stabilized by stochastic control and the partial subsystem was controlled. In other words, the space S of the Markov chain was divided into two proper subspaces S1 and S2, i.e., S = S1 ∪S2. In summary, Mao et al. [27] considered the controlled stochastic system (1) dx(t) = f(x(t), r(t), t)dt + u(r(t), t)dB(t), (1) dx(t) = f(x(t), r(t), t)dt + u(r(t), t)dB(t), where u(i, t) = 0 for i ∈S1 while u(i, t) = u(i, x(t)) was a feedback control for i ∈S2. New methods and suﬃcient conditions on the stochastic stabilization for Markovian jump systems were provided in [28]. With some applications, two examples on stabilization and destabilization by noise in the plane were presented in [29]. where u(i, t) = 0 for i ∈S1 while u(i, t) = u(i, x(t)) was a feedback control for i ∈S2. New methods and suﬃcient conditions on the stochastic stabilization for Markovian jump systems were provided in [28]. With some applications, two examples on stabilization and destabilization by noise in the plane were presented in [29]. We should of course point out that the corresponding problem based on discrete-time state observations has already been studied by some authors. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:2 Recently, Mao [30] was the ﬁrst to study this stabilization problem. He also obtained a bound τ ∗on τ for the controlled system to be stable as long as τ < τ∗ (plus some other conditions of course). Here τ > 0 is the duration between two consecutive observations. From the point of control cost, it is clearly better to have a larger τ ∗. Inﬂuenced by [30], a number of recent papers (e.g., [31, 32]) have signiﬁcantly improved the bound τ ∗. Mao et al. [31] established a better bound on τ ∗by considering a couple of important classes of hybrid stochastic systems and using their special features. On the other hand, a better bound on τ ∗was also obtained in [32] by making use of Lyapunov functionals. In particular, Song et al. [33] pointed out that the discrete-time feedback control in controlled hybrid stochastic systems was based on not only the discrete-time observations of the state, x(kτ) (k = 0, 1, 2, . . .) but also it was still dependent on the discrete-time observations of the mode, r(kτ), on k = 0, 1, 2, . . .. Observing that all the papers mentioned above were concerned with the mean square stabilization by the discrete-time feedback control in the drift part, Mao [34] discussed the following almost sure exponential stabilization by discrete-time feedback control in the diﬀusion part. Given an unstable nonlinear system ˙x(t) = f(x(t)), Mao designed a feedback control Ax([t/τ]τ), based on the discrete-time state observations, in the diﬀusion part so that the corresponding closed-loop system dx(t) = f(x(t))dt + Ax([t/τ]τ)dB(t) (2) dx(t) = f(x(t))dt + Ax([t/τ]τ)dB(t) (2) dx(t) = f(x(t))dt + Ax([t/τ]τ)dB(t) dx(t) = f(x(t))dt + Ax([t/τ]τ)dB(t) (2) s almost surely exponentially stable. Here B(t) was a scalar Brownian motion, f : Rn →Rn satisﬁed as almost surely exponentially stable. Here B(t) was a scalar Brownian motion, f : Rn →Rn sat \|f(x) −f(y)\| ⩽α\|x −y\|, ∀x, y ∈Rn, for some α > 0 and f(0) = 0, and A was an n × n real-valued matrix such that \|Ax\|2 ⩽ρ1\|x\|2 and \|xTAx\|2 ⩾ρ2\|x\|2, ∀x ∈Rn, for some positive numbers ρ1 and ρ2 satisfying ρ2 −0.5ρ1 > α. Mao [34] showed that there was a positive number τ ∗such that the controlled system (2) was almost surely exponentially stable provided that τ < τ∗. 1 Introduction In recent years, stochastic systems have been considered by many researchers since many practical systems can be modeled using these kinds of systems. Many signiﬁcant results for stochastic systems have been reported (see [1–13]). Markovian jump systems are a special class of hybrid stochastic systems, which can be found in some engineering systems including power systems, manufacturing systems, ecosystems, and so forth. The literature in this area is huge and lots of papers are open access, thus we only mention a few [14–18]. Shaikhet [19] provided the suﬃcient conditions of asymptotic mean square stability for Markovian systems with delay. Mao [20] discussed the problem of exponential stability of general nonlinear Markovian jump systems. info.scichina.com link.springer.com c⃝The Author(s) 2018. This article is published with open access at link.springer.com Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:2 To the best of the authors’ knowledge, the problem of almost sure exponential stabilization for hybrid stochastic systems has received little attention, in particular, in the framework of stochastic feedback control based on discrete-time observations of mode and state. for some positive numbers ρ1 and ρ2 satisfying ρ2 −0.5ρ1 > α. Mao [34] showed that there was a positive number τ ∗such that the controlled system (2) was almost surely exponentially stable provided that τ < τ∗. To the best of the authors’ knowledge, the problem of almost sure exponential stabilization for hybrid stochastic systems has received little attention, in particular, in the framework of stochastic feedback control based on discrete-time observations of mode and state. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:3 These motivate us to consider the following more general problem: if the given unstable system is expressed as a hybrid stochastic system ˙x(t) = f(x(t), r(t)), can we design a discrete-time feedback control u(x([t/τ]τ), r([t/τ]τ)), based on the discrete-time observations of both state and mode, in the diﬀusion part so that the following closed-loop system dx(t) = f(x(t), r(t))dt + u(x([t/τ]τ), r([t/τ]τ))dB(t) (3) (3) is almost surely exponentially stable? Here B(t) is an m-dimensional Brownian motion, r(t) is a Markov chain in a ﬁnite state space S while f : Rn × S →Rn and u : Rn × S →Rn×m. We highlight a number of key features. is almost surely exponentially stable? Here B(t) is an m-dimensional Brownian motion, r(t) is a Markov chain in a ﬁnite state space S while f : Rn × S →Rn and u : Rn × S →Rn×m. We highlight a number of key features. is almost surely exponentially stable? Here B(t) is an m-dimensional Brownian motion, r(t) is a Markov chain in a ﬁnite state space S while f : Rn × S →Rn and u : Rn × S →Rn×m. We highlight a number of key features. • The stabilization is in the sense of almost sure exponential stability and there is far less known about this than the mean square stabilization. • The stabilization is in the sense of almost sure exponential stability and there is far less known about this than the mean square stabilization. • The controlled system is a hybrid stochastic delay system. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:2 • The discrete-time feedback control u(x([t/τ]τ), r([t/τ]τ)) is based on the discrete-time observations of both state and mode. • The discrete-time feedback control u(x([t/τ]τ), r([t/τ]τ)) is based on the discrete-time observations of both state and mode. 2 Preliminaries and notation Consider the following unstable hybrid stochastic system Consider the following unstable hybrid stochastic system ˙x(t) = f(x(t), r(t)), t ⩾0, (7) (7) ˙x(t) = f(x(t), r(t)), t ⩾0, with the initial conditions x(0) = x0 ∈Rn and r(0) = r0 ∈S, where f : Rn × S →Rn, x(t) is the state, and r(t) is the mode. We are required to design a stochastic feedback control u(x([t/τ]τ), r([t/τ]τ))dB(t) based on the observations of state x([t/τ]τ) and mode r([t/τ]τ) at the discrete times 0, τ, 2τ, . . . such that the corresponding closed-loop system dx(t) = f(x(t), r(t))dt + u(x([t/τ]τ), r([t/τ]τ))dB(t), t ⩾0, (8) (8) dx(t) = f(x(t), r(t))dt + u(x([t/τ]τ), r([t/τ]τ))dB(t), t ⩾0, becomes almost surely exponentially stable, where the positive constant τ > 0 denotes the duration between two consecutive observations, [t/τ] is the integer part of t/τ, and u : Rn × S →Rn×m is the control input. For the existence and uniqueness of the solution to the controlled system, we impose the following assumption. becomes almost surely exponentially stable, where the positive constant τ > 0 denotes the duration between two consecutive observations, [t/τ] is the integer part of t/τ, and u : Rn × S →Rn×m is the control input. For the existence and uniqueness of the solution to the controlled system, we impose the following assumption. There exist two positive constants K1 and K2 such that Assumption 1. There exist two positive constants K1 and K2 such that \|f(x, i) −f(y, i)\| ⩽K1\|x −y\| and \|u(x, i) −u(y, i)\| ⩽K2\|x −y\| (9) (9) for all (x, y, i) ∈Rn × Rn × S. for all (x, y, i) ∈Rn × Rn × S. This assumption guarantees that for any initial state x(0) = x0 ∈Rn and mode r(0) = r0 ∈S, the controlled system (8) has a unique solution x(t) on t ∈R+ and E\|x(t)\|2 < ∞for all t ⩾0. In fact, for t ∈[0, τ], system (8) becomes This assumption guarantees that for any initial state x(0) = x0 ∈Rn and mode r(0) = r0 ∈S, the controlled system (8) has a unique solution x(t) on t ∈R+ and E\|x(t)\|2 < ∞for all t ⩾0. In fact, for t ∈[0, τ], system (8) becomes dx(t) = f(x(t), r(t))dt + u(x(0), r(0))dB(t) with the initial state x(0) = x0 and mode r(0) = r0. 2 Preliminaries and notation Throughout this paper, the notation is fairly standard. Here \| · \| denotes the Euclidean norm in Rn. For a vector or matrix A, AT denotes its transpose and \|A\| = p trace(ATA) represents the trace norm of matrix A. For a symmetric matrix A, λmax(A) and λmin(A) represent the largest and smallest eigen- value, respectively. Here (Ω, F, {Ft}t⩾0, P) is a complete probability space, where {Ft}t⩾0 satisﬁes the conditions that it is right continuous and F0 contains all P-null sets. In addition, we use B(t), t ⩾0, as an m-dimensional Brownian motion. The continuous-time Markov chain r(t), t ⩾0, takes discrete values in a given ﬁnite set S = {1, 2, . . ., N} and has the generator Γ = (γij)N×N given by P{r(t + ∆) = j\|r(t) = i} = ( γij∆+ o(∆), if i ̸= j, 1 + γii∆+ o(∆), if i = j, with ∆> 0 and γii = −P j̸=i γij. Here γij ⩾0 denotes the transition rate from i to j. The notation π = (π1, π2, . . . , πN) ∈R1×N represents a stationary (probability) distribution. Furthermore, one can ﬁnd the linear equation πΓ = 0 subject to PN j=1 πj = 1 and πj > 0 for all j ∈S. In particular, we recall that −γii = P j̸=i γij > 0. We state a lemma that estimates the probability of jumps. Lemma 1. For any t ⩾0, v > 0, and i ∈S, P(r(s) ̸= i, for some s ∈[t, t + v] r(t) = i) ⩽1 −e−ˆγv, (4) (4) where where ˆγ = max i∈S (−γii). (5) ˆγ = max i∈S (−γii). (5) To show this lemma, deﬁne the stopping time ζi = inf{s ⩾t : r(s) ̸= i}l, given r(t) = i, and let inf ∅= ∞where ∅denotes the empty set as usual. It is well known (see [14]) that ζi −t has the exponential probability distribution with parameter −γii. Hence, P(r(s) ̸= i for some s ∈[t, t + v]\|r(t) = i) = P(ζi −t ⩽v\|r(t) = i) = Z v 0 −γiieγiisds Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:4 = 1 −eγiiv ⩽1 −e−ˆγv (6) = 1 −eγiiv ⩽1 −e−ˆγv (6) as desired. as desired. dx(t) = f(x(t), r(t))dt + u(x(τ), r(τ))dB(t) dx(t) = f(x(t), r(t))dt + u(x(τ), r(τ))dB(t) with the initial state x(τ) and mode r(τ) at t = τ. It is easy to see that this hybrid stochastic system has a unique solution x(t) on t ∈[τ, 2τ] with E\|x(t)\|2 < ∞. Repeating this procedure, we can see what we have just claimed. Let us denote the solution by x(t; x0, r0). We see easily show that if x0 = 0, then x(t; 0, r0) = 0 for all t ⩾0 almost surely. This is known as the trivial solution. with the initial state x(τ) and mode r(τ) at t = τ. It is easy to see that this hybrid stochastic system has a unique solution x(t) on t ∈[τ, 2τ] with E\|x(t)\|2 < ∞. Repeating this procedure, we can see what we have just claimed. Let us denote the solution by x(t; x0, r0). We see easily show that if x0 = 0, then x(t; 0, r0) = 0 for all t ⩾0 almost surely. This is known as the trivial solution. The purpose of this paper is to ﬁnd suﬃcient conditions on the coeﬃcient f and the control input u as well as to obtain a positive bound τ ∗such that the controlled system (8) is almost surely exponentially stable as long as τ ⩽τ ∗. By the almost sure exponential stability, we mean that lim sup 1 t log(\|x(t; x0, r0)\|) < 0 almost surely, lim sup 1 t log(\|x(t; x0, r0)\|) < 0 almost surely, for any (x0, r0) ∈Rn × S (see [7,8,23,25]). We also observe that when τ →0, the controlled system (8) becomes the corresponding hybrid stochastic system for any (x0, r0) ∈Rn × S (see [7,8,23,25]). We also observe that when τ →0, the controlled system (8) becomes the corresponding hybrid stochastic system for any (x0, r0) ∈Rn × S (see [7,8,23,25]). We also observe that when τ →0, the controlled system (8) becomes the corresponding hybrid stochastic system dy(t) = f(y(t), r(t))dt + u(y(t), r(t))dB(t) (10) (10) on t ⩾0 with the initial condition (y(0), r(0)) = (x0, r0). Under Assumption 1, system (10) has a unique solution (see [17,20]). Denote the unique solution by y(t; x0, r0) on t ⩾0. on t ⩾0 with the initial condition (y(0), r(0)) = (x0, r0). Under Assumption 1, system (10) has a unique solution (see [17,20]). 2 Preliminaries and notation It is easy to show (see [27, Theorem 3.13]) that this hybrid stochastic system has a unique solution x(t) on t ∈[0, τ] with E\|x(t)\|2 < ∞. For t ∈[τ, 2τ], system (8) becomes with the initial state x(0) = x0 and mode r(0) = r0. It is easy to show (see [27, Theorem 3.13]) that this hybrid stochastic system has a unique solution x(t) on t ∈[0, τ] with E\|x(t)\|2 < ∞. For t ∈[τ, 2τ], system (8) becomes dx(t) = f(x(t), r(t))dt + u(x(τ), r(τ))dB(t) Denote the unique solution by y(t; x0, r0) on t ⩾0. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:5 3 Main results We see clearly from the discussion in the previous section that the conditions we need to impose should at least guarantee the almost sure exponential stability of the corresponding hybrid stochastic system (10). Although there are many useful criteria on the almost sure exponential stability, we use that established by [20]. Accordingly, we impose the following assumptions. Assumption 2. For each i ∈S, there exist constant triples αi ∈R, ρi ⩾0, and σi ⩾0 such that xTf(x, i) ⩽αi\|x\|2, \|u(x, i)\| ⩽ρi\|x\|, \|xTu(x, i)\| ⩾σi\|x\|2, (11) (11) for all x ∈Rn. Set ˆα = maxi∈S αi and ˆρ = maxi∈S ρi. Assumption 3. There is a constant p ∈(0, 1) such t for all x ∈Rn. Set ˆα = maxi∈S αi and ˆρ = maxi∈S ρi. Assumption 3. There is a constant p ∈(0, 1) such th for all x ∈Rn. Set α = maxi∈S αi and ρ = maxi∈S ρi. Assumption 3. There is a constant p ∈(0, 1) such that the N × N matrix Assumption 3. There is a constant p ∈(0, 1) such that the N × N matrix A(p) := diag(θ1(p), . . . , θN(p)) −Γ (12) A(p) := diag(θ1(p), . . . , θN(p)) −Γ (12) is a nonsingular M-matrix, where θi(p) := p(2 −p)σ2 i 2 −pρ2 i 2 −pαi (13) (13) σi, ρi are the constants speciﬁed in Assumption 2. and αi, σi, ρi are the constants speciﬁed in Assumption 2. on these assumptions. First, we point out that Assumptions 1 and 2 force Let us make some comments on these assumptions. First, we point out that Assumptions 1 and 2 force Let us make some comments on these assumptions. First, w f(0, i) = 0 and u(0, i) = 0, ∀i ∈S, (14) (14) which meet the stability purpose in this paper. In fact, g(0, i) = 0 follows from the second inequality in (11). To show that f(0, i) = 0 for all i ∈S, we assume otherwise that there were some i ∈S such that z := f(0, i) ̸= 0. Choose a constant b such that 0 < b < 1/(K1 + \|αi\|) and let x = bz. Then, by the ﬁrst inequality in (11), bzTf(bz, i) ⩽αib2\|z\|2 ⩽\|αi\|b2\|z\|2. bzTf(bz, i) ⩽αib2\|z\|2 ⩽\|αi\|b2\|z\|2. bzTf(bz, i) ⩽αib2\|z\|2 ⩽\|αi\|b2\|z\|2. On the other hand, by Assumption 1, bzTf(bz, i) = b\|z\|2 + bzT(f(bz, i) −f(0, i)) ⩾b\|z\|2 −K1b2\|z\|2. 3 Main results Hence, b\|z\|2 −K1b2\|z\|2 ⩽\|αi\|b2\|z\|2. This implies that 1 ⩽(K1 + \|αi\|)b, but this is a contradiction as b < 1/(K1 + \|ai\|). We therefore must have (14). Recalling that y(t; x0, r0) denotes the solution of the hybrid stochastic system (10), we can hence highlight a signiﬁcant property given in Mao [20, Lemma 2.1], which then leads to but this is a contradiction as b < 1/(K1 + \|ai\|). We therefore must have (14). Recalling that y(t; x0, r0) denotes the solution of the hybrid stochastic system (10), we can hence highlight a signiﬁcant property given in Mao [20, Lemma 2.1], which then leads to P{y(t; x0, r0) ̸= 0 on t ⩾0} = 1, ∀x0 ̸= 0. (15) (15) That is, if any initial solution of system (10) is a nonzero state, almost all the trajectories of system (10) will never converge to the origin. Thus, Lyapunov functions can be chosen in a variety of ways. That is, if any initial solution of system (10) is a nonzero state, almost all the trajectories of system (10) will never converge to the origin. Thus, Lyapunov functions can be chosen in a variety of ways. We also emphasize that we are only interested in the case when ˆα ⩾0 in this paper; otherwise, the given hybrid system (7) is already stable (see [28]) and there is no need to stabilize it using feedback control. We should also point out that we always have ˆα ⩽K1 and ˆρ ⩽K2, but we might have ˆα < K1 and ˆρ < K2 in many cases. For example, consider the scalar case where f(x, i) = −x + ai sin(x) and u(x, i) = x −bi sin(x) where ai ∈[1, 2], bi ∈(0, 1], but a1 = 2 and b1 = 1. It is easy to see that K1 = 3 and K2 = 2. On the other hand, xf(x, i) = −x2 + ai sin(x)x ⩽(ai −1)\|x\|2 so ˆα = 1 ⩽K1. Moreover, 0 ⩽\|u(x, i)\| ⩽\|x\| for \|x\| ⩽π whereas 0 ⩽\|u(x, i)\| ⩽\|x\| + bi ⩽(1 + bi/π)\|x\| for \|x\| > π so we have Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:6 ˆρ = (1 + 1/π) < K2. We also observe that once Assumption 2 holds, the veriﬁcation of Assumption 3 depends very much on the choice of p ∈(0, 1). 3 Main results In Appendix A, we give some easier conditions that guarantee the existence of such a p and, hence, for Assumption 3 to hold. The following lemma shows that the corresponding hybrid stochastic system (10) is exponentially stable in the pth moment and, hence, by [27, Theorem 5.9], the hybrid stochastic system is also almost surely exponentially stable. Lemma 2. Let Assumptions 1–3 hold. Deﬁne (c1, . . . , cN)T = A−1(p)(1, . . . , 1)T, (16) (16) so all ci are positive by the theory of M-matrices [2, 27] or by Lemma A1 in the Appendix A and let cmin = min1⩽i⩽N ci and cmax = max1⩽i⩽N ci. Then the solution of the hybrid stochastic system (10) satisﬁes E\|y(t; x0, r0)\|p ⩽M\|x0\|pe−γt, ∀t ⩾0, (17) (17) Proof. For x0 = 0, that is y(t; 0, r0) = 0, we can deduce that the assertion is natural. Fix x0 ̸= 0 and r0 ∈S arbitrarily and write y(t; x0, r0) = y(t). Recalling (15), we have that y(t) ̸= 0 for all t ⩾0 almost surely. Deﬁne the Lyapunov function V (y, t, i) = ci\|y\|peγt, for (y, t, i) ∈(Rn −{0}) × R+ × S. We can therefore apply the generalized Itˆo formula (see [27, Theorem 1.45]) to obtain that apply the generalized Itˆo formula (see [27, Theorem 1.45]) to obtain that EV (y(t), t, r(t)) = V (x0, 0, r(0)) + E Z t 0 LV (y(s), s, r(s))ds, (18) t ⩾0 h LV (Rn {0}) R S R i d ﬁ d b EV (y(t), t, r(t)) = V (x0, 0, r(0)) + E Z t 0 LV (y(s), s, r(s))ds, (18) (18) for t ⩾0, where LV : (Rn −{0}) × R+ × S →R is deﬁned by LV ( t i) γt \| \|p + \| \|p−2 Tf( i) + pci \| \|p−2\| ( i)\|2 for t ⩾0, where LV : (Rn −{0}) × R+ × S →R is deﬁned by for t ⩾0, where LV : (Rn −{0}) × R+ × S →R is deﬁned by LV (y, t, i) = eγt γci\|y\|p + pci\|y\|p−2yTf(y, i) + pci 2 \|y\|p−2\|u(y, i)\|2 −p(2 −p)ci 2 \|y\|p−4\|yTu(y, i)\|2 + N X j=1 γijcj\|y\|p . By Assumption 2 and then using deﬁnition (13) of θi(p), we have LV (y, t, i) ⩽eγt\|y\|p  1 −ciθi(p) + N X j=1 γijcj  . 3 Main results ciθi(p) − N X j=1 γijcj = 1, ∀i ∈S. Hence, we have LV (y, t, i) ⩽0. LV (y, t, i) ⩽0. Substituting this into (18) yields Substituting this into (18) yields EV (y(t), t, r(t)) ⩽V (x0, 0, r(0)). p cmineγtE\|y(t)\|p ⩽cmax\|x0\|p, cmineγtE\|y(t)\|p ⩽cmax\|x0\|p, cmineγtE\|y(t)\|p ⩽cmax\|x0\|p, which is the desired assertion (17). which is the desired assertion (17). which is the desired assertion (17). To simplify our notation, we let δt = [t/τ]τ for t ⩾0 and set tk = kτ for k = 0, 1, 2, . . . from now on. which is the desired assertion (17). To simplify our notation we let δt = [t/τ]τ for t ⩾0 and set tk = kτ for k = 0 1 2 from now on To simplify our notation, we let δt = [t/τ]τ for t ⩾0 and set tk = kτ for k = 0, 1, 2, . . . from now on. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:7 Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:7 Lemma 3. Let Assumptions 1 and 2 hold. Then for any initial condition (x0, r0) ∈Rn × S, E\|x(t; x0 r0)\|2 ⩽\|x0\|2e(2ˆα+ˆρ2)t (1 Lemma 3. Let Assumptions 1 and 2 hold. Then for any initial condition (x0, r0) ∈Rn × S, E\|x(t; x0, r0)\|2 ⩽\|x0\|2e(2ˆα+ˆρ2)t E\|x(t; x0, r0)\|2 ⩽\|x0\|2e(2ˆα+ˆρ2)t (19) E\|x(t; x0, r0)\|2 ⩽\|x0\|2e(2ˆα+ˆρ2)t (19) (19) and and E\|x(t; x0, r0) −x(δt; x0, r0)\|2 ⩽2τ(K2 1τ + ˆρ2)\|x0\|2e(2ˆα+ˆρ2)t ⩾0 (20) (20) for all t ⩾0. ⩾ Proof. Fix any (x0, r0) ∈Rn × S and write x(t; x0, r0) = x(t). By the Itˆo formula and Assumption 2, it follows that Proof. Fix any (x0, r0) ∈Rn × S and write x(t; x0, r0) = x(t). By the Itˆo formula and Assumption 2, it follows that E\|x(t)\|2 ⩽\|x0\|2 + E Z t 0 (2ˆα\|x(s)\|2 + ˆρ2\|x(δs)\|2)ds ⩽\|x0\|2 + (2ˆα + ˆρ2) Z t 0 sup 0⩽u⩽s E\|x(u)\|2 ds, for t ⩾0. As the second term on the right-hand side is increasing in t, we can obtain sup 0⩽u⩽t E\|x(t)\|2 ⩽\|x0\|2 + (2ˆα + ˆρ2) Z t 0 sup 0⩽u⩽s E\|x(u)\|2 ds, which implies the desire assertion (19) by the well-known Gronwall inequality. 3 Main results Moreover, by Assump- tions 1 and 2, we further derive that which implies the desire assertion (19) by the well-known Gronwall inequality. Moreover, by Assump- tions 1 and 2, we further derive that E\|x(t) −x(δt)\|2 ⩽2τK2 1 Z t δt E\|x(s)\|2ds + 2τ ˆρ2E\|x(δt)\|2. This, together with (19), implies another assertion (20). This, together with (19), implies another assertion (20). Lemma 4. Let Assumptions 1 and 2 hold and p ∈(0, 1). Then for any initial condition (x0, r0) ∈Rn×S, /2 Let Assumptions 1 and 2 hold and p ∈(0, 1). Then for any initial condition (x0, r0) ∈Rn×S, Lemma 4. Let Assumptions 1 and 2 hold and p ∈(0, 1). Then for any initial condition (x E\|x(t; x0, r0) −y(t; x0, r0)\|p ⩽\|x0\|pep(K1+1.5K2 2)t H(τ) e(2ˆα+ˆρ2)t −1 p/2 , (21) E\|x(t; x0, r0) −y(t; x0, r0)\|p ⩽\|x0\|pep(K1+1.5K2 2)t H(τ) e(2ˆα+ˆρ2)t −1 p/2 , (21) (21) for all t ⩾0, where H(τ) = 6K2 2 τ(K2 1τ + ˆρ2) + 2(1 −e−ˆγτ) 2ˆα + ˆρ2 . Proof. Fix any (x0, r0) ∈Rn × S and set x(t; x0, r0) = x(t) and y(t; x0, r0) = y(t). By applying the Itˆo formula and Assumption 1, it can be veriﬁed that E\|x(t) −y(t)\|2 ⩽2K1E Z t 0 \|x(s) −y(s)\|2ds + J1(t), (22) (22) for t ⩾0, where J1(t) = E Z t 0 \|u(x(δs), r(δs)) −u(y(s), r(s))\|2ds ⩽3E Z t 0 \|u(x(δs), r(δs)) −u(x(δs), r(s))\|2 +\|u(x(δs), r(s)) −u(x(s), r(s))\|2 +\|u(x(s), r(s)) −u(y(s), r(s))\|2 ds ⩽3E Z t 0 \|u(x(δs), r(δs)) −u(x(δs), r(s))\|2 +K2 2\|x(δs) −x(s)\|2 + K2 2\|x(s) −y(s)\|2 ds. Substituting this into (22) yields E\|x(t) −y(t)\|2 ⩽(2K1 + 3K2 2) Z t 0 E\|x(s) −y(s)\|2ds + 3K2 2J2(t) + 3J3(t), (23) E\|x(t) −y(t)\|2 ⩽(2K1 + 3K2 2) Z t 0 E\|x(s) −y(s)\|2ds + 3K2 2J2(t) + 3J3(t), (23) (23) Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:8 where where J2(t) = Z t 0 E\|x(δs) −x(s)\|2ds, J3(t) = Z t 0 E\|u(x(δs), r(δs)) −u(x(δs), r(s))\|2ds. ere J2(t) = Z t 0 E\|x(δs) −x(s)\|2ds, J2(t) = Z t 0 E\|x(δs) −x(s)\|2ds, J3(t) = Z t 0 E\|u(x(δs), r(δs)) −u(x(δs), r(s))\|2ds. However, by Lemma 3, However, by Lemma 3, J2(t) ⩽2τ(K2 1τ + ˆρ2)\|x0\|2 2ˆα + ˆρ2 h e(2ˆα+ˆρ2)t −1 i . J2(t) ⩽2τ(K2 1τ + ˆρ2)\|x0\|2 2ˆα + ˆρ2 h e(2ˆα+ˆρ2)t −1 i . 3 Main results (24) To estimate J3(t), we let κ = κ(t) = [t/τ]. Then J3(t) = κ X k=0 Z t∧tk+1 tk E\|u(x(δs), r To estimate J3(t), we let κ = κ(t) = [t/τ]. Then J3(t) = κ X k=0 Z t∧tk+1 tk E\|u(x(δs), r(δs)) −u(x(δs), r(s))\|2ds. (25) (25) By Assumption 2, we can derive that, for tk ⩽s ⩽t ∧tk+1, By Assumption 2, we can derive that, for tk ⩽s ⩽t ∧tk+1, By Assumption 2, we can derive that, for tk ⩽s ⩽t ∧tk+1, By Assumption 2, we can derive that, for tk ⩽s ⩽t ∧tk+1, E\|u(x(δs), r(δs)) −u(x(δs), r(s))\|2 = E\|u(x(tk), r(tk)) −u(x(tk), r(s))\|2 = E h E \|u(x(tk), r(tk)) −u(x(tk), r(s))\|2\|Ftk i ⩽E h 4ˆρ2\|x(tk)\|2E I{r(s)̸=r(tk)}\|Ftk i = E " 4ˆρ2\|x(tk)\|2E X i∈S I{r(tk)=i}I{r(s)̸=i} Ftk !# = E " 4ˆρ2\|x(tk)\|2 X i∈S I{r(tk)=i}P(r(s) ̸= i r(tk) = i # . (26) (26) However, by Lemma 1, However, by Lemma 1, However, by Lemma 1, P(r(s) ̸= i r(tk) = i ⩽P(r(¯s) ̸= i for some ¯s ∈[tk, t ∧tk+1] r(tk) = i ⩽1 −e−ˆγτ. Hence, E\|u(x(δs), r(δs)) −u(x(δs), r(s))\|2 ⩽E h 4ˆρ2\|x(tk)\|2(1 −e−ˆγτ) i = 4ˆρ2(1 −e−ˆγτ)E\|x(tk)\|2. (27) E\|u(x(δs), r(δs)) −u(x(δs), r(s))\|2 ⩽E h 4ˆρ2\|x(tk)\|2(1 −e−ˆγτ) i = 4ˆρ2(1 −e−ˆγτ)E\|x(tk)\|2. (27) E\|u(x(δs), r(δs)) −u(x(δs), r(s))\|2 ⩽E h 4ˆρ2\|x(tk)\|2(1 −e−ˆγτ) i = 4ˆρ2(1 −e−ˆγτ)E\|x(tk)\|2. (27) (27) Substituting this into (25), we obtain Substituting this into (25), we obtain J3(t) ⩽4ˆρ2(1 −e−ˆγτ) κ X k=0 Z t∧tk+1 tk E\|x(tk)\|2ds. (28) (28) However, by Lemma 3, we then have However, by Lemma 3, we then have κ X k=0 Z t∧tk+1 tk E\|x(tk)\|2ds ⩽ κ X k=0 Z t∧tk+1 tk \|x0\|2e(2ˆα+ˆρ2)tkds ⩽ κ X k=0 Z t∧tk+1 tk \|x0\|2e(2ˆα+ˆρ2)sds = \|x0\|2 Z t 0 e(2ˆα+ˆρ2)sds = \|x0\|2 2ˆα + ˆρ2 h e(2ˆα+ˆρ2)t −1 i . Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:9 Putting this into (28) gives J3(t) ⩽4ˆρ2(1 −e−ˆγτ)\|x0\|2 2ˆα + ˆρ2 h e(2ˆα+ˆρ2)t −1 i . (29) (29) Substituting (24) and (29) into (23), we obtain E\|x(t) −y(t)\|2 ⩽(2K1 + 3K2 3) Z t 0 E\|x(s) −y(s)\|2ds + \|x0\|2H(τ) h e(2ˆα+ˆρ2)t −1 i . (30) (30) Using the well-known Gronwall inequality, we have E\|x(t) −y(t)\|2 ⩽\|x0\|2H(τ)e(2K1+3K2 3)th e(2ˆα+ˆρ2)t −1 i . Finally, we can obtain the desired assertion (21) by applying the H¨older inequality. Lemma 5. Let Assumptions 1–3 hold. Choose a free parameter ε ∈(0, 1). 3 Main results Let ¯τ > 0 be the unique root to the equation Finally, we can obtain the desired assertion (21) by applying the H¨older inequality. Lemma 5. Let Assumptions 1–3 hold. Choose a free parameter ε ∈(0, 1). Let ¯τ > 0 be the unique root to the equation a 5. Let Assumptions 1–3 hold. Choose a free parameter ε ∈(0, 1). Let ¯τ > 0 be the unique he equation ep(K1+1.5K2 2)(τ+log(M/ε)/γ) H(τ) e(2ˆα+ˆρ2)(τ+log(M/ε)/γ) −1 p/2 = 1 −ε, (31) (31) where H(τ) and γ, M have been given in Lemmas 2 and 4, respectively. Then, for each τ ∈(0, ¯τ], there exists a pair of positive integer ¯κ and number λ such that, for every initial condition (x0, r0) ∈Rn × S, the solution of system (8) satisﬁes where H(τ) and γ, M have been given in Lemmas 2 and 4, respectively. Then, for each τ ∈(0, ¯τ], there exists a pair of positive integer ¯κ and number λ such that, for every initial condition (x0, r0) ∈Rn × S, the solution of system (8) satisﬁes E\|x(k¯κτ; x0, r0)\|p ⩽\|x0\|pe−λk¯κτ, ∀k = 1, 2, 3, . . .. (32) (32) Proof. It is easy to see that the term on the left-hand side of (31) is a continuous increasing function of τ ⩾0 and is equal to zero when τ = 0 whereas it tends to inﬁnity as τ →∞, thus Eq. (31) must have a unique root ¯τ > 0. Fix τ ∈(0, ¯τ] and (x0, r0) ∈Rn × S arbitrarily and write x(kτ; x0, r0) = xk for k = 0, 1, 2, . . .. Let ¯κ be the smallest positive integer that is no less than log(M/ε) γτ , namely Proof. It is easy to see that the term on the left-hand side of (31) is a continuous increasing function of τ ⩾0 and is equal to zero when τ = 0 whereas it tends to inﬁnity as τ →∞, thus Eq. (31) must have a unique root ¯τ > 0. Fix τ ∈(0, ¯τ] and (x0, r0) ∈Rn × S arbitrarily and write x(kτ; x0, r0) = xk for k = 0, 1, 2, . . .. 3 Main results Let ¯κ be the smallest positive integer that is no less than log(M/ε) γτ , namely log(M/ε) γτ ⩽¯κ < 1 + log(M/ε) γτ , (33) (33) where γ and M have been deﬁned in Lemma 2. This implies Me−γ¯κτ ⩽ε. (34) Me−γ¯κτ ⩽ε. (34) Write y(¯κτ; x0, r0) = y¯κ. By Lemma 2, Write y(¯κτ; x0, r0) = y¯κ. By Lemma 2, E\|y¯κ\|p ⩽M\|x0\|pe−γ¯κτ ⩽ε\|x0\|p. (35) E\|y¯κ\|p ⩽M\|x0\|pe−γ¯κτ ⩽ε\|x0\|p. (35) By the elementary inequality (a + b)p ⩽ap + bp for any a, b ⩾0 and 0 < p < 1, one can obtain E\|x¯κ\|p ⩽E\|y¯κ\|p + E\|y¯κ −x¯κ\|p. By the elementary inequality (a + b)p ⩽ap + bp for any a, b ⩾0 and 0 < p < 1, one can obtain E\|x¯κ\|p ⩽E\|y¯κ\|p + E\|y¯κ −x¯κ\|p. E\|x¯κ\|p ⩽E\|y¯κ\|p + E\|y¯κ −x¯κ\|p. E\|x¯κ\|p ⩽E\|y¯κ\|p + E\|y¯κ −x¯κ\|p. Using (35) and Lemma 4, we obtain that Using (35) and Lemma 4, we obtain that E\|x¯κ\|p ⩽\|x0\|pn ε + ep(K1+1.5K2 2)¯κτ H(τ) e(2ˆα+ˆρ2)¯τ −1 p/2o . (36) Noting from (33) that ¯κτ < τ + log(M/ε)/γ, we have ε + ep(K1+1.5K2 2)¯κτ H(τ) e(2ˆα+ˆρ2)¯τ −1 p/2 < ε + ep(K1+1.5K2 2)(τ+log(M/ε)/γ) H(τ) e(2ˆα+ˆρ2)(τ+log(M/ε)/γ) −1 p/2 ⩽1, (36) ε + ep(K1+1.5K2 2)¯κτ H(τ) e(2ˆα+ˆρ2)¯τ −1 p/2 < ε + ep(K1+1.5K2 2)(τ+log(M/ε)/γ) H(τ) e(2ˆα+ˆρ2)(τ+log(M/ε)/γ) −1 p/2 ⩽1, < ε + ep(K1+1.5K2 2)(τ+log(M/ε)/γ) H(τ) e(2ˆα+ˆρ2)(τ+log(M/ε)/γ) −1 p/2 ⩽1 Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:10 Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:10 where Eq. (31) has been used. We may therefore write ε + ep(K1+1.5K2 2)¯κτ H(τ) e(2ˆα+ˆρ2)¯τ −1 p/2 = e−λ¯κτ, for some λ > 0. It then follows from (36) that E\|x¯κ\|p ⩽\|x0\|pe−λ¯κτ. (37) E\|x¯κ\|p ⩽\|x0\|pe−λ¯κτ. (37) Let us now discuss the solution x(t) of hybrid stochastic system (8) on t ⩾¯κτ. This can be regarded as the solution of the hybrid stochastic system (8) with initial condition (x¯κ, r(¯κτ)) at time t = ¯κτ. Owing to the time-homogeneous property of hybrid stochastic system (8), we can thus easily show that E(\|x2¯κ\|p\|F¯κτ) ⩽\|x¯κ\|pe−λ¯κτ. This implies E\|x2¯κ\|p ⩽E\|x¯κ\|pe−λ¯κτ ⩽\|x0\|pe−2λ¯κτ. Repeating this procedure, we have E\|xk¯κ\|p ⩽E\|x(k−1)¯κ\|pe−λ¯κτ ⩽\|x0\|pe−λk¯κτ, ∀k = 1, 2, 3, . . . as desired. The proof is, hence, complete. as desired. The proof is, hence, complete. 3 Main results we are in the position to present and prove our main results in this section. Now, we are in the position to present and prove our main results in this section. Theorem 1. Let Assumptions 1–3 hold. Then there exists a positive number τ ∗such that the stochastic controlled hybrid system (8) is almost surely exponentially stable provided τ ⩽τ ∗. Proof. Choose a free parameter ε ∈(0, 1). We note that the unique root of (31) is ¯τ > 0. Let τ ∗= ¯τ. ( ] ( ) ( ) ( ) Theorem 1. Let Assumptions 1–3 hold. Then there exists a positive number τ ∗such that the stochastic controlled hybrid system (8) is almost surely exponentially stable provided τ ⩽τ ∗. Theorem 1. Let Assumptions 1–3 hold. Then there exists a positive number τ ∗such that the stochastic controlled hybrid system (8) is almost surely exponentially stable provided τ ⩽τ ∗. Proof. Choose a free parameter ε ∈(0, 1). We note that the unique root of (31) is ¯τ > 0. Let τ ∗= ¯τ. First, we set τ ∈(0, τ ∗] and (x0, r0) ∈R × S, and then we write x(t; x0, r0) = x(t). In addition, let ¯κ and xk¯κ be the same as deﬁned in the proof of Lemma 5. For any t ⩾0, we can ﬁnd a unique integer k such that t ∈[k¯κτ, (k + 1)¯κτ). By the time-homogeneous property of system (8), we see from Lemma 3 that y y ( ) y p y p ⩽ Proof. Choose a free parameter ε ∈(0, 1). We note that the unique root of (31) is ¯τ > 0. Let τ ∗= ¯τ. First, we set τ ∈(0, τ ∗] and (x0, r0) ∈R × S, and then we write x(t; x0, r0) = x(t). In addition, let ¯κ and xk¯κ be the same as deﬁned in the proof of Lemma 5. For any t ⩾0, we can ﬁnd a unique integer k such that t ∈[k¯κτ, (k + 1)¯κτ). By the time-homogeneous property of system (8), we see from Lemma 3 that E \|x(t)\|2Fk¯κτ ⩽\|xk¯κ\|2e(2ˆα+ˆρ2)(t−k¯κτ) ⩽\|xk¯κ\|2e(2ˆα+ˆρ2)¯κτ. An application of the H¨older inequality yields An application of the H¨older inequality yields E \|x(t)\|pFk¯κτ ⩽C1\|xk¯κ\|p, E \|x(t)\|pFk¯κτ ⩽C1\|xk¯κ\|p, where C1 = e(ˆα+0.5ˆρ2)p¯κτ. This, together with Lemma 5, implies where C1 = e(ˆα+0.5ˆρ2)p¯κτ. 3 Main results E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ⩽2C3e−λtk. Consequently, E sup tk⩽t⩽tk+1 \|x(t)\|p = E max 0⩽l⩽z−1 sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p E sup tk⩽t⩽tk+1 \|x(t)\|p = E max 0⩽l⩽z−1 sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ⩽ z−1 X l=0 E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ⩽C4e−λtk, (40) E sup tk⩽t⩽tk+1 \|x(t)\|p = E max 0⩽l⩽z−1 sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ⩽ z−1 X l=0 E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ⩽C4e−λtk, (40) for all k = 0, 1, 2, . . ., where C4 = 2zC3. This implies ⩽ z−1 X l=0 E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ⩽C4e−λtk, (40) (40) l=0 k ⩽⩽k ( ) for all k = 0, 1, 2, . . ., where C4 = 2zC3. This implies P sup tk⩽t⩽tk+1 \|x(t)\|p ⩾e−0.5λtk ⩽C4e−0.5λtk, for all k = 0, 1, 2, . . ., where C4 = 2zC3. This implies P sup tk⩽t⩽tk+1 \|x(t)\|p ⩾e−0.5λtk ⩽C4e−0.5λtk, for all i ⩾1. The Borel-Cantelli lemma shows that, with probability 1, for all i ⩾1. The Borel-Cantelli lemma shows that, with probability 1, sup tk⩽t⩽tk+1 \|x(t)\|p < e−0.5λtk sup tk⩽t⩽tk+1 \|x(t)\|p < e−0.5λtk holds for all but ﬁnitely many k. That is, for almost all ω ∈Ω, there is an integer k0 = k0(ω) such that sup \|x(t, ω)\|p < e−0.5λtk, ∀k ⩾k0(ω). sup t ⩽t⩽t \|x(t, ω)\|p < e−0.5λtk, ∀k ⩾k0(ω). sup tk⩽t⩽tk+1 \|x(t, ω)\|p < e−0.5λtk, ∀k ⩾k0(ω). sup tk⩽t⩽tk+1 \|x(t, ω)\|p < e−0.5λtk, ∀k ⩾k0(ω). sup tk⩽t⩽tk+1 \|x(t, ω)\|p < e−0.5λtk, ∀k ⩾k0(ω). Therefore, for tk ⩽t ⩽tk+1 and k ⩾k0, Therefore, for tk ⩽t ⩽tk+1 and k ⩾k0, 1 t log(\|x(t, ω)\|) < −0.5λtk ptk+1 = −0.5λk p(k + 1). Letting t →∞, we obtain for almost all ω ∈Ω. This completes the proof. 3 Main results This, together with Lemma 5, implies E\|x(t)\|p ⩽C1E\|xk¯κ\|p ⩽C1\|x0\|pe−λk¯κτ ⩽C2\|x0\|pe−λt, (38) E\|x(t)\|p ⩽C1E\|xk¯κ\|p ⩽C1\|x0\|pe−λk¯κτ ⩽C2\|x0\|pe−λt, (38) (38) where C2 = C1eλ¯κτ. In other words, we have shown that the controlled system (8) is exponentially stable in the pth moment. However, this is not yet what we require. In the remainder of this proof, we show that this pth moment exponential stability yields the almost sure exponential stability as desired. We should of course point out that [27, Theorem 5.9] shows this implication for hybrid stochastic systems, but our controlled system (8) is, in fact, a hybrid stochastic delay system. In the area of hybrid stochastic delay systems, Mao et al. [27, Theorem 7.24] showed this implication for p ⩾1, but here we have p ∈(0, 1). Let z be a positive integer suﬃciently large for τK1 z p ⩽0.5. (39) (39) Set ε = τ/z. Let integers k ⩾0 and 0 ⩽l ⩽z −1 be arbitrary. For t ∈[tk + lε, tk + (l + 1)ε], it follows from system (8) that Set ε = τ/z. Let integers k ⩾0 and 0 ⩽l ⩽z −1 be arbitrary. For t ∈[tk + lε, tk + (l + 1)ε], it follows from system (8) that x(t) = x(tk + lε) + Z t tk+lε f(x(s), r(s))ds + Z t tk+lε u(x(tk), r(tk))dB(s). Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:11 Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:11 By the Burkholder–Davis–Gandy inequality (see [25]) and inequality (38), we then derive that ! By the Burkholder–Davis–Gandy inequality (see [25]) and inequality (38), we then derive that E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ! ⩽E\|x(tk + lε)\|p + E sup tk+lε⩽t⩽tk+(l+1)ε Z t tk+lε f(x(s), r(s))ds p ! + E sup tk+lε⩽t⩽tk+(l+1)ε Z t tk+lε u(x(tk), r(tk))dB(s) p ! ⩽E\|x(tk + lε)\|p + εpKp 1E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ! + cpE εˆρ2\|x(tk)\|2p/2 ⩽C3e−λtk + εpKp 1E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ! , ⩽E\|x(tk + lε)\|p + εpKp 1E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ! + cpE εˆρ2\|x(tk)\|2p/2 ⩽C3e−λtk + εpKp 1E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ! , where C3 = C2\|x0\|p(1 + cpεp/2ˆρp) and cp is the constant from the Burkholder-Davis-Gandy inequality. By (39) we hence have where C3 = C2\|x0\|p(1 + cpεp/2ˆρp) and cp is the constant from the Burkholder-Davis-Gandy inequality. By (39), we hence have E sup tk+lε⩽t⩽tk+(l+1)ε \|x(t)\|p ⩽2C3e−λtk. 4.1 Observable in all modes We ﬁrst consider the case where the state x(t) is observable in every mode i ∈S. For each i ∈S, choose a matrix Di ∈Rn×n such that ∥Di∥= 1 and λmin(Di + DT i ) ⩾ √ 3. (44) ∥Di∥= 1 and λmin(Di + DT i ) ⩾ √ 3. (44) Choose a nonnegative number δi such that Choose a nonnegative number δi such that δ2 i > 4αi. (45) at δ2 i > 4αi. (45) δ2 i > 4αi. (45) δ2 i > 4αi. (45) Let Ai δiDi. Noting that \|u(x, i)\| = \|Aix\| ⩽δi\|x\| and \|xTu(x, i)\| = \|xTAix\| = 0.5xT(Ai + AT i )x\| ⩾ p 3/4δi\|x\|2 \|u(x, i)\| = \|Aix\| ⩽δi\|x\| and \|xTu(x, i)\| = \|xTAix\| = 0.5xT(Ai + AT i )x\| ⩾ p 3/4δi\|x\|2 \|u(x, i)\| = \|Aix\| ⩽δi\|x\| and \|xTu(x, i)\| = \|xTAix\| = 0.5xT(Ai + AT i )x\| ⩾ p 3/4δi\|x\|2 for all x ∈Rn, we see that the 2nd and 3rd inequality in (11) hold with ρi = δi and σi = p 3/4δi. By (45) for all x ∈Rn, we see that the 2nd and 3rd inequality in (11) hold with ρi = δi and σi = p 3/4δi. By (45), ∈Rn, we see that the 2nd and 3rd inequality in (11) hold with ρi = δi and σi = p 3/4δi. By for all x ∈Rn, we see that the 2nd and 3rd inequality in (11) hold with ρi = δi and σi = p 3 (45), σ2 i −0.5ρ2 i −αi = 0.25δ2 i −αi > 0. σ2 i −0.5ρ2 i −αi = 0.25δ2 i −αi > 0. σ2 i −0.5ρ2 i −αi = 0.25δ2 i −αi > 0. We can therefore ﬁnd a p ∈(0, 1) suﬃciently small for θi(p) = p (2 −p)σ2 i 2 −ρ2 i 2 −αi > 0, ∀i ∈S. (46) (46) ently, recalling (12), we have Consequently, recalling (12), we have Consequently, recalling (12), we have A(p)1 = (θ1(p), . . . , θN(p))T > 0, A(p)1 = (θ1(p), . . . , θN(p))T > 0, where 1 = (1, . . . , 1)T ∈RN. By Lemma A1, A(p) is a nonsingular M-matrix. In other words, we have veriﬁed Assumption 3 under condition (45). 4.1 Observable in all modes In summary, we can conclude by using Theorem 1 that if we choose Di and δi to meet conditions (44) and (45), respectively, and let Ai = δiDi for each i ∈S, there is a positive scalar τ ∗such that the stochastic controlled hybrid system (43) is almost surely exponentially stable provided that τ ⩽τ ∗. where 1 = (1, . . . , 1)T ∈RN. By Lemma A1, A(p) is a nonsingular M-matrix. In other words, we have veriﬁed Assumption 3 under condition (45). In summary, we can conclude by using Theorem 1 that if we choose Di and δi to meet conditions (44) and (45), respectively, and let Ai = δiDi for each i ∈S, there is a positive scalar τ ∗such that the stochastic controlled hybrid system (43) is almost surely exponentially stable provided that τ ⩽τ ∗. 4 Design of linear feedback controls Consider the following unstable hybrid stochastic system ˙x(t) = f(x(t), r(t)), t ⩾0, (41) (41) ˙x(t) = f(x(t), r(t)), t ⩾0, Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:12 with the initial state x(0) = x0 ∈Rn and mode r(0) = r0 ∈S, where f : Rn × S →Rn. As before, we assume that f meets conditions (9) and (11), namely there are constants K1 > 0 and αi ∈R (i ∈S) such th t with the initial state x(0) = x0 ∈Rn and mode r(0) = r0 ∈S, where f : Rn × S →Rn. As before, we assume that f meets conditions (9) and (11), namely there are constants K1 > 0 and αi ∈R (i ∈S) such that \|f(x, i) −f(y, i)\| ⩽K1\|x −y\| and xTf(x, i) ⩽αi\|x\|2, (42) \|f(x, i) −f(y, i)\| ⩽K1\|x −y\| and xTf(x, i) ⩽αi\|x\|2, (42) for x, y ∈Rn and i ∈S. Instead of nonlinear feedback controls, we are now looking for linear feedback controls. To avoid the notation becoming complicated, we set B(t) be a scalar Brownian motion in this section (and leave the multi-dimensional case to the reader). The linear feedback control function we look for is of the form u(x, i) = A(i)x so the controlled system becomes dx(t) = f(x(t), r(t)) + A(r[t/τ])x([t/τ])dB(t), (43) (43) dx(t) = f(x(t), r(t)) + A(r[t/τ])x([t/τ])dB(t), (43) where A(i) ∈Rn×n for i ∈S and we will often write A(i) = Ai. Noting that \|u(x, i) −u(y, i)\| ⩽∥Ai∥\|x −y\|, ∀x, y ∈Rn, i ∈S, where A(i) ∈Rn×n for i ∈S and we will often write A(i) = Ai. Noting that \|u(x i) −u(y i)\| ⩽∥Ai∥\|x −y\| ∀x y ∈Rn i ∈S where A(i) ∈Rn×n for i ∈S and we will often write A(i) = Ai. Noting that \|u(x, i) −u(y, i)\| ⩽∥Ai∥\|x −y\|, ∀x, y ∈Rn, i ∈S, we see that the second inequality in (9) holds with K2 = maxi∈S ∥Ai∥. 4.2 Observable in some modes We therefore see that the 2nd and 3rd inequality in (11) hold with ρi = δi and σi = p 3/4δi for i ∈S2 whereas ρi = σi = 0 for i ∈S1. Deﬁne Choose a matrix Di satisfying condition (44) and let Ai = δiDi. We therefore see that the 2nd and 3rd inequality in (11) hold with ρi = δi and σi = p 3/4δi for i ∈S2 whereas ρi = σi = 0 for i ∈S1. Deﬁne ξ = ( ¯ Ntimes z }\| { 1, · · · , 1, N−¯ Ntimes z }\| { β, . . . , β )T, and set (ζ1, . . . , ζN)T := A(p)ξ. (ζ1, . . . , ζN)T := A(p)ξ. Then, for i ∈S1, ζi = −pαi − ¯ N X j=1 γij −β N X j= ¯ N+1 γij = −pαi + (1 −β) N X j= ¯ N+1 γij > 0 by (48), whereas for i ∈S2, ζi = βp (2 −3p)δ2 i 8 −αi − ¯ N X j=1 γij −β N X j= ¯ N+1 γij = βp (2 −3p)δ2 i 8 −αi −(1 −β) ¯ N X j=1 γij > 0 > 0 by (49). By Lemma A1, A(p) is a nonsingular M-matrix. In other words, we have to design Ai to meet Assumption 3 under condition (47). We can therefore conclude by using Theorem 1 that if we design Ai as described above, there is a positive scalar τ ∗such that the stochastic controlled hybrid system (43) is almost surely exponentially stable provided that τ ⩽τ ∗. 4.2 Observable in some modes Let us now consider the case where the state of the underlying system is observable only for some modes but not all. To describe this situation, let us divide the space S of the Markov chain into two proper subspaces S1 and S2 (namely S = S1 ∪S2 and S1 ∩S2 = ∅). Assume that the state x(t) is not observable when the system is in any mode i ∈S1, but is fully observable in any mode i ∈S2. Without any loss of generality, let us assume that S1 = {1, . . . , ¯N} and S2 = { ¯N + 1, . . . , N} for some 1 ⩽¯N < N. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:13 Let us now design our stochastic feedback control function u(x, i). Given that the system is not observable in any mode i ∈S1, it is reasonable to think it is not controllable in these modes so we must have u(x, i) = 0 for i ∈S1. For i ∈S2, we seek the linear control function u(x, i) = Aix as in the last subsection. Hence, the controlled system can still be described by the hybrid system (43) as long as we set Ai = 0 for i ∈S1. n Ai for i ∈S2, we impose an additional condition that To design Ai for i ∈S2, we impose an additional condition that for each i ∈S1, there is a j ∈S2 such that γij > 0. (47) (47) In layman’s terms, this condition means that given that the Markov chain is at state i ∈S1 at any time t, it could jump to a state j ∈S2 directly in very short time with a positive probability. Based on this condition, we can choose a pair of numbers p ∈(0, 2/3) and β ∈(0, 1) such that (1 −β) N X j= ¯ N+1 γij > pαi, ∀i ∈S1. (48) (48) We can then, for each i ∈S2, ﬁnd a nonnegative number δi such that βpδ2 i (2 −3p) 8 > (1 −β) ¯ N X j=1 γij + βpαi. (49) (49) Choose a matrix Di satisfying condition (44) and let Ai = δiDi. 5 Conclusion Inﬂuenced by Mao [34], we have discussed the almost sure stabilization of a given unstable hybrid dif- ferential equation ˙x(t) = f(x(t), r(t)) by nonlinear discrete-time stochastic feedback control u(x([t/τ]τ), Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:14 r([t/τ]τ))dB(t). We have shown that there is a positive number τ ∗such that the stochastically controlled system dx(t) = f(x(t), r(t))dt + u(x([t/τ]τ), r([t/τ]τ))dB(t) is almost surely stable provided that τ < τ∗ under the global Lipschitz condition plus the condition that guarantees the almost sure exponential sta- bility of the corresponding hybrid stochastic system dx(t) = f(x(t), r(t))dt + u(x(t), r(t))dB(t). As a special but important case, we have discussed in more detail how to design the linear feedback controls. Acknowledgements The authors would like to thank Leverhulme Trust (Grant No. RF-2015-385), Royal Society (Grant No. WM160014, Royal Society Wolfson Research Merit Award), Royal Society and Newton Fund (Grant No. NA160317, Royal Society–Newton Advanced Fellowship), Engineering and Physics Sciences Research Council (Grant No. EP/K503174/1), National Natural Science Foundation of China (Grant Nos. 61503190, 61473334, 61403207), Natural Science Foundation of Jiangsu Province (Grant Nos. BK20150927, BK20131000), and Ministry of Education (MOE) of China (Grant No. MS2014DHDX020) for their ﬁnancial support. The ﬁrst author would like to thank Chinese Scholarship Council for awarding him the scholarship to visit the University of Strathclyde for one year. Open access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and source are credited. References 1 Black F, Scholes M. The pricing of options and corporate liabilities. J Political Economy, 1973, 81: 637–654 2 Berman A, Plemmons R J. Nonnegative Matrices in the Mathematical Sciences. Philadelphia: Society for Industrial and Applied Mathematics, 1994 3 Krishnamurthy V, Wang X, Yin G. Spreading code optimization and adaptation in CDMA via discrete stochastic approximation. IEEE Trans Inform Theor, 2004, 50: 1927–1949 Luo Q, Gong Y Y, Jia C X. Stability of gene regulatory networks with L´evy noise. Sci China Inf Sci, 2017, 60: 0722 4 Luo Q, Gong Y Y, Jia C X. Stability of gene regulatory networks with Levy noise. Sci China Inf Sci, 2017, 60: 072204 5 Luo Q, Shi L L, Zhang Y T. Stochastic stabilization of genetic regulatory networks. Neurocomputing, 2017, 266: 123–127 5 Luo Q, Shi L L, Zhang Y T. Stochastic stabilization of genetic regulatory networks. Neurocomputing, 123–127 Mao W H, Deng F Q, Wan A H. Robust H2/H∞global linearization ﬁlter design for nonlinear stochastic time-varyi delay systems. Sci China Inf Sci, 2016, 59: 032204 y y , , 7 Mao X R. Stability of Stochastic Diﬀerential Equations With Respect to Semimartingales. London: Longman Scientiﬁc and Technical, 1991 8 Mao X R. Exponential Stability of Stochastic Diﬀerential Equations. New York: Marcel Dekker, 1994 9 Tang M, Meng Q. Stochastic evolution equations of jump type with random coeﬃcients: existence, un optimal control. Sci China Inf Sci, 2017, 60: 118202 10 Willsky A S, Levy B C. Stochastic stability research for complex power systems. DOE Contract LIDS MIT Report No. ET-76-C-01-2295, 1979 11 Sworder D D, Rogers R O. An LQ-solution to a control problem associated with a solar thermal central receiver. IEEE Trans Automat Contr, 1983, 28: 971–978 12 Yin G, Liu R H, Zhang Q. Recursive algorithms for stock liquidation: a stochastic optimization approach. SIAM J Optim, 2002, 13: 240–263 13 Zhang Q. Stock trading: an optimal selling rule. SIAM J Control Optim, 2001, 40: 64–87 14 Anderson W J. Continuous-Time Markov Chains. New York: Springer, 1991 15 Ji Y, Chizeck H J, Feng X, et al. Stability and control of discrete-time jump linear systems. Contr Theor Adv Tech, 1991, 7: 247–270 16 Mariton M. Jump Linear Systems in Automatic Control. New York: Marcel Dekker, 1990 17 Skorohod A V. Asymptotic Methods in the Theory of Stochastic Diﬀerential Equations. Providence: American Math- ematical Society, 1989 18 Yin G, Zhang Q. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:15 Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:15 2007 26 Appleby J A D, Mao X R. Stochastic stabilisation of functional diﬀerential equations. Syst Control Lett 1069–1081 27 Mao X R, Yuan C G. Stochastic Diﬀerential Equations With Markovian Switching. London: Imperial Col 2006 28 Mao X R, Yin G G, Yuan C G. Stabilization and destabilization of hybrid systems of stochastic diﬀerential equations. Automatica, 2007, 43: 264–273 Scheutzow M K R. Stabilization and destabilization by noise in the plane. Stochastic Anal Appl, 1993, 11: 97–113 30 Mao X R. Stabilization of continuous-time hybrid stochastic diﬀerential equations by discrete-time feedback control. Automatica, 2013, 49: 3677–3681 , , 31 Mao X R, Liu W, Hu L J, et al. Stabilization of hybrid stochastic diﬀerential equations by feedback control based on discrete-time state observations. Syst Control Lett, 2014, 73: 88–95 32 You S, Liu W, Lu J, et al. Stabilization of hybrid systems by feedback control based on discrete-time state observations. SIAM J Control Optim, 2015, 53: 905–925 Song G, Zheng B C, Luo Q, et al. Stabilisation of hybrid stochastic diﬀerential equations by feedback control bas on discrete-time observations of state and mode. IET Control Theor Appl, 2017, 11: 301–307 34 Mao X R. Almost sure exponential stabilization by discrete-time stochastic feedback control. IEEE Trans Automat Contr, 2016, 61: 1619–1624 1) Minkowski H. Diophantische Approximationen. Teubner, 1907. Appendix A Once Assumption 2 holds, the veriﬁcation of Assumption 3 depends very much on the choice of p ∈(0, 1). In this appendix, we give some easier qualiﬁcations that guarantee the existence of such a p and, hence, for Assumption 3 to hold. For this purpose, we need the theory of M-matrices (see [2,27]). For a vector or matrix A, A > 0 means that all elements of A are positive. Moreover, a square matrix A = [aij]N×N is called a Z-matrix if it has nonpositive oﬀ-diagonal entries, namely aij ⩽0 for all i ̸= j. Here, we cite a useful lemma on M-matrices. L A1 ([20]) If A [ ] i Z Lemma A1 ([20]). If A = [aij]N×N is a Z-matrix, the following statements are equivalent: (1) A is a nonsingular M-matrix. Lemma A1 ([20]). If A = [aij]N×N is a Z-matrix, the following statements are equivalent: (1) A is a nonsingular M-matrix. ( ) g (2) A is semi-positive; that is, there exists x > 0 in RN such that Ax > 0. (3) A−1 exists and its elements are all nonnegative. (4) All the leading principal minors of A are positive; that is (4) All the leading principal minors of A are positive; that is a11 · · · a1k ... ... ak1 · · · akk > 0, for every k = 1, 2, . . . , N. We also need another classical result. We also need another classical result. Lemma A2 (Minkowski1)) If a Z matrix A = [aij]N N has all positive row sums that is We also need another classical result. Lemma A2 (Minkowski1)). If a Z-matrix A = [aij]N×N has all positive row sums, that is Lemma A2 (Minkowski1)). If a Z-matrix A = [aij]N×N has all positive row sums, that is N X j=1 aij > 0, ∀i = 1, 2, . . . , N, N X j=1 aij > 0, ∀i = 1, 2, . . . , N, detA > 0. References Continuous-Time Markov Chains and Applications: a Singular Perturbation Approach. New York: Springer-Verlag, 1998 Shaikhet L. Stability of stochastic hereditary systems with Markov switching. Theory Stoch Proc, 1996, 2: 180–18 20 Mao X R. Stability of stochastic diﬀerential equations with Markovian switching. Stochastic Processes Their Appl, 1999, 79: 45–67 Arnold L, Crauel H, Wihstutz V. Stabilization of linear systems by noise. SIAM J Control Optim, 1983, 21: 451–4 21 Arnold L, Crauel H, Wihstutz V. Stabilization of linear systems by noise. SIAM J Control Optim, 1983, 21: 451–461 22 Basak G K, Bisi A, Ghosh M K. Stability of a random diﬀusion with linear drift. J Math Anal Appl, 1996, 202: 604–622 22 Basak G K, Bisi A, Ghosh M K. Stability of a random diﬀusion with linear drift. J Math Anal Appl, 1996, 202: 604–622 23 Khasminskii R Z. Stochastic Stability of Diﬀerential Equations. Leiden: Sijthoﬀand Noordhoﬀ, 1981 24 Mao X R. Stochastic stabilization and destabilization. Syst Control Lett, 1994, 23: 279–290 25 Mao X R. Stochastic Diﬀerential Equations and Their Applications. 2nd ed. Chichester: Horwood Publishin 25 Mao X R. Stochastic Diﬀerential Equations and Their Applications. 2nd ed. Chichester: Horwood Publishing Limited, Minkowski H. Diophantische Approximationen. Teubner, 1907. Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:15 detA > 0. detA > 0. Now, let us propose a condition that −(α1 + 0.5ρ2 1 −σ2 1) −γ12 · · · −γ1N −(α2 + 0.5ρ2 2 −σ2 2) −γ22 · · · −γ2N ... ... · · · ... −(αN + 0.5ρ2 N −σ2 N) −γN2 · · · −γNN > 0, (A1) (A1) where αi, ρi, σi are the constants speciﬁed in Assumption 2. This condition can be veriﬁed straightaway once Assumption 2 holds. It was shown in [28] that under an additional condition that It was shown in [28] that under an additional condition that for some u ∈S, γiu > 0, for all i ̸= u, (A2) for some u ∈S, γiu > 0, for all i ̸= u, (A2) condition (A1) is equivalent to the following simpler condition condition (A1) is equivalent to the following simpler condition condition (A1) is equivalent to the following simpler condition N X i=1 πi(αi + 0.5ρ2 i −σ2 i ) < 0, (A3) N X i=1 πi(αi + 0.5ρ2 i −σ2 i ) < 0, (A3) (A3) Song G F, et al. Sci China Inf Sci July 2018 Vol. 61 070213:16 where (π1, . . . , πN) are the stationary distribution of the Markov chain as deﬁned in Section 2. In this paper, we replace condition (A2) by a slightly weaker one that where (π1, . . . , πN) are the stationary distribution of the Markov chain as deﬁned in Section 2. In this paper, we replace condition (A2) by a slightly weaker one that tationary distribution of the Markov chain as deﬁned in Section 2. In this paper, we replace weaker one that for some u ∈S, γiu ∨(σ2 i −0.5ρ2 i −αi) > 0, for all i ̸= u. (A4) some u ∈S, γiu ∨(σ2 i −0.5ρ2 i −αi) > 0, for all i ̸= u. (A4) We do not yet know whether (A1) is equivalent to (A3) under this weaker condition. However, the following proposition is good enough for use in this paper. Proposition A1. If conditions (A1) and (A4) hold, Assumption 3 is satisﬁed. f h l f l h h d Proposition A1. If conditions (A1) and (A4) hold, Assumption 3 is satisﬁed. Proof. Without loss of generality, we may assume that the state u = N in condition (A4), namely p ( ) ( ) , p 3 Proof. detA > 0. Without loss of generality, we may assume that the state u = N in condition (A4), namely γiN ∨(σ2 i −0.5ρ2 i −αi) > 0, for all 1 ⩽i ⩽N −1. (A5) (A5) If not, by switching state u with N, we need to reorder the states of the Markov chain r(t) that is, rename state u as N and N as u. Consequently, the determinant in the left-hand side of (A1) will switch the uth row with the Nth row and then switch the uth column with the Nth column, but these do not change the value of the determinant, namely the determinant remains positive. Moreover, given a nonsingular M-matrix, it is easy to show that the new matrix remains a nonsingular M-matrix by switching the uth column with the Nth column and then switching the uth row with the Nth row. y g g By [27, Lemma 5.2], the derivative dA(0)/dp is equal to the determinant on the left-hand side of (A1), whence dA(0)/dp > 0. It is also easy to see A(0) = 0. Consequently, for all p ∈(0, 1) suﬃciently small, we have detA(p) > 0. (A6) detA(p) > 0. (A6) n the other hand, for each i = 1, 2, . . . , N −1, either γiN > 0 or γiN = 0. In the case when γiN > 0, we clearly ha , , , , , γiN γiN γiN , y θi(p) > −γiN, for all suﬃciently small p ∈(0, 1); whereas in the case when γiN = 0, condition (A5) implies σ2 i −0.5ρ2 i −αi > 0 whence θi(p) > −γiN, for all suﬃciently small p ∈(0, 1); ( ) γ ( ) whereas in the case when γiN = 0, condition (A5) implies σ2 i −0.5ρ2 i −αi > 0 whence θi(p) > 0 = −γiN, for all suﬃciently small p ∈(0, 1). θi(p) > 0 = −γiN, for all suﬃciently small p ∈(0, 1). In other words, we always have θi(p) > −γiN, i = 1, 2, . . . , N −1, (A7) θi(p) > −γiN, i = 1, 2, . . . , N −1, (A7) for all p ∈(0, 1) suﬃciently small. Fix any p ∈(0, 1) suﬃciently small for both (A6) and (A7) to hold. detA > 0. Consider the deriving principal sub-matrix, Ak(p) :=   θ1(p) −γ11 −γ12 · · · −γ1k −γ21 θ1(p) −γ22 · · · −γ2k ... ... · · · ... −γk1 −γk2 · · · θk(p) −γkk   of A(p), for k = 1, 2, . . . , N −1. Obviously, Ak(p) is a Z-matrix. Moreover, for every i = 1, 2, . . . , k, the ith row of this sub-matrix has its sum k N θi(p) − k X j=1 γij = θi(p) + N X j=k+1 γij ⩾θi(p) + γiN > 0 by (A7). By Lemma A2, detAk(p) > 0. Thus, we should point out that all the deriving principal minors of A(p) are positive. By applying Lemma A1, we can obtain that A(p) is a nonsingular M-matrix. This completes the proof.
https://openalex.org/W4390759634	https://journal.ilmudata.co.id/index.php/ijmst/article/download/232/65	Indonesian	null	Analisis Pengaruh Pertumbuhan Angkatan Kerja, Inflasi dan Suku Bunga terhadap Jumlah UMKM	Indonesian Journal of Multidisciplinary on Social and Technology	2,023	cc-by	4,086	Analisis Pengaruh Pertumbuhan Angkatan Kerja, Inflasi dan Suku Bunga terhadap Jumlah UMKM Mutia Alius1, Alima Shofia 2, Hadigufri Triha 3, Trinda Farhan Satria 4, Beni Harma 5, Jefri Rahmad Mulia 6 1,2,3,4 Program Studi Teknik Industri, Universitas Adzkia, Padang, Indonesia 5 Program Studi Teknik Industri, Fakultas Teknik, Universitas Putra Indonesia “YPTK”, Padang, Indonesia 6 Program Studi Informatika, Universitas Adzkia, Padang, Indonesia 1mutiaalius.ti@adzkia.ac.id, 2alimashofia_ti@adzkia.ac.id, 3hadigufri@adzkia.ac.id, 4trindafarhansatria@adzkia.ac.id, 5beniharma@upiyptk.ac.id, 6jefrirm.if@adzkia.ac.id Abstrak Peran UMKM dalam pertumbuhan ekonomi di Indonesia sangatlah penting karena banyaknya tenaga kerja yang terserap pada sektor ini. Salah satu cara pemerintah untuk menghidupkan kembali laju pertumbuhan ekonomi Indonesia adalah dengan menekan laju inflasi dan menurunkan suku bunga. Jika dilihat dari hasil penelitian yang ada, i . Penelitian lainnya . Penelitian ini ingin melihat pengaruh dari Pertumbuhan Angkatan Kerja, Inflasi dan Suku Bunga terhadap Jumlah UMKM dengan metode regresi linier berganda dimana analisis regresi linear berganda dilakukan untuk mengetahui arah dan seberapa besar pengaruh variabel independent (bebas) terhadap variabel dependen (terikat). Jumlah Kata kunci: UMKM, regresi linier berganda, angkatan kerja, inflasi, suku bunga. Indonesian Journal of Multidisciplinary on Social and Technology Homepage: https://journal.ilmudata.co.id/index.php/ijmst Vol. 1 No. 3 (2023) pp: 290-296 P-ISSN: 2986-6790, e-ISSN: 2986-6782 Indonesian Journal of Multidisciplinary on Social and Technology Homepage: https://journal.ilmudata.co.id/index.php/ijmst Vol. 1 No. 3 (2023) pp: 290-296 P-ISSN: 2986-6790, e-ISSN: 2986-6782 Analisis Pengaruh Pertumbuhan Angkatan Kerja, Inflasi dan Suku Bunga terhadap Jumlah UMKM 290 1. Pendahuluan ; (3) (Triana, 2022) : (1) ; (2) 290 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 (Malentang et al., 2022). . T mengurangi Studi lain menyimpulkan bahwa untuk setiap kenaikan suku bunga 1%, permintaan menurun sebesar 35,36% (Effendi, 2016) Berdasarkan Badan Pusat Statistik (BPS), tenaga kerja merupakan penduduk dalam usia kerja (15 tahun ke atas) yang berpotensi memproduksi barang dan jasa. BPS membagi tenaga kerja sebagai berikut: (1) Tenaga kerja penuh (full employed), yaitu tenaga kerja yang mempunyai jumlah jam kerja lebih dari 35 jam dalam seminggu dengan hasil kerja tertentu sesuai dengan uraian tugas. (2) Tenaga kerja tidak penuh atau setengah pengangguran (under employed), yaitu tenaga kerja dengan jam kerja kuran dari 35 jam dalam seminggu; dan (3) Tenaga kerja yang belum bekerja atau sementara tidak bekerja (unemployed), yaitu tenaga kerja dengan jam kerja 0 > 1 jam perminggu. Sementara itu, angkatan kerja (labor force) merupakan bagian dari penduduk yang mampu dan bersedia melakukan pekerjaan (Eliza, 2015). Angkatan kerja menurut International Labor Organization (ILO) merupakan penduduk yang berumur > 10 tahun yang mampu terlibat dalam proses produksi, dimana mere aktif dalam kegiatan yang menghasilkan barang atau jasa, atau mereka yang selama seminggu melakukan pekerjaan dengan maksud untuk memperoleh penghasilan (Tasyim et al., 2021). Sejalan dengan hal tersebut, dalam penyerapan tenaga kerja, lapangan pekerjaan atau unit usaha akan menampung semua tenaga kerja sesuai dengan jumlah tenaga kerja yang ada, di mana lapangan kerja merupakan kegiatan usaha atau instansi tempat seseorang sedang bekerja atau pernah bekerja. Maka dapat diasumsikan bahwa seiring dengan bertambahnya angkatan kerja, maka UMKM sebagai lapangan kerja akan terus bertambah. Dalam Teori Keynes, inflasi terjadi disebabkan oleh gaya hidup masyarakat yang melampaui batas kemampuan ekonominya shingga menyebabkan pengeluaran agregat terlalu besar (Ramadani et al., 2020). Inflasi merupakan masalah serius bagi kestabilan perekonomian suatu negara, dimana akan menyebabkan harga-harga barang menjadi naik, turunnya nilai mata uang, meningkatnya angka pengangguran, dan menurunnya tingkat kesejahteraan masyarakat (Nova, 2022; Rosa & Idwar, 2019). Kenaikan inflasi membuat biaya keuangan dalam UMKM semakin tinggi, sehingga dapat berpengaruh pada kelangsungan usaha. 1. Pendahuluan Hal ini menunjukkan bahwa inflasi memiliki pengaruh yang kuat terhadap DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) 291 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 pertumbuhan UMKM (Romadhoni et al., 2020). Penelitian lainnya juga melihat adanya hubungan inflasi dan Angkatan kerja dengan tingkat pengangguran (Rizal, 2021). pertumbuhan UMKM (Romadhoni et al., 2020). Penelitian lainnya juga melihat adanya hubungan inflasi dan Angkatan kerja dengan tingkat pengangguran (Rizal, 2021). pertumbuhan UMKM (Romadhoni et al., 2020). Penelitian lainnya juga melihat adanya hubungan inflasi dan Angkatan kerja dengan tingkat pengangguran (Rizal, 2021). 2. Metode Penelitian Perhitungan X1Y, X2Y, dan X3Y Tahun X1Y X2Y X3Y 2014 27.088.768.996 8.813 5.586 2015 30.019.169.916 647 5.729 2016 31.481.730.508 3.927 4.694 2017 34.274.715.672 1.713 3.704 2018 4.599.718.537 260 521 2019 5.137.525.520 195 639 2020 5.587.920.349 249 498 2021 14.746.907.448 147 1.081 Jumlah 152.936.456.946 15.953 22.452 Gambar 2. Bagan Penelitian DOI: https://doi.org/10.31004/ijmst.v1i3.232 p g j Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) 2. Metode Penelitian Berdasarkan pendahuluan di atas, maka dapat digambarkan kerangka pemikiran pada Gambar 1 sebagai berikut: (1) ; (2) ; (3) ; (4) . Gambar 1. Kerangka Pemikiran Jumlah Angkatan Kerja Inflasi Suku bunga Jumlah UMKM di Kota Padang Gambar 1. Kerangka Pemikiran Jumlah Angkatan Kerja Inflasi Suku bunga Jumlah UMKM di Kota Padang Jumlah Angkatan Kerja Jumlah UMKM di Kota Padang . Gambar 1. Kerangka Pemikiran Penelitian ini menggunakan metode regresi linier berganda. Regresi linear berganda merupakan model regresi yang melibatkan lebih dari satu variabel independent (bebas). Analisis regresi linear berganda dilakukan untuk mengetahui arah dan seberapa besar pengaruh variabel independent (bebas) terhadap variabel dependen (terikat). Bentuk umum dari persamaan regresi linier berganda sebagai berikut: Penelitian ini menggunakan metode regresi linier berganda. Regresi linear berganda merupakan model regresi yang melibatkan lebih dari satu variabel independent (bebas). Analisis regresi linear berganda dilakukan untuk mengetahui arah dan seberapa besar pengaruh variabel independent (bebas) terhadap variabel dependen (terikat). Bentuk umum dari persamaan regresi linier berganda sebagai berikut: P lain (Rosa & Idwar, 2019) dkk pada tahun (Triha et al., 2023) Y = b0 + b1X1 + b2X2 + b3X3 + … + ℇ (1) (1) Y = b0 + b1X1 + b2X2 + b3X3 + … + ℇ Y = b0 + b1X1 + b2X2 + b3X3 + … + ℇ Pada penelitian ini ada 3 variabel bebas yaitu: (1) Pertumbuhan Angkatan Kerja (X1); (2) Inflasi di Kota Padang (X2); dan (3) Suku Bunga (X3). Jumlah UMKM (Y) adalah variabel terikat. Penelitian ini dilakukan di Kota Padang dimana data yang diolah adalah data sekunder yang bersumber dari BPS Kota Padang dan Bank Indonesia. Metodologi penelitian ini dijelaskan dalam bagan yang dapat dilihat pada Gambar 2 sebagai berikut: DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) 292 (Triha et al., 2023) M Mulai Studi Pendahuluan Studi Literatur yang berhubungan dengan UMKM, Angkatan kerja, inflasi dan suku bunga A Mulai Studi Pendahuluan Studi Literatur yang berhubungan dengan UMKM, Angkatan kerja, inflasi dan suku bunga Mulai M Studi Pendahuluan Studi Literatur yang berhubungan dengan UMKM, Angkatan kerja, inflasi dan suku bunga DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) 292 292 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. DOI: https://doi.org/10.31004/ijmst.v1i3.232 2. Metode Penelitian 3, 2023 tia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 Selanjutnya data pertumbuhan angkatan kerja Kota Padang, inflasi Kota Padang dan suku bunga (BI) 2014- 2021 dapat dilihat pada Tabel 2. Gambar 2. Bagan Penelitian A Pengumpulan Data Data yang digunakan adalah data sekunder yang diambil dari website BPS Kota Padang dan Bank Indonesia tahun 2014 – 2021. Jumlah data yang digunakan sebanyak 8 periode. Pengolahan Data Pengolahan data menggunakan Ms. Excel dan analisis statistik Uji R Square, Uji T dan Uji f menggunakan analysis toolpak - regression excel Analisis Analisis yang dilakukan adalah melihat hubungan antara Angkatan kerja, inflasi dan tingkat suku bunga terhadap jumlah UMKM Penutup Kesimpulan Selesai Tabel 2. Pertumbuhan Angkatan Kerja Kota Padang, Inflasi Kota Padang dan Suku Bunga (BI) 2014-2021 Tahun Jumlah Angkatan Kerja (X1) Inflasi (X2) Suku Bunga (X3) 2014 365.758 11,90% 7,54% 2015 394.092 0,85% 7,52% 2016 402.430 5,02% 6,00% 2017 422.196 2,11% 4,56% 2018 450.467 2,55% 5,10% 2019 452.048 1,72% 5,63% 2020 476.663 2,12% 4,25% 2021 480.324 0,48% 3,52% Jumlah 3.443.978 27% 44% Langkah berikutnya adalah melakukan perhitungan untuk menemukan nilai-nilai yang dibutuhkan dalam matrik penolong seperti dibawah ini. A = [ 𝑛 ∑𝑋1 ∑𝑋2 ∑𝑋3 ∑𝑋1 ∑𝑋1 2 ∑𝑋1𝑋2 ∑𝑋1𝑋3 ∑𝑋2 ∑𝑋3 ∑𝑋1𝑋2 ∑𝑋1 𝑋3 ∑𝑋2 2 ∑𝑋2𝑋3 ∑𝑋2𝑋3 ∑𝑋3 2 ] H = [ ∑𝑌 ∑𝑋1𝑌 ∑𝑋2𝑌 ∑𝑋3𝑌] Selanjutnya dilakukan perhitungan untuk menghitung X1Y, X2Y, dan X3Y yang dapat dilihat pada Tabel 3. Selanjutnya dilakukan perhitungan untuk menghitung X1Y, X2Y, dan X3Y yang dapat dilihat pada Tabel 3. 0.31004/ijmst.v1i3.232 ution 4.0 International (CC BY 4.0) 93 Tabel 3. Perhitungan X1Y, X2Y, dan X3Y Tahun X1Y X2Y X3Y 2014 27.088.768.996 8.813 5.586 2015 30.019.169.916 647 5.729 2016 31.481.730.508 3.927 4.694 2017 34.274.715.672 1.713 3.704 2018 4.599.718.537 260 521 2019 5.137.525.520 195 639 2020 5.587.920.349 249 498 2021 14.746.907.448 147 1.081 Jumlah 152.936.456.946 15.953 22.452 Selanjutnya dilakukan perhitungan untuk menghitung X1X2, X1X3, dan X2X3 yang dapat dilihat pada Tabel 4. Tabel 4. Perhitungan X1X2, X1X3, dan X2X3 Tahun X1X2 X1X3 X2X3 2014 43525 27584,24917 0,008975 2015 3350 29639,0025 0,000639 2016 20202 24145,82611 0,003012 2017 8908 19262,6925 0,000963 Tabel 3. Perhitungan X1Y, X2Y, dan X3Y Tabel 3. 3. Hasil dan Pembahasan dan dapat dilihat pada Tabel 1 DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) 293 dan dapat dilihat pada Tabel 1 Tabel 1. Jumlah UMKM Kota Padang 2014-2021 Tahun Jumlah UMKM (Y) 2014 74.062 2015 76.173 2016 78.229 2017 81.182 2018 10.211 2019 11.365 2020 11.723 2021 30.702 Jumlah 373.647 2019 5.137.525.520 195 639 2020 5.587.920.349 249 498 2021 14.746.907.448 147 1.081 Jumlah 152.936.456.946 15.953 22.452 Selanjutnya dilakukan perhitungan untuk menghitung X1X2, X1X3, dan X2X3 yang dapat dilihat pada Tabel 4. Tabel 4. Perhitungan X1X2, X1X3, dan X2X3 Tahun X1X2 X1X3 X2X3 2014 43525 27584,24917 0,008975 2015 3350 29639,0025 0,000639 2016 20202 24145,82611 0,003012 2017 8908 19262,6925 0,000963 Tabel 1. Jumlah UMKM Kota Padang 2014-2021 Tabel 1. Jumlah UMKM Kota Padang 2014-2021 Tahun Jumlah UMKM (Y) 2014 74.062 2015 76.173 2016 78.229 2017 81.182 2018 10.211 2019 11.365 2020 11.723 2021 30.702 Jumlah 373.647 Tabel 1. Jumlah UMKM Kota Padang 2014-2021 Selanjutnya dilakukan perhitungan untuk menghitung X1X2, X1X3, dan X2X3 yang dapat dilihat pada Tabel 4. Tabel 4. Perhitungan X1X2, X1X3, dan X2X3 Tabel 4. Perhitungan X1X2, X1X3, dan X2X3 Tahun X1X2 X1X3 X2X3 2014 43525 27584,24917 0,008975 2015 3350 29639,0025 0,000639 2016 20202 24145,82611 0,003012 2017 8908 19262,6925 0,000963 293 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia g Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 p y gy 2018 11487 22992,58646 0,001302 2019 7775 25427,7 0,000968 2020 10105 20258,1775 0,000901 2021 2306 16911,4075 0,000169 Jumlah 107.658 186.222 0,016928 Berikut adalah Matrik A0 A0 = [ 373.647 3.443.978 0,27 0,44 152.936.456.946 1.494.473.809.485 107.658 186.222 15.953 22.452 107.658 186.222 0,01862 0,016928 0,016928 0,02584 ] dengan det (A0) = 131524965655 Selanjutnya adalah Matrik A1 Berikut adalah Matrik A0 A0 = [ 373.647 3.443.978 0,27 0,44 152.936.456.946 1.494.473.809.485 107.658 186.222 15.953 22.452 107.658 186.222 0,01862 0,016928 0,016928 0,02584 ] dengan det (A0) = 131524965655 Selanjutnya adalah Matrik A1 A1 = [ 8 373.647 0,27 0,44 3.443.978 152.936.456.946 107.658 186.222 0,27 0,44 15.953 22.452 0,01862 0,016928 0,016928 0,02584 ] dengan det (A1) = -242087,4701 Selanjutnya adalah Matrik A1 Selanjutnya dilakukan perhitungan untuk menghitung X12, X22, dan X32 yang dapat dilihat pada Tabel 5. Selanjutnya dilakukan perhitungan untuk menghitung X12, X22, dan X32 yang dapat dilihat pada Tabel 5. A1 = [ 8 373.647 0,27 0,44 3.443.978 152.936.456.946 107.658 186.222 0,27 0,44 15.953 22.452 0,01862 0,016928 0,016928 0,02584 ] Tabel 5. Perhitungan X1 2, X2 2, dan X3 2 Tahun X1 2 X2 2 X3 2 2014 133.778.914.564 0,0142 0,0057 2015 155.308.504.464 0,0001 0,0057 2016 161.950.255.103 0,0025 0,0036 2017 178.249.462.416 0,0004 0,0021 2018 202.920.518.089 0,0007 0,0026 2019 204.347.394.304 0,0003 0,0032 2020 227.207.615.569 0,0004 0,0018 2021 230.711.144.976 0,0000 0,0012 Jumlah 1.494.473.809.485 0,01862 0,02584 Tabel 5. Tabel 1. Jumlah UMKM Kota Padang 2014-2021 Perhitungan X1 2, X2 2, dan X3 2 dengan det (A1) = -242087,4701 Selanjutnya adalah Matrik A2 A2 = [ 8 3.443.978 373.647 0,44 3.443.978 1.494.473.809.485 152.936.456.946 186.222 0,27 0,44 107.658 186.222 15.953 22.452 0,016928 0,02584 ] dengan det (A2) = -59725515451 Selanjutnya adalah Matrik A3 A3 = [ 8 3.443.978 0,27 373.647 3.443.978 1.494.473.809.485 107.658 152.936.456.946 0,27 0,44 107.658 186.222 0,01862 0,016928 15.953 22.452 ] dengan det (A3) = -332982875900 Setelah didapatkan variabel-variabel yang dibutuhkan untuk mengisi matrik penolong, maka selanjutnya variabel-variabel tersebut dipindahkan ke dalam matrik A dan matrik H untuk mendapatkan persamaan regresi berganda. Setelah didapatkan semua determinan matrik, maka dilanjutkan ke tahap penyusunan persamaan regresi linier berganda dengan format Y = b0 + b1X1 + b2X2 + b3X3. Untuk mendapatkan nilai dari b0, b1, b2, dan b3 dilakuan perhitungan sebagai berikut: Susunan Matrik A dan Matrik H A = [ 8 3.443.978 0,27 0,44 3.443.978 1.494.473.809.485 107.658 186.222 0,27 0,44 107.658 186.222 0,01862 0,016928 0,016928 0,02584 ] dengan det A = 148670,8833 H = [ 373.647 152.936.456.946 15 953 ] Susunan Matrik A dan Matrik H Susunan Matrik A dan Matrik H b0 = Det (A0) / Det (A) = 131524965655 / 148670,8833 = 884672,0 b1 = Det (A1) / Det (A) = -242087,4701 / 148670,8833 = -1,6 b2 = Det (A2) / Det (A) = -59725515451 / 148670,8833 = -401729,7 b3 = Det (A3) / Det (A) = -332982875900 / 148670,8833 = -2239731,6 b0 = Det (A0) / Det (A) = 131524965655 / 148670,8833 = 884672,0 b1 = Det (A1) / Det (A) = -242087,4701 / 148670,8833 = -1,6 b2 = Det (A2) / Det (A) = -59725515451 / 148670,8833 = -401729,7 b3 = Det (A3) / Det (A) = -332982875900 / 148670,8833 = -2239731,6 b0 = Det (A0) / Det (A) = 131524965655 / 148670,8833 = 884672,0 b1 = Det (A1) / Det (A) = -242087,4701 / 148670,8833 = -1,6 H = [ 373.647 152.936.456.946 15.953 22.452 ] H = [ 373.647 152.936.456.946 15.953 22.452 ] b2 = Det (A2) / Det (A) = -59725515451 / 148670,8833 = -401729,7 b3 = Det (A3) / Det (A) = -332982875900 / 148670,8833 = -2239731,6 Langkah selanjutnya adalah menyusun matrik A0, A1, A2, dan A3 berdasarkan matrik A dan matrik H dengan format pada Gambar 3. Gambar 3. DOI: https://doi.org/10.31004/ijmst.v1i3.232 Tabel 1. Jumlah UMKM Kota Padang 2014-2021 Format Matrik A0, A1, A2, dan A3 Maka didapatkan persamaan regresi linier berganda sebagai berikut: Y = b0 + b1X1 + b2X2 + b3X3 = 884672 + (-1,6) X1 + (-401729,7) X2 + (-2239731,6) X3 = 884672 - 1,6 X1 - 401729,7 X2 - 2239731,6 X3 = 884672 - 1,6 X1 - 401729,7 X2 - 2239731,6 X3 Setelah didapatkan persamaan diatas, maka dilakukan analisi regresi linier berganda untuk melihat pengaruh dari masing-masing variabel bebas terhadap variabel terikat dimana uji R2, Uji t dan Uji f dilakukan dengan tingkat kepercayaan 95%. Hasil analisis statistik tersebut dapat dilihat pada Tabel 6, Tabel 7, dan Tabel 8 berikut: Gambar 3. Format Matrik A0, A1, A2, dan A3 DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) 294 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 Mutia Alius, Alima Shofia, Hadigufri Triha, Trinda Farhan Satria, Beni Harma , Jefri Rahmad Mulia Indonesian Journal of Multidisciplinary on Social and Technology Vol. 1 No. 3, 2023 Tabel 6. Uji R2 Regression Statistics Multiple R 0,96872595 R Square 0,938429965 Adjusted R Square 0,892252439 Standard Error 10999,15458 Observations 8 simultan atau bersama-sama, Angkatan Kerja (X1), Inflasi (X2) dan Suku Bunga (X3) berpengaruh signifikan terhadap Jumlah UMKM (Y) dimana nilai R square = 93,84%. Sementara sisanya 6% dipengaruhi oleh variabel diluar dari yang diteliti. Jika dilihat secara terpisah, Angkatan Kerja (X1) dan Suku Bunga (X3) secara signifikan memberikan pengaruh terhadap Jumlah UMKM dimana P-value untuk Angkatan Kerja 0,0029 < 0,05 dan P-value untuk Suku Bungan 0,022 < 0,05. Sedangkan variabel Inflasi (X2) cukup memberikan pengaruh yang signifian terhadap Jumlah UMKM karena nilai dari P-value sedikit diatas 0,05 yaitu 0,06. simultan atau bersama-sama, Angkatan Kerja (X1), Inflasi (X2) dan Suku Bunga (X3) berpengaruh signifikan terhadap Jumlah UMKM (Y) dimana nilai R square = 93,84%. Sementara sisanya 6% dipengaruhi oleh variabel diluar dari yang diteliti. Jika dilihat secara terpisah, Angkatan Kerja (X1) dan Suku Bunga (X3) secara signifikan memberikan pengaruh terhadap Jumlah UMKM dimana P-value untuk Angkatan Kerja 0,0029 < 0,05 dan P-value untuk Suku Bungan 0,022 < 0,05. Sedangkan variabel Inflasi (X2) cukup memberikan pengaruh yang signifian terhadap Jumlah UMKM karena nilai dari P-value sedikit diatas 0,05 yaitu 0,06. Perhatikan nilai R Square yang ada pada Tabel 6. Effendi, I. (2016). ANALISIS PENGARUH SUKU BUNGA TERHADAP KREDIT USAHA MIKRO KECIL DAN MENENGAH (UMKM) PADA BANK PEMBANGUNAN DAERAH (BPD) DI PROVINSI JAMBI. Jurnal Ilmiah Universitas Batanghari Jambi. Effendi, I. (2016). ANALISIS PENGARUH SUKU BUNGA TERHADAP KREDIT USAHA MIKRO KECIL DAN MENENGAH (UMKM) PADA BANK PEMBANGUNAN DAERAH (BPD) DI PROVINSI JAMBI. Jurnal Ilmiah Universitas Batanghari Jambi. Tabel 7. Uji Signifikansi f Eliza, Y. (2015). Pengaruh Investasi, Angkatan Kerja Dan Pengeluaran Pemerintah Terhadap Pertumbuhan Ekonomi Di Provinsi Sumatera Barat. Jurnal Pekbis, 7(3), 200–210. Halim, A. (2020). Pengaruh Pertumbuhan Usaha Mikro, Kecil Dan Menengah Terhadap Pertumbuhan Ekonomi Kabupaten Mamuju. Jurnal Ilmiah Ekonomi Pembangunan, 1(2), 157– 172. https://stiemmamuju.e-journal.id/GJIEP/article/view/39 Berdasarkan uji signifikansi f yang dilakukan, didapatkan kesimpulan bahwa secara simultan atau bersama-sama, Angkatan Kerja (X1), Inflasi (X2) dan Suku Bunga (X3) berpengaruh signifikan terhadap Jumlah UMKM (Y). Hal ini dibuktikan dengan nilai sign-f yang diperoleh lebih kecil dari 5% atau (0,0069 < 0,05). Jihad, N. M. (2014). Analisis Pengaruh Suku Bunga Dan Pemberian Kredit Terhadap Unit , Tenaga Kerja Dan Omzet Umkm. Khotijah, N. Z., Suharti, T., & Yudhawati, D. (2020). PENGARUH TINGKAT SUKU BUNGA DAN INFLASI TERHADAP PROFITABILITAS. Manager : Jurnal Ilmu Manajemen. https://doi.org/10.32832/manager.v3i1.3831 Tabel 8. Uji t Coefficients t Stat P-value Intercept 884671,99 6,284232338 0,003274665 Angkatan Kerja (X1) -1,63 -6,44165563 0,002988269 Inflasi (X2) -401729,741 -2,534273422 0,064371445 Suku Bunga (X3) -2239731,603 -3,633157288 0,022098026 Tabel 8. Uji t Tabel 8. Uji t Malentang, E. Y., Walewangko, E. N., & Siwu, H. F. D. (2022). Pengaruh pengangguran dan pendidikan terhadap tingkat kemiskinan di Kota Manado. Jurnal Berkala Ilmiah Efisiensi. Nova, N. (2022). ANALISIS HUBUNGAN KEBIJAKAN MONETER DENGAN INFLASI DI INDONESIA. Jurnal Ekonomika. https://doi.org/10.51179/eko.v17i1.1073 Purba, W., Nainggolan, P., & Panjaitan, P. D. (2022). Analisis Pengaruh Inflasi dan Pertumbuhan Ekonomi Terhadap Pengangguran di Provinsi Sumatera Utara. Jurnal Ekuilnomi. https://doi.org/10.36985/ekuilnomi.v4i1.336 Berdasarkan uji- t, dapat diketahui bahwa variabel Angkatan Kerja (X1) dan Suku Bunga (X3) secara signifikan memberikan pengaruh terhadap Jumlah UMKM. Hal ini terlihat dari nilai P-value untuk variabel Angkatan Kerja dan Suku Bunga kecil da ri 5%. Sedangkan variabel Inflasi (X2) cukup memberikan pengaruh yang signifian terhadap Jumlah UMKM karena nilai dari P-value sedikit diatas 0,05 yaitu 0,06. Ramadani, A. T., Junaidi, & Zulfa Eliza. (2020). Pengaruh Pertumbuhan Umkm (Usaha Mikro,Kecil Dan Menengah), Inflasi Dan Tingkat Pengangangguran Terhadap Pertumbuhan Ekonomi Di Indonesia. Jurnal Investasi Islam, 5, 153–173. Rizal, M. (2021). PENGARUH INFLASI DAN ANGKATAN KERJA TERHADAP TINGKAT PENGANGGURAN DI KOTA MAKASSAR. EKONOMI. Romadhoni, D., Amril, A., & Emilia, E. (2020). Analisis pengaruh pertumbuhan ekonomi, inflasi dan suku bunga terhadap pertumbuhan UMKM di Provinsi Jambi. E-Journal Perdagangan Industri Dan Moneter, 8(3), 127–134. https://doi.org/10.22437/pim.v8i3.13743 Tabel 1. Jumlah UMKM Kota Padang 2014-2021 Berdasarkan nilai yang ada, dapat dilihat bahwa Angkatan Kerja (X1), Inflasi (X2) dan Suku Bunga (X3) mempengaruhi Jumlah UMKM (Y) sebesar 93,84%. Sementara sisanya 6% dipengaruhi oleh variabel diluar dari yang diteliti. Reference (20 success factor of lean six sigma im Tabel 7. Uji Signifikansi f ANOVA df F Significance F Regression 3 20,32222266 0,006960275 Residual 4 Total 7 Reference DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) Tabel 7. Uji Signifikansi f ANOVA df F Significance F Regression 3 20,32222266 0,006960275 Residual 4 Total 7 Berdasarkan uji signifikansi f yang dilakukan, didapatkan kesimpulan bahwa secara simultan atau bersama-sama, Angkatan Kerja (X1), Inflasi (X2) dan Suku Bunga (X3) berpengaruh signifikan terhadap Jumlah UMKM (Y). Hal ini dibuktikan dengan nilai sign-f yang diperoleh lebih kecil dari 5% atau (0,0069 < 0,05). Tabel 8. Uji t Coefficients t Stat P-value Intercept 884671,99 6,284232338 0,003274665 Angkatan Kerja (X1) -1,63 -6,44165563 0,002988269 Inflasi (X2) -401729,741 -2,534273422 0,064371445 Suku Bunga (X3) -2239731,603 -3,633157288 0,022098026 Berdasarkan uji- t, dapat diketahui bahwa variabel Angkatan Kerja (X1) dan Suku Bunga (X3) secara signifikan memberikan pengaruh terhadap Jumlah UMKM. Hal ini terlihat dari nilai P-value untuk variabel Angkatan Kerja dan Suku Bunga kecil da ri 5%. Sedangkan variabel Inflasi (X2) cukup memberikan pengaruh yang signifian terhadap Jumlah UMKM karena nilai dari P-value sedikit diatas 0,05 yaitu 0,06. 4. Kesimpulan Penelitian ini berhasil mendapatkan persamaan regresi liner berganda untuk menentukan Jumlah UMKM (Y) berdasarkan Angkatan Kerja (X1), Inflasi (X2) dan Suku Bunga (X3) yaitu Y = 884672 - 1,6X1 - 401729,7X2 - 2239731,6X3. Dari hasil analisis regresi yang dilakukan, dapat disimpulkan bahwa secara Effendi, I. (2016). ANALISIS PENGA TERHADAP KREDIT USAHA MENENGAH (UMKM) PADA BA DAERAH (BPD) DI PROVINSI Universitas Batanghari Jambi. Eliza, Y. (2015). Pengaruh Investasi, Pengeluaran Pemerintah Terhadap P Provinsi Sumatera Barat. Jurnal P Halim, A. (2020). Pengaruh Pertumbuhan Menengah Terhadap Pertumbuha Mamuju. Jurnal Ilmiah Ekonomi P 172. https://stiemmamuju.e-journal Jihad, N. M. (2014). Analisis Pengaruh Su Kredit Terhadap Unit , Tenaga Ke Khotijah, N. Z., Suharti, T., & Yudhawati TINGKAT SUKU BUNGA DAN PROFITABILITAS. Manager : J https://doi.org/10.32832/manager Malentang, E. Y., Walewangko, E. N., & Pengaruh pengangguran dan pend kemiskinan di Kota Manado. Jurna Nova, N. (2022). ANALISIS HUBU MONETER DENGAN INFLASI D Ekonomika. https://doi.org/10.51 Purba, W., Nainggolan, P., & Panjaitan Pengaruh Inflasi dan Pertumbu Pengangguran di Provinsi Sumatera https://doi.org/10.36985/ekuilnom Ramadani, A. T., Junaidi, & Zulfa E Pertumbuhan Umkm (Usaha Mikr Inflasi Dan Tingkat Penganganggur Ekonomi Di Indonesia. Jurnal Inve Rizal, M. (2021). PENGARUH INFLA KERJA TERHADAP TINGKAT KOTA MAKASSAR. EKONOMI. Romadhoni, D., Amril, A., & Emilia, E. ( pertumbuhan ekonomi, inflasi da pertumbuhan UMKM di Prov Perdagangan Industri Dan Mo https://doi.org/10.22437/pim.v8i3 Rosa, Y. Del, & Idwar. (2019). PENGARU BUNGA TERHADAP OMSET UM Menara Ekonomi. Shofia, A., Bakhtiar, A., & Prastawa, H. Wicaksono, D. A. I., & Kramadibrata, B. S. (2022). Analisis Pengaruh Inflasi Dan Tingkat Suku Bunga Terhadap Kinerja Umkm Binaan Komunitas Jakpreneur Cipayung. Journal Ilmu Sosial, Manajemen, Dan Akuntansi (JISMA), 1(4), 487–500. https://melatijournal.com/index.php/jisma/article/view/186 improvement of business performance on low-cost hotel industry: A literature review. AIP Conference Proceedings. https://doi.org/10.1063/5.0000707 DOI: https://doi.org/10.31004/ijmst.v1i3.232 Lisensi: Creative Commons Attribution 4.0 International (CC BY 4.0) 296 Health Sains. https://doi.org/10.46799/jsa.v3i4.426 4. Kesimpulan Penelitian ini berhasil mendapatkan persamaan regresi liner berganda untuk menentukan Jumlah UMKM (Y) berdasarkan Angkatan Kerja (X1), Inflasi (X2) dan Suku Bunga (X3) yaitu Y = 884672 - 1,6X1 - 401729,7X2 - 2239731,6X3. Dari hasil analisis regresi yang dilakukan, dapat disimpulkan bahwa secara Rosa, Y. Del, & Idwar. (2019). PENGARUH INFLASI DAN SUKU BUNGA TERHADAP OMSET UMKM KOTA PADANG. Menara Ekonomi. Shofia, A., Bakhtiar, A., & Prastawa, H. (2020). The impact of critical success factor of lean six sigma implementation towards the 295 improvement of business performance on low-cost hotel industry: A literature review. AIP Conference Proceedings. https://doi.org/10.1063/5.0000707 Health Sains. https://doi.org/10.46799/jsa.v3i4.426 Health Sains. https://doi.org/10.46799/jsa.v3i4.426 Triha, H., Shofia, A., Putri, R. M., Alius, M., & Satria, T. F. (2023). Correlation between the Number of MSMEs and the Growth of the Workforce with the Regression Analysis Model. Journal of Computer Scine and Information Technology, 9, 7– 12. https://doi.org/10.35134/jcsitech.v9i1.40 Sitanggang, H., & Munthe, K. (2019). Pengaruh Inflasi, Suku Bunga Dan Nilai Tukar Terhadap Return Saham Pada Perusahaan Manufaktur Di Bursa Efek Indonesia Pada Periode 2013- 2016. Jurnal Manajemen Dan Bisnis, 18, 101–113. https://doi.org/10.54367/jmb.v18i2.424 Wicaksono, D. A. I., & Kramadibrata, B. S. (2022). Analisis Pengaruh Inflasi Dan Tingkat Suku Bunga Terhadap Kinerja Umkm Binaan Komunitas Jakpreneur Cipayung. Journal Ilmu Sosial, Manajemen, Dan Akuntansi (JISMA), 1(4), 487–500. https://melatijournal.com/index.php/jisma/article/view/186 Tasyim, D., Kawung, G. M. V, & ... (2021). Pengaruh Jumlah Unit Usaha Umkm Dan Pdrb Terhadap Penyerapan Tenaga Kerja Di Sulawesi Utara. Jurnal EMBA, 9(3), 391–400. Triana, L. (2022). Analisa Strategi Peningkatan Pendapatan Umkm Selama Pandemi Covid 19 Di Kabupaten Pandeglang. Jurnal
https://openalex.org/W4320715078	https://proceedings.unisba.ac.id/index.php/BCSME/article/download/5633/2566	Indonesian	null	Optimasi Pencampuran Batubara Dalam Upaya Meningkatkan Kualitas di PT Bukit Asam, Tbk. Tanjung Enim, Sumatera Selatan	Bandung Conference Series. Mining Engineering	2,023	cc-by	4,697	Kata Kunci: Pencampuran, Kualitas Batubara, Market Brand. Alditiya Nurrosyidi, Solihin, Sri Widayati Alditiya Nurrosyidi, Solihin, Sri Widayati Prodi Teknik Pertambangan, Fakultas Teknik, Universitas Islam Bandung, Indonesia. Prodi Teknik Pertambangan, Fakultas Teknik, Universitas Islam Bandung, Indonesia. solihin@unisba.ac.id, solihin@unisba.ac.id, solihin@unisba.ac.id, *alditiyanurrosyidi856@gmail.com, widayati_teknik@yahoo.com Abstract. PT Bukit Asam, Tbk. is a coal commodity mining industry company located in Lawang Kidul District, Muara Enim Regency, South Sumatra Province. Data collection of coal products resulting from mining called mine brand. Mining is carried out from the Air Laya Mine pit (TAL), Muara Tiga Besar pit (MTB), and Bangko pit which produces 16 Mine brands namely AL 49, AL 51, AL 53, AL 55, AL 59, AL 61, AL 63, AL 65, AL 67HV, AL 71MV, AL 69HV, AL 71LV, BB 47, BB 51, BB 53, and MT 47. Mine brands must meet market brand criteria according to consumer needs. In May 2022, consumer demand for market brand BA 48 was 538,231 tons with criteria of total moisture ≤ 30%, ash content ≤ 8%, and calorific value ≥ 4,800 kcal/kg, and market brand BA 50 was 311,814 tons with criteria of total moisture ≤ 28%, ash content ≤ 7%, and calorific value ≥ 5,000 kcal/kg. To meet consumer demand, it is necessary to maximize coal blending carried out by the company, so a coal blending optimization simulation is carried out using the simplex method calculation in a linear program with POM-QM For Windows software as a tool for applying the simplex method. The results of BA 48 market brand coal blending obtained a proportion of 60.73% BB 51, and 39.27% MT 47 with total moisture parameters of 29.81%, ash content of 3.14%, and calorific value of 4,800 kcal/kg, while for BA 50 market brand coal blending obtained a proportion of 52.50% AL 49 and 47.50% BB 53 with total moisture parameters of 27.75%, ash content of 2.81%, and calorific value of 5,000 kcal/kg. Keywords: Blending, Coal Quality, Market Brand. Bandung Conference Series: Mining Engineering Bandung Conference Series: Mining Engineering https://doi.org/10.29313/bcsme.v3i1.5633 Corresponding Author Email: solihin@unisba.ac.id A. Pendahuluan Batubara merupakan bahan bakar hidrokarbon padat yang terbentuk dari proses penggambutan dan pembatubaraan di dalam suatu cekungan (daerah rawa) dalam jangka waktu geologis yang meliputi aktivitas bio-geokimia terhadap akumulasi flora di alam yang mengandung selulosa dan lignin (Sukandarrumidi, 1995). Batubara terdiri atas unsur- unsur utama, yaitu: karbon, hidrogen, oksigen, dan unsur-unsur tambahan seperti belerang dan nitrogen (Muchjidin, 2006). Perbedaan kualitas batubara dalam satu seam yang sama sangat mungkin terjadi baik itu secara lateral maupun vertikal. Keadaan ini dapat disebabkan oleh perbedaan proses pengendapan, komposisi penyusun, serta akumulasi pengotor yang terikut saat proses pembatubaraan atau coalification (Saputra, dkk, 2014). Selain itu proses pengambilan serta penanganan batubara dari front penambangan juga menyebabkan terjadinya perbedaan ataupun perubahan kualitas batubara. PTBA sebagai BUMN perusahaan tambang yang bergerak di bidang pertambangan batubara memiliki kualitas batubara yang bervariasi. Adanya perbedaan kualitas batubara yang ditambang serta tanggung jawab dalam memenuhi kriteria kontrak dan komitmen menjadi suatu tantangan tersendiri bagi perusahaan (Saputra, dkk, 2014). Strategi untuk menyelesaikan masalah tersebut salah satunya adalah dengan melakukan pencampuran batubara. Pencampuran batubara merupakan suatu pekerjaan mencampurkan dua jenis batubara atau lebih yang kualitasnya berbeda untuk memperoleh satu jenis batubara dengan kualitas yang sesuai dengan spesifikasi dalam kontrak (Muchjidin, 2006). Proses pencampuran batubara dapat membantu meningkatkan kualitas dan nilai kalori batubara sesuai dengan kebutuhan konsumen baik domestik maupun ekspor. Pencampuran batubara merupakan salah satu cara untuk menyesuaikan parameter kualitas batubara yang dijual agar sesuai dengan kriteria yang diharapkan pembeli (Prasetyo, dkk, 2016). Proses pencampuran batubara seringkali tidak sesuai dengan target yang diharapkan sehingga nilai jualnya akan menjadi lebih rendah. Hal ini dikarenakan oleh banyaknya seam batubara dengan kualitas yang berbeda sehingga pencampuran menjadi sulit karena tercampurnya batubara dengan material pengotor, dan stockpile management yang kurang baik (Anwar dan Arief, 2011). PTBA memiliki kendala yang sering terjadi dalam proses pencampuran batubara karena pembagian dari dua atau lebih mine brand yang belum sesuai, sehingga mengakibatkan parameter kualitas batubara yaitu nilai kelembaban, kandungan abu, dan nilai kalori tidak memenuhi permintaan konsumen. Maka dilakukan penelitian untuk menentukan pembagian batubara dengan kualitas yang berbeda dalam upaya meningkatkan dan mengoptimalkan kualitas batubara, sehingga dapat memenuhi kuantitas dan kualitas batubara yang sesuai dengan kriteria permintaan konsumen. Berdasarkan latar belakang yang telah diuraikan, maka perumusan masalah dalam penelitian ini sebagai berikut: “Bagaimana upaya memenuhi spesifikasi batubara untuk dilakukan proses pencampuran batubara dalam memenuhi market brand BA 48 dan BA 50 pada bulan Mei 2022?”. Keywords: Blending, Coal Quality, Market Brand. Keywords: Blending, Coal Quality, Market Brand. Abstrak. PT. Bukit Asam, Tbk. merupakan perusahaan industri pertambangan komoditas batubara berlokasi di Kecamatan Lawang Kidul, Kabupaten Muara Enim, Provinsi Sumatera Selatan. Pengambilan data produk batubara hasil dari penambangan yang disebut mine brand. Penambangan dilakukan dari pit Tambang Air Laya (TAL), pit Muara Tiga Besar (MTB), dan pit Bangko yang menghasilkan 16 Mine brand yaitu AL 49, AL 51, AL 53, AL 55, AL 59, AL 61, AL 63, AL 65, AL 67HV, AL 71MV, AL 69HV, AL 71LV, BB 47, BB 51, BB 53, dan MT 47. Mine brand harus memenuhi kriteria market brand sesuai kebutuhan konsumen. Pada bulan Mei 2022 permintaan konsumen untuk market brand BA 48 sebanyak 538.231 ton dengan kriteria total kelembaban ≤ 30%, kandungan abu ≤ 8%, dan nilai kalori ≥ 4.800 kkal/kg, dan market brand BA 50 sebanyak 311.814 ton dengan kriteria total kelembaban ≤ 28%, kandungan abu ≤ 7%, dan nilai kalori ≥ 5.000 kkal/kg. Untuk memenuhi permintaan konsumen perlu memaksimalkan pencampuran batubara yang dilakukan oleh perusahaan, maka dilakukan simulasi optimasi pencampuran batubara yang dilakukan menggunakan perhitungan metode simpleks dalam program linier dengan software POM-QM For Windows sebagai alat penerapan metode simpleks. Hasil dari pencampuran batubara market brand BA 48 mendapatkan proporsi sebesar 60,73% BB 51, dan 39,27% MT 47 dengan parameter total kelembaban 29,81%, kandungan abu 3,14%, dan nilai kalori 4.800 kkal/kg, sedangkan untuk pencampuran batubara market brand BA 50 mendapatkan proporsi sebesar 52,50% AL 49 dan 47,50% BB 53 dengan parameter total kelembaban 27,75%, kandungan abu 2,81%, dan nilai kalori 5.000 kkal/kg. Kata Kunci: Pencampuran, Kualitas Batubara, Market Brand. Corresponding Author Email: solihin@unisba.ac.id 74 75 Optimasi Pencampuran Batubara Dalam Upaya Meningkatkan Kualitas di PT. Bukit Asam, Tbk … \| 75 A. Pendahuluan Selanjutnya, terdapat tujuan dalam penelitian ini diuraikan dalam pokok-pokok sebagai berikut: 1. Mengetahui spesifikasi dan jumlah tonase market brand BA 48 dan BA 50 yan diinginkan konsumen di bulan Mei 2022. 2. Menentukan perbandingan proporsi tonase pada pencampuran batubara sesuai dengan market brand BA 48 dan market brand BA 50 dengan menerapkan perhitungan metode simpleks. p 3. Merekomendasikan alternatif pencampuran mine brand untuk memenuhi market brand BA 48 dan BA 50 dengan memaksimalkan penggunaan semua mine brand yang optimal untuk dilakukan pencampuran. B. Metodologi Penelitian Metodologi penelitian yang dilakukan penulis dengan menggunakan alisis berdasarkan perhitungan dari metode simpleks yang terdiri dari beberapa tahapan. Tahapan pertama dilakukan pengambilan 2 jenis data yaitu data primer dan data sekunder. Data primer yang Mining Engineering 76 \| Alditiya Nurrosyidi, et al. digunakan yaitu peringkat batubara dan tonase dari front penambangan, dan data sekunder yang digunakan yaitu data curah hujan dan data permintaan market brand oleh konsumen meliputi tonase maupun spesifikasi peringkat batubara berupa nilai kelembaban, kandungan abu, dan nilai kalori. Metode analisa yang dilakukan dari hasil optimasi kegiatan pencampuran batubara akan menghasilkan suatu rekomendasi rasio pencampuran batubara mine brand dalam memenuhi kebutuhan market brand, apabila sesuai akan dilanjutkan dengan proses pengiriman, sebaliknya apabila tidak sesuai dapat dilakukannya proses optimasi pencampuran batubara kembali dan di analisis kembali dari kualitas dan konsumen merasa tertarik dengan promosi yang dilakukan perusahaan. Secara umum batubara didefinisikan sebagai batuan organik berwarna gelap yang terbentuk dari jasad tumbuh-tumbuhan. Kandungan utama batubara adalah atom karbon, hidrogen, dan oksigen (Pasymi, 2008). Batubara dapat didefinisikan sebagai satuan sedimen yang terbentuk dari dekomposisi tumpukan tanaman selama kira-kira 300 juta tahun. Dekomposisi tanaman ini terjadi karena proses biologi dengan mikroba dimana banyak oksigen dalam selulosa diubah menjadi karbondioksida dan air (Purnamasari Yunita, 2000). Batubara merupakan bahan mineral yang heterogen baik secara kimia maupun fisika, yang tersusun dari unsur utama karbon, hidrogen, oksigen, sulfur, dan nitrogen. Tingkat pembatubaraan secara umum berhubungan dengan karakteristik batubara. Batubara dilihat dari tingginya tingkat pembatubaraan maka nilai karbon akan meningkat sedangkan nilai hidrogen, oksigen, dan sulfur akan bekurang. Batubara dengan tingkat pembatubaraan yang rendah seperti lignit dan sub-bituminus lebih lembut dengan materi yang rapuh, berwarna suram seperti tanah, memiliki nilai kelembaban yang tinggi dan nilai karbon yang rendah. Semakin tinggi peringkat batubara, umumnya akan semakin keras dan kompak, serta warna akan semakin hitam mengkilat. Selain itu, kelembaban batubara akan berkurang sedangkan nilai karbon akan meningkat, sehingga memiliki kandungan energi yang juga semakin besar. Menurut Saputra, dkk (2014), Pencampuran dilakukan pada batubara yang beda nilai kalori, kelembaban, kandungan abu, serta kandungan sulfur, sehingga kualitas batubara hasil campuran merupakan perpaduan dari parameter kualitas batubara yang dicampur. Teknik pencampuran batubara dilakukan dengan menggunakan teknik pengaturan tumpukan (stacking) pada stockpile dan teknik pengaturan laju batubara pada conveyor. Menurut R.N.F. Majid, dkk (2019), Proses pencampuran di lapangan kadangankala tidak mencapai titik optimum dan tidak sesuai dengan target yang diinginkan sehingga mengurangi kualitas dari batubara itu sendiri. Vol. 3 No. 1 (2023), Hal: 74-81 B. Metodologi Penelitian Aspek-aspek penyebab ketidaktercapaian target nilai kalori yang ditetapkan adalah ketidaktercapai kualitas batubara sesuai permintaan. Selain itu juga dapat disebabkan oleh ketidaksesuain rasio komposisi berat batubara yang akan dilakukan pencampuran, sehingga nilai kalori diperoleh di bawah target yang akan diminta oleh konsumen. Secara teoritis parameter kualitas campurannya dapat didekati dengan persamaan berikut. sa aa be ut. KBc = "(KB1×PB1)+(KB2×PB2)+...(KBn×PBn)" /"PBC" Keterangan : KB1 : Parameter kualitas batubara 1 KB2 : Parameter kualitas batubara 2 KBn : Parameter kualitas batubara ke-n KBc : Parameter kualitas batubara hasil campuran PB1 : Persentase batubara 1 (MT) PB2 : Persentase batubara 2 (MT) PBn : Persentase batubara ke-n (MT) PBc : Persentase batubara hasil campuran KBc = "(KB1×PB1)+(KB2×PB2)+...(KBn×PBn)" /"PBC" Keterangan : KBc = "(KB1×PB1)+(KB2×PB2)+...(KBn×PBn)" /"PBC" Keterangan : KB1 : Parameter kualitas batubara 1 KB2 : Parameter kualitas batubara 2 KBn : Parameter kualitas batubara ke-n KBc : Parameter kualitas batubara hasil campuran PB2 : Persentase batubara 2 (MT) PBn : Persentase batubara ke-n (MT) PBc : Persentase batubara hasil campuran C. Hasil Penelitian dan Pembahasan PTBA memiliki 3 lokasi utama pit penambangan perusahaan melakukan pengendalian kualitas batubara berupa pengambilan sampel pada setiap stockpile. Parameter kualitasnya meliputi total kelembaban, total moisture bawaan, kandungan abu, zat terbang, karbon Vol. 3 No. 1 (2023), Hal: 74-81 ISSN: 2828-2140 ISSN: 2828-2140 Optimasi Pencampuran Batubara Dalam Upaya Meningkatkan Kualitas di PT. Bukit Asam, Tbk … \| 77 77 tertambat, kadar sulfur dan nilai kalori. PTBA melakukan penyesuaian terhadap produk batubara hasil dari proses penambangan sebesar 3% sesuai pada Tabel 1 hal tersebut dilakukan untuk mengantisipasi penurunan peringkat batubara atau yang disebut dengan dilusi. Komposisi pencampuran ditentukan berdasarkan nilai dari kelembaban, kandungan abu dan nilai kalori. Hal ini dilakukan untuk mengantisipasi penurunan peringkat batubara selama proses pengangkutan batubara. Tabel 1. Parameter Kualitas Mine Brand Mine Brand Parameter Kualitas Batubara TM (%, ar) IM (%. ar) ASH (%, ar) VM (%, ar) FC (%, ar) TS (%, ar) CV (Kkal/kg, ar) AL-49 28,36 10,12 3,33 34,84 34,35 0,21 4.848 AL-51 28,33 10,61 3,48 34,72 34,35 0,37 4.902 AL-53 26,41 10,19 3,21 35,71 35,48 0,72 5.099 AL-55 22,76 7,96 5,42 36,37 36,21 2,45 5.300 AL-59 19,45 8,18 2,86 38,41 39,82 0,42 5.739 AL-61 17,69 6,11 2,85 38,25 41,66 0,25 5.866 AL-63 15,70 6,01 3,47 39,07 42,04 0,27 6.062 AL-65 13,26 6,40 3,18 40,46 43,44 0,27 6.299 AL-67 HV 11,55 4,94 2,93 40,74 45,09 0,33 6.537 AL-69 HV 9,56 4,48 2,67 40,33 47,50 0,38 6.800 Al-71 MV 3,58 1,11 5,43 31,61 58,76 1,79 7.554 Al-71 LV 3,83 0,98 6,44 24,12 64,63 2,26 7.451 BB-47 30,78 10,15 1,77 33,52 34,79 0,37 4.777 BB-51 28,53 9,89 2,99 34,74 34,60 0,47 4.950 BB-53 27,08 9,24 2,24 35,19 36,26 0,77 5.168 MT-47 31,81 10,53 3,39 33,64 32,18 0,47 4.568 Sumber: PT Bukit Asam Tbk., 2022. Tabel 1. Parameter Kualitas Mine Brand Tabel 1. Parameter Kualitas Mine Brand Pada bulan Mei 2022 dilakukannya perencanaan produksi batubara sebanyak 2.630.000 ton. Kegiatan penambangan dilakukan dengan metode tambang terbuka pada 3 lokasi penambangan yaitu TAL, MTB dan tambang BB. Kegiatan rencana produksi PTBA ditujukan untuk memenuhi kebutuhan batubara secara domestik dan ekspor dengan beberapa konsumen yang berbeda. C. Hasil Penelitian dan Pembahasan Untuk stock awal batubara bulan Mei 2022, pada live stockpile dan temporary stockpile TAL yang terdiri dari AL 49, AL 51, AL 53, AL 55, AL 59, AL 61, AL 63, AL 65, AL 67 HV, AL 69 HV, AL 71 MV dan AL 71 LV jumlah produk batubara hasil dari proses penambangan di bulan Mei 2022 sebesar 227.378 ton. Sehingga mendapatkan total tonase batubara dibulan Mei 2022 sebesar 2.857.378 ton. Parameter kualitas yang diinginkan pasar lebih memperhatikan pada total kelembaban, kandungan abu dan nilai kalori sesuai dengan kontrak jual beli PTBA dengan konsumen yang mana kriteria parameter kualitas batubara BA 48 dan BA 50 terdapat pada Tabel 2 di bawah ini. Tabel 2. Permintaan Kualitas Batubara Konsumen Parameter Market brand Satuan BA-48 BA-50 Total Moisture 30,00 28,00 (%, ar) Ash 8,00 7,00 (%, ar) Calorivic Value 4.800,00 5.000,00 (Kkal/Kg, ar) Sumber: PT Bukit Asam Tbk., 2022. Tabel 2. Permintaan Kualitas Batubara Konsumen Tabel 2. Permintaan Kualitas Batubara Konsumen Permintaan batubara market brand BA 48 pada bulan Mei 2022 sebanyak 398.381 ton yang akan dikirim ke Pelabuhan Tarahan dan melalui Pelabuhan Kertapati sebanyak 139.850 ton. Parameter kualitas yang dituju dan diinginkan konsumen pada BA 48 yaitu kelembaban air ≤ 30%, kandungan abu ≤ 8%, dan nilai kalori ≥ 4.800 Kkal/kg. Kemudian dilakukan pencampuran batubara dari 3 produk batubara hasil proses penambangan untuk mendapatkan Mining Engineering 78 \| Alditiya Nurrosyidi, et al. proporsi yang optimal dengan menggunakan software POM-QM For Windows yang menerapkan perhitungan program linier metode simpleks. Market brand BA 48 dapat dipenuhi dengan pencampuran produk batubara hasil proses penambangan AL 51, BB 51, dan MT 47 dengan tonase diasumsikan sebesar 1 ton dan parameter sesuai pada Tabel 3 sehingga dapat diketahui perbandingan tonase yang dibutuhkan dari jumlah proporsi yang didapatkan dari software tersebut. Tabel 3. Permintaan Kualitas Batubara Konsumen Jenis Batubara TM ASH CV (%, ar) (%, ar) (Kkal/kg, ar) AL 51 28,33 3,48 4.902 BB 51 28,53 2,99 4.950 MT 47 31,81 3,39 4.568 Kriteria Permintaan ≤ 30,00 ≤ 8,00 ≥ 4.800 Sumber: PT Bukit Asam Tbk., 2022. Tabel 3. Permintaan Kualitas Batubara Konsumen Setelah dilakukan perhitungan optimasi dengan menggunakan program linier perhitungan simpleks, sehingga dihasilkan output proporsi untuk memenuhi market brand BA 48. Sehingga mendapatkan perbandingan proporsi pencampuran sesuai pada Tabel 4 sebagai berikut : Tabel 4. Permintaan Kualitas Batubara Konsumen Tabel 4. C. Hasil Penelitian dan Pembahasan Permintaan Kualitas Batubara Konsumen Jenis Batubara Proporsi TM ASH CV (%, ar) (%, ar) (Kkal/kg, ar) AL 51 0,00 % 28,33 3,48 4902 BB 51 60,73 % 28,53 2,99 4950 MT 47 39,27 % 31,81 3,39 4568 Hasil Pencampuran 29,81 3,14 4.800 Kriteria Permintaan ≤ 30,00 ≤ 8,00 ≥ 4800 Sumber: Hasil Pengolahan Data, 2022 Permintaan market brand BA 48 pada bulan Mei 2022 sebesar 538.231 ton yang akan dilakukan pengiriman melalui Pelabuhan Tarahan sebanyak 398.381 ton dan untuk pengiriman melalui Dermaga Kertapati sebanyak 139.850 ton untuk konsumen luar negeri. Untuk memenuhi permintaan pengiriman batubara BA 48 di bulan Mei 2022 sebesar 398.381 ton ke Pelabuhan Tarahan maka jumlah batubara produk batubara hasil proses penambangan BB 51 dan MT 47 dapat dihitung menggunakan perhitungan di bawah ini yaitu: Kebutuhan (Ton) mine brand BB 51 = Permintaan x Proporsi BB 51 = 398.381 Ton x 60,73% = 241.937 Ton Kebutuhan (Ton) mine brand MT 47 = Permintaan x Proporsi MT 47 = 398.381 Ton x 39,27% = 156.444 Ton P i i i b b BA 48 di b l M i 2022 b 39 = 398.381 Ton x 60,73% = 241.937 Ton Kebutuhan (Ton) mine brand MT 47 = Permintaan x Proporsi MT 47 = 398.381 Ton x 39,27% = 156.444 Ton Permintaan pengiriman batubara BA 48 di bulan Mei 2022 sebesar 398.381 ton ke aga Kertapati maka j mlah bat bara prod k bat bara hasil proses penambangan AL 51 Permintaan pengiriman batubara BA 48 di bulan Mei 2022 sebesar 398.381 ton k Permintaan pengiriman batubara BA 48 di bulan Mei 2022 sebesar 398.381 ton ke ermaga Kertapati maka jumlah batubara produk batubara hasil proses penambangan AL 51, p g Dermaga Kertapati maka jumlah batubara produk batubara hasil proses penambangan AL 51 BB 51, dan MT 47 dapat dihitung menggunakan perhitungan di bawah ini yaitu: Kebutuhan (Ton) mine brand BB 51 = Permintaan x Proporsi BB 51 = 139.850 Ton x 60,73% = 84.931 Ton Kebutuhan (Ton) mine brand MT 47 = Permintaan x Proporsi MT 47 = 139.850 Ton x 39,27% = 54.919 Ton BB 51, dan MT 47 dapat dihitung menggunakan perhitungan di bawah ini yaitu: Kebutuhan (Ton) mine brand BB 51 = Permintaan x Proporsi BB 51 = 139.850 Ton x 60,73% = 84.931 Ton Kebutuhan (Ton) mine brand MT 47 = Permintaan x Proporsi MT 47 = 139.850 Ton x 39,27% = 54.919 Ton = 54.919 Ton Permintaan batubara market brand BA 50 pada bulan Mei 2022 sebanyak 82.917 ton yang dikirim ke PLTU Bukit Asam, 121.111 ton yang akan dikirim ke Pelabuhan Tarahan dan melalui Pelabuhan Kertapati sebanyak 107.786 ton. Parameter kualitas yang dituju dan Vol. 3 No. 1 (2023), Hal: 74-81 ISSN: 2828-2140 ISSN: 2828-2140 Optimasi Pencampuran Batubara Dalam Upaya Meningkatkan Kualitas di PT. Bukit Asam, Tbk … \| 79 79 diinginkan konsumen pada BA 50 yaitu kelembaban air ≤ 30%, kandungan abu ≤ 7%, dan nilai kalori ≥ 5000 Kkal/kg. Kemudian dilakukannya pencampuran batubara dari 2 produk batubara hasil proses penambangan untuk mendapatkan proporsi yang optimal dengan menggunakan software POM-QM For Windows yang menerapkan perhitungan program linier metode simpleks. Market brand BA 50 dapat dipenuhi dengan pencampuran produk batubara hasil proses penambangan AL 49 dan BB 53 dengan tonase diasumsikan sebesar 1 ton dan parameter sesuai pada Tabel 5 sehingga dapat diketahui perbandingan tonase yang dibutuhkan dari persenan proporsi yang didapatkan dari software tersebut. Tabel 5. Permintaan Kualitas Batubara Konsumen Jenis Batubara TM ASH CV (%, ar) (%, ar) (Kkal/kg, ar) AL 49 28,36 3,33 4.848 BB 53 27,08 2,24 5.168 Kriteria Permintaan ≤ 28,00 ≤ 7,00 ≥ 5.000 Sumber: PT Bukit Asam Tbk., 2022. Tabel 5. Permintaan Kualitas Batubara Konsumen Tabel 5. Permintaan Kualitas Batubara Konsumen Setelah dilakukan perhitungan optimasi dengan menggunakan program linier perhitungan simpleks, sehingga dihasilkan output proporsi untuk memenuhi Market brand BA 50. Sehingga mendapatkan perbandingan proporsi pencampuran sesuai pada Tabel 6 sebagai berikut : Tabel 6. Permintaan Kualitas Batubara Konsumen Tabel 6. Permintaan Kualitas Batubara Konsumen Jenis Batubara Proporsi TM ASH CV (%, ar) (%, ar) (Kkal/kg, ar) AL 49 52,50% 28,36 3,33 4.848 BB 53 47,50% 27,08 2,24 5.168 Hasil Pencampuran 27,75 2,81 5.000 Kriteria Permintaan ≤ 28,00 ≤ 7,00 ≥ 5.000 Sumber: Hasil Pengolahan Data, 2022. Permintaan market brand BA 50 pada bulan Mei 2022 sebesar 311.814 ton yang akan dilakukan pengiriman melalui Pelabuhan Tarahan dan Dermaga Kertapati. Kemudian pengiriman melalui Pelabuhan Tarahan yaitu sebanyak 121.111 ton, sedangkan pengiriman melalui Dermaga Kertapati sebanyak 107.786 ton. = 54.919 Ton Untuk memenuhi permintaan pengiriman batubara BA 50 di bulan Mei 2022 sebesar 121.111 ton ke Pelabuhan Tarahan maka jumlah batubara produk batubara hasil proses penambangan AL 49 dan BB 53 dapat dihitung menggunakan perhitungan di bawah ini yaitu: gg p g y Kebutuhan (Ton) mine brand AL 49 = Permintaan x Proporsi AL 49 Kebutuhan (Ton) mine brand AL 49 = Permintaan x Proporsi AL 49 = 121.111 Ton x 52,50% = 63.583 Ton Kebutuhan (Ton) mine brand BB 53 = Permintaan x Proporsi BB 53 = 121.111 Ton x 47,50% = 57.528 Ton P i t i i b t b BA 50 di b l M i 2022 b 107 786 t k Kebutuhan (Ton) mine brand AL 49 = Permintaan x Proporsi AL 49 = 121.111 Ton x 52,50% = 63.583 Ton Kebutuhan (Ton) mine brand BB 53 = Permintaan x Proporsi BB 53 = 121.111 Ton x 47,50% = 57.528 Ton Permintaan pengiriman batubara BA 50 di bulan Mei 2022 sebesar 107.786 ton ke Dermaga Kertapati maka jumlah batubara produk batubara hasil proses penambangan AL 49 dan BB 53 dapat dihitung menggunakan perhitungan di bawah ini yaitu: Kebutuhan (Ton) mine brand AL 49 = Permintaan x Proporsi AL 49 = 107.786 Ton x 52,50% = 56.588 Ton Kebutuhan (Ton) mine brand BB 53 = Permintaan x Proporsi BB 53 Dermaga Kertapati maka jumlah batubara produk batubara hasil proses penambangan AL 49 dan BB 53 dapat dihitung menggunakan perhitungan di bawah ini yaitu: Dermaga Kertapati maka jumlah batubara produk batubara hasil proses penam dan BB 53 dapat dihitung menggunakan perhitungan di bawah ini yaitu: Dermaga Kertapati maka jumlah batubara produk batubara hasil proses penambangan AL dan BB 53 dapat dihitung menggunakan perhitungan di bawah ini yaitu: dan BB 53 dapat dihitung menggunakan perhitungan di bawah ini yaitu: Kebutuhan (Ton) mine brand AL 49 = Permintaan x Proporsi AL 49 = 107.786 Ton x 52,50% = 56.588 Ton Kebutuhan (Ton) mine brand BB 53 = Permintaan x Proporsi BB 53 = 107.786 Ton x 47,50% = 51.198 Ton Kebutuhan (Ton) mine brand AL 49 = Permintaan x Proporsi AL 49 = 107.786 Ton x 52,50% Mining Engineering 80 \| Alditiya Nurrosyidi, et al. D. Kesimpulan Berdasarkan hasil penelitian yang dilakukan mengenai optimasi pencampuran batubara dalam upaya meningkatkan kualitasnya di PT. BA, Tanjung Enim, Sumatera Selatan, dapat diambil kesimpulan sebagai berikut : p y g kesimpulan sebagai berikut : 1. Permintaan market brand BA 48 pada bulan Mei 2022 sebesar 538.231 ton yang akan dilakukan pengiriman melalui Pelabuhan Tarahan dan Dermaga Kertapati. Pengiriman melalui Pelabuhan Tarahan yaitu sebanyak 398.381 ton sedangkan untuk pengiriman melalui Dermaga Kertapati sebanyak 139.850 ton dengan parameter kualitas yang diinginkan konsumen pada BA 48 yaitu kelembaban air ≤ 30%, kandungan abu ≤ 8%, dan nilai kalori ≥ 4800 Kkal/kg. Kemudian permintaan batubara market brand BA 50 pada bulan Mei 2022 sebanyak 311.814 ton yang akan dikirim melalul Pelabuhan Tarahan sabanyak 121.111 ton, melalui Pelabuhan Kertapati sebanyak 107.786 ton, dan PLTU Bukit Asam sebanyak 82.917 ton. Parameter kualitas yang dituju dan diinginkan konsumen pada BA 50 yaitu kelembaban air ≤ 28%, kandungan abu ≤ 7%, dan nilai kalori ≥ 5000 Kkal/kg. 2. Berdasarkan hasil optimasi pencampuran batubara untuk mendapatkan batubara market brand BA 48 dan BA 50 menggunakan POM-QM For Windows dengan metode program linier perhitungan simpleks. Maka dihasilkan proporsi untuk market brand BA 48 sebesar 60,73 % BB 51 dan 39,27 % MT 47 yang menghasilkan kualitas batubara dengan nilai kelembaban 29,81 %, kandungan abu 3,14 %, dan nilai kalori 4.800 Kkal/kg. Sedangkan market brand BA 50 sebesar 52,50% AL 49 : 47,50% BB 53 dan menghasilkan kualitas batubara dengan kelembaban air 27,25 %, kandungan abu 2,81 %, dan nilai kalori 5.000 Kkal/kg g 3. Rekomendasi alternatif pencampuran batubara dilakukan guna meminimalisir adanya kehabisan atau kekurangan stock produk batubara hasil proses penambangan untuk mendapatkan kualitas market brand BA 48 dan BA 50. Rekomendasi yang didapatkan untuk market brand BA 48 yaitu 43,69% AL 53 : 56,31% MT 47, 31,69% AL 55 : 68,31% MT 47, dan 82,86% AL 49 : 17,14% MT 47. Kemudian, untuk market brand BA 50 rekomendasi alternatif pencampuran batubara yaitu 63,16% AL 51 : 36,84% BB 53, 77,06% BB 51 : 22,94% BB 53 dan 66,37% AL 49 : 33,63% AL 55. Acknowledge g 1. Ibunda Suminten dan Ayahanda Nuryanto tercinta sebagai tanda bakti, hormat, dan rasa terima kasih yang tiada terhingga kupersembahkan karya kecil ini kepada Ibu dan Ayah yang telah memberikan kasih sayang, segala dukungan, dan cinta kasih yang tiada terhingga yang tiada mungkin dapat kubalas hanya dengan selembar kertas yang bertuliskan kata cinta dalam kata persembahan. Semoga ini menjadi langkah awal untuk membuat Ibu dan Ayah bahagia karna kusadar, selama ini belum bisa berbuat yang lebih. 2. Skripsi ini Saya dedikasikan untuk orang-orang yang sangat Saya sayangi yaitu adik pertama Daffa Mukhtar Rosyidi, adik kedua Muhamad Varel Rosyidi dan keponakan Muhammad Faisal. Terkadang, ketika Saya kehilangan kepercayaan pada diri Saya sendiri, kalian di sini untuk percaya pada Saya dengan cara memberi canda dan tawa. 3. Skripsi ini Saya persembahkan untuk teman atau sahabat Saya selama 4 tahun di perkuliahan yaitu Naufal Abdan S, Muhammad Aziz R, Roby Ardianta, Wahyu Guntoro dan Rizky Ade P. Saya bahkan tidak bisa menjelaskan betapa bersyukurnya Saya memiliki kalian dalam hidup Saya. Daftar Pustaka [1] Anwar, Ubaidillah and Arief, A.Taufik. 2011. “Model Matematika Untuk Optimasi Nilai Kalori Batu bara Pencampuran di PT. Batu bara Bukit Kendi Tanjung Enim- Sumatera Selatan. Palembang”. Prosiding AVoER III. [1] Anwar, Ubaidillah and Arief, A.Taufik. 2011. “Model Matematika Untuk Optimasi Nilai Kalori Batu bara Pencampuran di PT. Batu bara Bukit Kendi Tanjung Enim- Sumatera Selatan. Palembang”. Prosiding AVoER III. [1] Anwar, Ubaidillah and Arief, A.Taufik. 2011. “Model Matematika Untuk Optimasi Nilai Kalori Batu bara Pencampuran di PT. Batu bara Bukit Kendi Tanjung Enim- Sumatera Selatan. Palembang”. Prosiding AVoER III. [2] g g [2] Amanda, Agzoci Zulva. 2020. "Optimalisasi Pencampuran Batubara Untuk Memenuhi Kriteria Permintaan Konsumen Dengan Menggunakan Metode Simplek di PT. Nusa [2] Amanda, Agzoci Zulva. 2020. "Optimalisasi Pencampuran Batubara Untuk Memenuhi Kriteria Permintaan Konsumen Dengan Menggunakan Metode Simplek di PT. Nusa Vol. 3 No. 1 (2023), Hal: 74-81 ISSN: 2828-2140 81 Optimasi Pencampuran Batubara Dalam Upaya Meningkatkan Kualitas di PT. Bukit Asam, Tbk … \| 81 Alam Lestari Kota Sawahlunto, Provinsi Sumatera Barat”. Sekolah Tinggi Teknologi Industri (STTIND) Padang. Billah, Mutasim. 2010. "Peningkatan Nilai Kalor Batubara Peringkat Rendah Dengan Menggunakan Minyak Tanah dan Minyak Residu". Teknik jurusan Teknik Kimia, Fakultas Teknologi Industri Universitas Pembangunan Nasional “Veteran” Jawa Timur. [4] Effendi Burlian, Zainal. 2020. “Ilmu Alamiah Dasar, Ilmu Budaya Dasar, dan Ilmu Sosial Dasar”. Buku Edisi I. ISBN: 978-623-7374-98-5 [5] Komariah, W. E. 2012. “Peningkatan Kualitas Batu bara Indonesia Peringkat Rendah Melalui Penghilang Moisture dengan Pemanasan Gelombang Mikro”. Tesis Teknik Kimia Universitas Indonesia. Jakarta. [6] Kuntaarsa, Abdullah dan Purwo Subagyo. 2020. "Desulfurisasi Batubara Dengan Metode Flotasi Menggunakan Gel Lidah Buaya". Jurnal Teknologi Technoscientia ISSN: 1979-8415 Vol. 12 No. 2. Lestari, Sonia dan Rijal Abdullah. 2017. “Optimalisasi Pencampuran Batu bara Untuk Memenuhi Kriteria Permintaan Konsumen Dengan Menggunakan Metode Simplek dan Evaluasi Biaya Pada Proses Pencampuran Batu bara Di Lokasi CV. Tahiti Coal, Talawi, Sawahlunto, Sumatera Barat”. Jurnal bina tambang Vol 3 No.3. [8] Majid, R.N.F, M. Yusuf, dan S. Komar. 2019. "STUDI PENGARUH ROTASI PER MENIT TERHADAP PARAMETER KUALITAS BLENDING BATUBARA MINE BRAND MT-46 DAN AL-55 PT. BUKIT ASAM, TBK". Jurnal Pertambangan Vol 3. No 3. Agustus 2019. ISSN 2549-1008. [9] [9] Mangestiono, Wiji. 2014. "Pembuatan Prototipe Pengering Batubara Untuk Mengurangi Moisture Content Pada PLTU". Gema Teknologi Vol. 18 No. 1. [10] Muchjidin. 2006. “Pengendalian Mutu dalam Industri Batu bara.” Bandung: Institut Teknologi Bandung. [11] Muhammad Armaghani Haqq. 2018. Vol. 3 No. 1 (2023), Hal: 74-81 Mada, Gajah Mada University Press: Yogyakarta. ISSN: 2828-2140 Daftar Pustaka “Rencana Teknis Pencampuran Batubara Untuk Pemenuhan Produk Market Brand BA-48 dan BA-55 Di PT. Bukit Asam Tbk, Tanjung Enim, Sumatera Selatan”. Fakultas Teknik, Jurusan Teknik Pertambangan. Universitas Islam Bandung. [12] Nafian, Muhammad Algadri dan Yan Rizal. 2021. “Geologi Batubara Daerah Tanjung Enim, Kabupaten Muara Enim, Provinsi Sumatera Selatan”. Bulletin Of Geology, Vol. 5, No. 2. [13] Pasymi. 2008. “Batubara (Jilid-1)”. Bung Hatta University Press. DRPM DIKTI Indonesia. [14] Purnamasari, Yunita. 2000. “Pembuatan Briket Dari Batubara Kualitas Rendah Dengan Proses Non Karbonisasi Dengan Menambahkan MgO dan MgCl2”. UPN Veteran: Jawa Timur. [15] Puspitasari, Mitha dan Mahreni. 2019. “Pencucian Batu Bara”. LPPM UPN "Veteran" Yogyakarta, Yogyakarta. ISBN 9786025534485. [16] Prasetyo, Agung Dwi, Agus Triantoro, Uyu Saismana, Wahyu Permadi dan Hafidz Noor Fikri. 2016. “Optimasi Pencampuran Batubara Melalui Simulasi Berdasarkan Kriteria Parameter Batubara”. Jurnal Himasapta, Vol. 1, No. 1, April 2016 : 11-16. [17] Ramayanti, Fitri. 2016. "Desulfurisasi Batubara Dengan Metode Elektrolisis Ditinjau dari Konsentrasi Elektrolit H2SO4 dan Kecepatan Pengadukan". Politeknik Negeri Sriwijaya. Palembang [18] Saputra, Dimas, Agus Triantoro dan Riswan. 2014. “Simulasi Pencampuran Batu bara di bawah standar kontrak Dalam Pencampuran dua jenis grade beda kualitas”. Jurnal Fisika FLUX, Vol.11 [19] Santoso, Binarko. 2015. “Petrologi Batu Bara Sumatra dan Kalimantan: Jenis, Peringkat dan Aplikasi”. Jakarta: LIPI Press. [20] Sukandarrumidi. 1995. “Batubara dan Gambut”. Fakultas teknik Universitas Gadja Mining Engineering Mining Engineering 82 \| Alditiya Nurrosyidi, et al. Mada, Gajah Mada University Press: Yogyakarta. Vol. 3 No. 1 (2023), Hal: 74-81 ISSN: 2828-2140 ISSN: 2828-2140
https://openalex.org/W4319263712	https://satellite-navigation.springeropen.com/counter/pdf/10.1186/s43020-022-00093-z	English	null	Wind speed retrieval using GNSS-R technique with geographic partitioning	Satellite Navigation	2,023	cc-by	9,955	© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Abstract In this paper, the effect of geographical location on Cyclone Global Navigation Satellite System (CYGNSS) observables is demonstrated for the first time. It is found that the observables corresponding to the same wind speed vary with geographic location regularly. Although latitude and longitude information is included in the conventional method, it cannot effectively reduce the errors caused by geographic differences due to the non-monotonic changes of observables with respect to latitude and longitude. Thus, an improved method for Global Navigation Satellite Sys- tem Reflectometry (GNSS-R) wind speed retrieval that takes geographical differences into account is proposed. The sea surface is divided into different areas for independent wind speed retrieval, and the training set is resampled by considering high wind speed. To balance between the retrieval accuracies of high and low wind speeds, the results with the random training samples and the resampling samples are fused. Compared with the conventional method, in the range of 0–20 m/s, the improved method reduces the Root Mean Square Error (RMSE) of retrieved wind speeds from 1.52 to 1.34 m/s, and enhances the correlation coefficient from 0.86 to 0.90; while in the range of 20–30 m/s, the RMSE decreases from 8.07 to 4.06 m/s, and the correlation coefficient increases from 0.04 to 0.45. Interestingly, the SNR observations are moderately correlated with marine gravities, showing correlation coefficients of 0.5–0.6, which may provide a useful reference for marine gravity retrieval using GNSS-R in the future. Keywords: CYGNSS, Geographical differences, Ocean wind speed, GNSS reflectometry, Marine gravit With a high temporal resolution and wide cover- age, Global Navigation Satellite System Reflectometry (GNSS-R) has developed as a valid remote sensing tech- nique for retrieving earth surface geophysical parameters over the past few decades (Garrison and Katzberg, 2000). Moreover, the implementation of Low Earth Orbit (LEO) satellite missions for GNSS-R, e.g., the United Kingdom- Disaster Monitoring Constellation (UK-DMC) (Gleason, 2006), TechDemoSat-1 (TDS-1) (Foti et al., 2015), and Cyclone Global Navigation Satellite System (CYGNSS), has further enhanced the advantages of temporal reso- lution and coverage (Ruf et al., 2013, 2016). These mis- sions have demonstrated the potential for monitoring ocean and land parameters, e.g., Significant Wave Height (SWH) (Roggenbuck et al., 2019), sea ice extent (Yan and Huang 2016), ocean wind speed (Hammond et al., 2020), soil moisture (Arroyo et al., 2014; Pan et al., 2020), above- ground biomass (Carreno-Luengo et al., 2020; Chen Wind speed retrieval using GNSS‑R technique with geographic partitioning Zheng Li, Fei Guo, Fade Chen, Zhiyu Zhang and Xiaohong Zhang https://satellite-navigation.springeropen.com/ Open Access Open Access ORIGINAL ARTICLE Introductionh The change of sea surface wind speed is one of the impor- tant factors affecting the marine environment. Timely and accurate monitoring of sea surface wind speed is of significance for maritime navigation safety and contrib- utes to improving our understanding of the marine cli- mate environment. Traditional wind speed monitoring mainly uses ground-based stations, buoys, and mete- orological remote sensing satellites like scatterometers and altimeters, etc. However, these methods suffer from limited measurement range and coverage, high cost, and high power consumption. Correspondence: Fei Guo fguo@whu.edu.cn School of Geodesy and Geomatics, Wuhan University, Wuhan 430079, China Li et al. Satellite Navigation (2023) 4:4 https://doi.org/10.1186/s43020-022-00093-z Li et al. Satellite Navigation (2023) 4:4 https://doi.org/10.1186/s43020-022-00093-z Satellite Navigation https://satellite-navigation.springeropen.com/ Li et al. Satellite Navigation (2023) 4:4 Li et al. Satellite Navigation Page 2 of 15 Page 2 of 15 differences is performed. The results show that when the wind speed is the same, the observables in different sea areas are not the same, and the same observables show different wind speeds in different sea areas. This geographical difference is the result of a combination of many influencing factors. As mentioned above, it is difficult to analyze all factors individually and exclude their influence on observables. However, simply train- ing all observables and wind speeds without eliminating the influence of other factors will inevitably lead to large retrieval errors. To solve this problem, the sea surface is divided into different areas for independent wind speed retrieval. Unlike the previous method, in each area the effect of these factors on the observables is basically the same, which can effectively reduce the retrieval error caused by these factors. Furthermore, in order to cor- rect the error caused by the low occurrence of the high wind samples, the random training samples and the resa- mpled samples are used for wind speed retrieval, respec- tively. Although this resampling method can improve the accuracy of high wind speed retrieval, the accuracy of low wind speed retrieval is reduced. To balance between the retrieval accuracies of high and low wind speeds, the results with the random training samples and the resam- pling samples are fused. The rest of this paper is arranged as follows. "Datasets and data filtering" section is the introduction of the experimental datasets and the data preprocessing. "The geographical difference analysis of CYGNSS observables" section analyzes the geographic differences of the observables and their relationship with marine gravity. "Method" section describes the proposed wind speed retrieval method, and its evaluation is given in "Evaluation section. Finally, the concluding remarks and suggestions for future work are presented in "Con- clusions" section. et al., 2021), etc. The focus of this paper is on wind speed retrieval. In 1993, Martin-Neira first proposed the concept of sea surface altimetry using Global Positioning System (GPS) direct and reflected signals (Martin-Neira, 1993). In 2000, a theoretical framework was developed to estab- lish the relationship between the properties of reflected GPS signal and the sea surface roughness (Zavorotny & Voronovich, 2000). In 2015, an ocean wind speed retrieval algorithm based on Technology Demonstra- tion Satellite (TDS-1) data was proposed by Foti et al., which demonstrated the performance of spaceborne GNSS-R for low wind speed retrieval (Foti et al., 2015). In 2017, Foti et al. showed that the spaceborne GNSS-R data derived from TDS-1 had the capability of monitor- ing hurricanes. The results confirmed that GNSS-R sig- nals can detect the ocean condition changes due to very strong and near-surface ocean wind associated with hur- ricanes (Foti et al., 2017). ( ) In 2014, a Minimum Variance (MV) wind speed esti- mator based on five observables derived from Delay- Doppler Map (DDM) was developed (Clarizia et al., 2014). Due to its good performance, this wind speed estimator has been used in the Level 2 ocean surface wind speed data product of the CYGNSS mission (Clar- izia & Ruf, 2016). In 2021, an improved CYGNSS wind speed retrieval method that combines the retrievals from DDM observables using Particle Swarm Optimi- zation (PSO) algorithm is proposed (Guo et al., 2021). Moreover, several studies have also shown that input- ting multiple observables into the Artificial Neural Net- work (ANN) can improve the accuracy of CYGNSS wind speed retrieval (Reynolds et al., 2020; Li et al., 2021; Asgarimehr et al., 2022). Although these observables are very sensitive to wind speed, they also are affected by many other factors, e.g., topography, swell and degree of wave development, long gravity waves, etc. (Chen-Zhang et al., 2016; Gleason et al., 2020). These factors make the relationship between CYGNSS observables and ocean wind speed complex and exhibit interactions. To obtain accurate wind speed retrieval, it is necessary to use a large amount of data to find out as many factors as pos- sible affecting wind speed and establish the relationship between the observables and wind speed. Datasets and data filtering Datasetsh The CYGNSS is designed to retrieve ocean surface wind speed with a constellation of eight small satellites. In addition to the advantages of high resolution and wide coverage with mean revisit time of 7.2 h over 0.25° × 0.25° latitude–longitude grids (Morris and Ruf, 2017), CYG- NSS can also estimate sea surface wind speed under all precipitating conditions and over the full dynamic range of wind speeds experienced in a Tropical Cyclone (TC) (Ruf et al., 2016). In this paper, the CYGNSS L1 level 3.1 version data (https://podaac.jpl.nasa.gov/dataset/CYG- NSS_L1_V3.1) in the 3 months from July to September 2019 is used for wind speed retrieval.hi However, due to the complex and changeable marine environment, it’s difficult to find out all the factors. To establish the accurate relationship between the CYGNSS observables and ocean wind speed, the unknown factors that affect observables are considered as a whole in this paper. Their combined influence on the observables is used for further analysis. The reference wind speed is derived from the fifth gen- eration ECMWF (European Center for Medium-range Weather Forecasts) Atmospheric Reanalysis, (ERA5) In this paper, an investigation on improving CYG- NSS wind speed retrieval by considering geographical Page 3 of 15 (2023) 4:4 Li et al. Satellite Navigation (2023) 4:4 Li et al. Satellite Navigation wind speed data provided by ECMWF (https://cds.clima te.copernicus.eu/cdsapp#!/dataset/reanalysisera5-single- levels). ERA5 data is the fifth-generation reanalysis data of ECMWF, which provides grid wind speed data hourly with a global spatial resolution of 0.25°. The final refer- ence wind speeds are obtained from the eastward com- ponent and the northward component of the wind speed 10 m above the surface in ERA5. After data preprocess- ing, the ECWMF data is matched with the CYGNSS dataset by applying the following criteria: the distance is less than 20 km, and the time difference is less than 30 min. other hand, the sea surface roughness can also reflect the change of wind speed. However, the sea surface rough- ness is affected not only by the wind speed, but also by other factors. Due to the influence of other factors, the relationship between the CYGNSS observables and the wind speed is complex and not an ideal one-to-one cor- respondence. Figure 1 shows the relationship between these CYGNSS observables and the wind speeds. Data filtering To ensure the quality of the CYGNSS data, the observa- bles must be checked for any anomaly before ocean wind speed retrieval. The initial quality control is mainly based on the following criteria (Asgarimehr et al., 2022; Li et al., 2021; Ruf & Balasubramaniam, 2019): (1) The receiver antenna gain in the direction of the specular point (sp_rx_gain) is larger than 0 dBi;h (2) The Normalized Bistatic Radar Cross Section (NBRCS) (ddm_nbrcs) is larger than 0; (3) The Leading Edge Slope (LES) (ddm_les) is larger than 0; (3) The Leading Edge Slope (LES) (ddm_les) is larger than 0; (4) The Signal Noise Ratio (SNR) (ddm_snr) is larger than 0 dB; (4) The Signal Noise Ratio (SNR) (ddm_snr) is larger than 0 dB; (5) The incident angle of all observables is less than 35°; h (6) Data with good overall quality, indicated by the quality_flags. Datasets and data filtering Datasetsh In order to reduce the influence of other potential factors on roughness as much as possible, these effects are treated as a whole in geographical analysis. g g p y To analyze the effect of geographical location on the CYGNSS observables, the observables are compared in the following criteria: the ocean is divided into the areas with 5° equal intervals in longitude and latitude, respec- tively; the average value of the observables correspond- ing to the same wind speed on each area. Figure 2 shows the correspondence between the wind speed and the three types of CYGNSS observables in different regions, including NBRCS, LES, and SNR. Top-left: The curve of NBRCS along longitude. Top-right: The curve of NBRCS along latitude. Middle-left: The curve of LES along lon- gitude. Middle-right: The curve of LES along latitude. Bottom-left: The curve of SNR along longitude. Bottom- right: The curve of SNR along latitude. It can be seen from Fig. 2 that the corresponding observables of the same wind speed at different geographic areas are different, and the changing trend of the observables corresponding to the different wind speeds is very similar. It should be noted that the observables of different wind speeds in the SNR analysis chart have multiple intersections, while less for LES and the least for NBRCS. This is mainly caused by the different sensitivity of the three types of observa- bles to wind speed. The sensitivity of the three types of observables to wind speed is NBRCS > LES > SNR, which is consistent with the previous finding (Li et al., 2021). In addition, Fig. 2 also tells that the higher the wind speed, the more intersection points will be. The reason The geographical difference analysis of CYGNSS observables Previous studies show that the change of sea surface roughness is mainly caused by the sea surface wind (Foti et al., 2017; Zavorotny & Voronovich, 2000). On the 25 20 15 10 5 ECMWF wind speed (m/s) 500 NBRCS 1000 0 25 20 15 10 5 50 100 150 200 LES 250 0 25 20 15 10 5 100 90 80 70 60 50 40 30 20 10 50 100 150 200 SNR (dB) Data density (%) 250 0 Fig. 1 The density map of the CYGNSS observables and ECMWF wind speed 25 20 15 10 5 ECMWF wind speed (m/s) 500 NBRCS 1000 0 0 25 20 15 10 5 100 90 80 70 60 50 40 30 20 10 50 100 150 200 SNR (dB) Data density (%) 250 0 Fig. 1 The density map of the CYGNSS observables and ECMWF wind speed Li et al. Satellite Navigation Page 4 of 15 may be the sensitivity of observables is reduced at high wind speeds. As the wind speed increases, the difference between the observables of different wind speeds gradu-h corresponding to the same observable will increase, which leads to an increase in retrieval error. Overall, Fig. The geographical difference analysis of CYGNSS observables 2 demonstrates that the observables LON 15 20 25 30 35 40 45 50 NBRCS 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s LAT 15 20 25 30 35 40 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s 3 4 5 6 7 8 9 10 LES 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s 3 4 5 6 7 8 9 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s 2 3 4 5 6 7 8 SNR 6 m/s 8 m/s 10 m/s 12 m/s 12 m/s 14 m/s 16 m/s 18 m/s 2 3 4 5 6 7 4 m/s 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 150°E, 155°E 40°S, 35°S 25°S, 20°S 10°S, 5°S 5°N, 10°N 20°N, 25°N 35°N, 40°N 90°E, 95°E 30°E, 35°E 30°W, 25°W 90°W, 85°W 150°W, 145°W LON LAT 150°E, 155°E 40°S, 35°S 25°S, 20°S 10°S, 5°S 5°N, 10°N 20°N, 25°N 35°N, 40°N 90°E, 95°E 30°E, 35°E 30°W, 25°W 90°W, 85°W 150°W, 145°W LON LAT 150°E, 155°E 40°S, 35°S 25°S, 20°S 10°S, 5°S 5°N, 10°N 20°N, 25°N 35°N, 40°N 90°E, 95°E 30°E, 35°E 30°W, 25°W 90°W, 85°W 150°W, 145°W Fig. The geographical difference analysis of CYGNSS observables 2 The variations of observables along longitude and latitude LON 15 20 25 30 35 40 45 50 NBRCS 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s 150°E, 155°E 90°E, 95°E 30°E, 35°E 30°W, 25°W 90°W, 85°W 150°W, 145°W LAT 15 20 25 30 35 40 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s 40°S, 35°S 25°S, 20°S 10°S, 5°S 5°N, 10°N 20°N, 25°N 35°N, 40°N 3 4 5 6 7 8 9 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s LAT 40°S, 35°S 25°S, 20°S 10°S, 5°S 5°N, 10°N 20°N, 25°N 35°N, 40°N 3 4 5 6 7 8 9 10 LES 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s 20 m/s LON 150°E, 155°E 90°E, 95°E 30°E, 35°E 30°W, 25°W 90°W, 85°W 150°W, 145°W LES 2 3 4 5 6 7 8 SNR 6 m/s 8 m/s 10 m/s 12 m/s 12 m/s 14 m/s 16 m/s 18 m/s LON 150°E, 155°E 90°E, 95°E 30°E, 35°E 30°W, 25°W 90°W, 85°W 150°W, 145°W 2 3 4 5 6 7 4 m/s 6 m/s 8 m/s 10 m/s 12 m/s 14 m/s 16 m/s 18 m/s LAT 40°S, 35°S 25°S, 20°S 10°S, 5°S 5°N, 10°N 20°N, 25°N 35°N, 40°N Fig. 2 The variations of observables along longitude and latitude corresponding to the same observable will increase, which leads to an increase in retrieval error. Overall, Fig. 2 demonstrates that the observables corresponding to the same wind speed vary with corresponding to the same observable will increase, which leads to an increase in retrieval error. may be the sensitivity of observables is reduced at high wind speeds. As the wind speed increases, the difference between the observables of different wind speeds gradu- ally decreases. This indicates that the wind speed range Overall, Fig. 2 demonstrates that the observables corresponding to the same wind speed vary with Li et al. The geographical difference analysis of CYGNSS observables 3 The comparison between marine gravity and CYGNSS observables in different locations LON 4 5 6 7 8 SNR (dB) −80 −45 −10 25 60 LAT 5.0 5.5 6.0 6.5 7.0 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) 180°W 60°E 0° 120°E 180°E 120°W 60°W 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N LON 4 5 6 7 8 SNR (dB) −80 −45 −10 25 60 LAT 5.0 5.5 6.0 6.5 7.0 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) 4 6 8 10 12 14 LES −80 −45 −10 25 60 6 8 10 12 14 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) 60 60 55 10 180°W 60°E 0° 120°E 180°E 120°W 60°W 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N LON LAT 180°W 60°E 0° 120°E 180°E 120°W 60°W 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N Marine gravity (mGal) LON LAT 4 6 8 10 12 14 LES −80 −45 −10 25 60 6 8 10 12 14 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) LON LAT 180°W 60°E 0° 120°E 180°E 120°W 60°W 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N 4 6 8 10 12 14 LES −80 −45 −10 25 60 LON 180°W 60°E 0° 120°E 180°E 120°W 60°W 6 8 10 12 14 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) LAT 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N Marine gravity (mGal) 20 30 40 50 60 NBRCS −80 −45 −10 25 60 LON 180°W 60°E 0° 120°E 180°E 120°W 60°W 30 35 40 45 50 55 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) LAT 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°Nf Fig. 3 The comparison between marine gravity and CYGNSS observables in different locations cause different roughness, thus keeping the CYGNSS observables in one-to-one correspondence with wind speed. However, the sea surface roughness may also be influenced by internal factors, such as marine gravity, which will counteract some of the wind effect on the wave. Therefore, when the wind speed is constant, the cause different roughness, thus keeping the CYGNSS observables in one-to-one correspondence with wind speed. However, the sea surface roughness may also be influenced by internal factors, such as marine gravity, which will counteract some of the wind effect on the wave. The geographical difference analysis of CYGNSS observables Satellite Navigation Page 5 of 15 LON 4 5 6 7 8 SNR (dB) −80 −45 −10 25 60 LAT 5.0 5.5 6.0 6.5 7.0 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) 4 6 8 10 12 14 LES −80 −45 −10 25 60 6 8 10 12 14 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) 20 30 40 50 60 NBRCS −80 −45 −10 25 60 30 35 40 45 50 55 −20 −15 −10 −5 0 5 10 Marine gravity (mGal) 180°W 60°E 0° 120°E 180°E 120°W 60°W 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N LON LAT 180°W 60°E 0° 120°E 180°E 120°W 60°W 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N LON LAT 180°W 60°E 0° 120°E 180°E 120°W 60°W 40°S 0° 30°S 20°S 10°S 10°N 20°N 30°N 40°N Fig. The geographical difference analysis of CYGNSS observables 5 The CYGNSS wind speed retrieval performance before and after the addition of marine gravity 1.50 1.55 1.60 1.65 RMSE(m/s) No marine gravity added Marine gravity added 0.75 0.80 0.85 0.90 Correlation coefficient Wind speed range: 0 - 20 m/s 1.59 0.85 0.86 1.61 p g Fig. 5 The CYGNSS wind speed retrieval performance before and after the addition of marine gravity − 15 × 10–3 to 10 × 10–3 cm/s2 in the latitude direction and − 80 × 10–3 to 60 × 10–3 cm/s2 in the longitude direction, respectively. The results show that CYGNSS may respond to large-scale changes in ocean gravity to a certain extent. Fig. 4 The correlation between SNR and marine gravity sea surface in different gravity zones may produce dif- ferent roughness. It is noted that in Fig. 3, SNR has the highest corre- lation with marine gravity, but the trend of SNR with marine gravity appears anomalous in the range of longi- tude 80°W–40°W and 130°E–150°E. The cause for these anomalous regions needs further study. The correlation coefficients (R) between SNR and marine gravity except for the anomalous region are given in Fig. 4, where the red dots are the marine gravity and the blue dots are SNR at different wind speeds. Since the amount of data decreases rapidly with increasing wind speed, the data are divided into three ranges according to wind speed (0–5 m/s, 5–10 m/s, 10–25 m/s), and the normalized correlation coefficients are given separately. It can be seen from these plots that the correlation coefficient decreases with increasing wind speed. This can be due to a reduction in data quality and quantity at high wind speeds. To understand the possible impacts of marine grav- ity on the geographical differences, the comparison of CYGNSS observables with marine gravity is pre- sented. The marine gravity is derived from CryoSat-2 and Jason-1 (https://topex.ucsd.edu/pub/global_grav_ 1min/ grav 1 min/) (Garcia et al., 2014). The following criteria are used to match the marine gravity data with the CYGNSS dataset: distance less than 20 km, and the time difference is not considered due to the marine gravity remains stable over time. Figure 3 shows the correlation between CYGNSS observables and marine gravity along longitude and latitude when the wind speed is 5 m/s, respectively. The geographical difference analysis of CYGNSS observables Therefore, when the wind speed is constant, the geographic locations. This phenomenon is likely related to the influencing factors of sea surface rough- ness, including external and internal factors. Wind speed is the main external factor. The waves on the ocean surface become rougher as the sea surface wind increases. Ideally, different wind speeds would Page 6 of 15 Li et al. Satellite Navigation (2023) 4:4 − 15 × 10–3 to 10 × 10–3 cm/s2 in the latitude direction and − 80 × 10–3 to 60 × 10–3 cm/s2 in the longitude direction, respectively. The results show that CYGNSS may respond to large-scale changes in ocean gravity to 0.4 0.6 0.8 1.0 MG 0.2 0.4 0.6 0.8 1.0 SNR 0.2 0.4 0.6 0.8 1.0 SNR LON 0.2 0.3 0.4 0.5 0.6 0.7 SNR Wind speed: 0 - 5 m/s Wind speed: 5 - 10 m/s Wind speed: 10 - 25 m/s R = 0.56 R = 0.54 R = 0.64 40°W 20°W 20°E 40°E 60°E 80°E 100°E 120°E 0° Fig. 4 The correlation between SNR and marine gravity 1.50 1.55 1.60 1.65 RMSE(m/s) No marine gravity added Marine gravity added 0.75 0.80 0.85 0.90 Correlation coefficient 6.4 6.5 6.6 6.7 6.8 RMSE(m/s) 0 0.1 0.2 0.3 0.4 0.5 Correlation coefficient Wind speed range: 20 - 30 m/s Wind speed range: 0 - 20 m/s 1.59 0.85 0.86 6.67 6.51 0.23 0.38 1.61 Fig. 5 The CYGNSS wind speed retrieval performance before and after the addition of marine gravity Page 6 of 15 Li et al. Satellite Navigation 0.4 0.6 0.8 1.0 MG 0.2 0.4 0.6 0.8 1.0 SNR 0.2 0.4 0.6 0.8 1.0 SNR LON 0.2 0.3 0.4 0.5 0.6 0.7 SNR Wind speed: 0 - 5 m/s Wind speed: 5 - 10 m/s Wind speed: 10 - 25 m/s R = 0.56 R = 0.54 R = 0.64 40°W 20°W 20°E 40°E 60°E 80°E 100°E 120°E 0° Fig. 4 The correlation between SNR and marine gravity 1.50 1.55 1.60 1.65 RMSE(m/s) No marine gravity added Marine gravity added 0.75 0.80 0.85 0.90 Correlation coefficient 6.4 6.5 6.6 6.7 6.8 RMSE(m/s) 0 0.1 0.2 0.3 0.4 0.5 Correlation coefficient Wind speed range: 20 - 30 m/s Wind speed range: 0 - 20 m/s 1.59 0.85 0.86 6.67 6.51 0.23 0.38 1.61 Fig. The geographical difference analysis of CYGNSS observables All data are divided into intervals of 0.5° along longitude and latitude, and the average values of various types of data are calculated in each interval. The left panel shows the variation of CYGNSS observables and marine gravity along the longitude direction. The right panel shows the varia- tion of CYGNSS observables and marine gravity along latitude. It can be seen that the correlation between CYGNSS observables and marine gravity along lon- gitude is significantly higher than that along latitude, and the variations range of marine gravity is about Method As mentioned above, given a wind speed, its correspond- ing CYGNSS observables show regular regional differ- ences. Therefore, the elimination of the geographical differences can contribute to expressing the relation- ship between the wind speed and CYGNSS observa- bles. In addition, existing research has shown that the Page 7 of 15 Li et al. Satellite Navigation 0 0.5 1.0 NBRCS 0 0.5 1.0 LES LON 0 0.5 1.0 SNR 180°W 180°E 140°W 100°W 60°W 60°E 100°E 140°E 20°W 20°E Fig. 6 Geographic partitioning based on the change of CYGNSS observables 0 0.5 1.0 NBRCS 0 0.5 1.0 LES LON 0 0.5 1.0 SNR 180°W 180°E 140°W 100°W 60°W 60°E 100°E 140°E 20°W 20°E Fig. 6 Geographic partitioning based on the change of CYGNSS observables Fig. 6 Geographic partitioning based on the change of CYGNSS observables Table 1 The specific information of each area Area Longitude range Wind speed range (m/s) Area-1 180°–152°W 20.97 Area-2 152°E–123°W 22.80 Area-3 123°W–107°W 24.73 Area-4 107°W–86°W 22.65 Area-5 86°W–71°W 29.69 Area-6 71°W–59°W 28.48 Area-7 59°W–52°W 19.36 Area-8 52°W–42°W 32.58 Area-9 42°W–22°W 30.22 Area-10 22°W–4°W 24.31 Area-11 4°W–4°E 18.51 Area-12 4°E–12°E 18.32 Area-13 12°E–23°E 21.15 Area-14 23°E–33°E 20.24 Area-15 33°E–44°E 21.23 Area-16 44°E–68°E 29.10 Area-17 68°E–89°E 22.65 Area-18 89°E–104°E 23.62 Area-19 104°E–121°E 19.43 Area-20 121°E–129°E 30.65 Area-21 129°E–145°E 24.81 Area-22 145°E–180° 23.08 Table 1 The specific information of each area low occurrence of the high wind samples will lead to the underestimation of high wind speeds (Guo et al., 2022; Li et al., 2021). Focusing on these two problems, a CYGNSS wind speed retrieval method considering the geographic differences and the low occurrence of the high wind sam- ples is presented in this section. Partitioning strategyh The conventional method attenuates the effect of geo- graphical differences by adding latitude and longi- tude information. However, the improvement in wind speed retrieval performance is limited due to the non- monotonic geographic variation of observations along latitude and longitude. As described in Sect. 3, after controlling for wind speed, the CYGNSS observables show a moderate correlation with marine gravity, so an attempt is made to add marine gravity to the data- set to improve retrieval performance. Figure 5 shows the comparison of the CYGNSS wind speed retrieval performance before and after the addition of marine gravity, except for the anomalous region. After the addition of marine gravity, in the range of 0–20 m/s, the Root Mean Square Error (RMSE) of retrieval wind speeds decreases from 1.61 to 1.59 m/s, and the corre- lation coefficient increases from 0.85 to 0.86; while in the range of 20–30 m/s, the RMSE decreases from 6.67 to 6.51, and the correlation coefficient increases from 0.23 to 0.38. The marginal effect can be attributed to the fact that additional parameters, such as tempera- ture, salinity, and seawater density, may also affect sea surface roughness in addition to marine gravity (Liu et al., 2021). These factors are potential contributors to the multiple values of the CYGNSS observables with constant wind speed. Since there are too many possible factors which are interconnected, it is quite challeng- ing to remove each factor’s influence separately. Fortu- nately, the combined effect of these factors is relatively stable in terms of geographic distribution. Thus, it is possible to improve the performance of CYGNSS wind speed retrieval by geographic partitioning.l (2) dSNR = 10 log10 Smax Navg (2) where Smax is the maximum value (in raw counts) in a single DDM bin and Navg is the average per-bin raw noise counts. In order to improve the sensitivity of SNR to wind speed, the bottom noise in DDM is removed and given as: (3) dSNR′ = 10log10 Smax −Navg Navg = 10log10 Smax Navg −1 (3) According to the instruments and geometric effects provided in the bistatic radar equation, the SNR in CYG- NSS can be corrected to: (4) dSNRC = dSNR′ + 10log10 R2 t R2 r + 10log10 cos2θ −10log10DEIRP −10log10Gr (4) NBRCS LES SNR Rt Rr Incident EIRP Sp_theta_orbit Wind speed Input layer Hidden layer Output layer Lon Lat 16 16 32 32 Fig. Theory of GNSS‑R wind speed retrievalh The scattered GPS signal power from the ocean can be described as a function of geometric parameters and sea surface roughness, which can be expressed as (Voronovich and Zavorotny 2017): (1) Y τ, f 2 = PtGt2T 2 i (4π)3 AGr2(τ)S2 f R2 t R2r σ0dA (1) where Y τ, f 2 is the function of the time delay τ and the frequency offset f , representing the Global Navi- gation Satellite System (GNSS) power; Pt is the GNSS transmitter power; Gt is the GNSS antenna gain; λ is the carrier wavelength; Ti is the coherent integration time; Rt is the distance from the transmitter to the specular reflection point; Rr is the distance from the receiver to the specular reflection point; 2 and S2 are the compo- nents of the Woodward Ambiguity Function (WAF) in delay and delay Doppler frequency, respectively; Gr is the receiver antenna gain; dA is the surface element of where Y τ, f 2 is the function of the time delay τ and the frequency offset f , representing the Global Navi- gation Satellite System (GNSS) power; Pt is the GNSS transmitter power; Gt is the GNSS antenna gain; λ is the carrier wavelength; Ti is the coherent integration time; Rt is the distance from the transmitter to the specular reflection point; Rr is the distance from the receiver to the specular reflection point; 2 and S2 are the compo- nents of the Woodward Ambiguity Function (WAF) in delay and delay Doppler frequency, respectively; Gr is the receiver antenna gain; dA is the surface element of Page 8 of 15 Li et al. Satellite Navigation (2023) 4:4 Li et al. Satellite Navigation the scattering area A ; σ0 symbolizes the NBRCS, which is related to the roughness of the glistening zone. Here, σ0 can be used to retrieve ocean wind speed. where θ is the incident angle, DEIRP is the GPS effective isotropic radiated power, and dSNRC represents the cor- rected SNR. Although SNR, LES, and NBRCS have high sensitiv- ity to wind speed, Eq. (4) tells that the inclusion of other observables can improve the wind speed retrieval accu- racy. Therefore, this study selects NBRCS, LES, SNR, Rt, Rr, incident, EIRP, and Sp_theta_orbit to retrieve the wind speed. Selection of observables In addition to NBRCS, the observables derived from DDM commonly used for wind speed retrieval are the LES of Integrated Delay Waveforms (IDW) and SNR (Garrison et al., 2002). The DDM_SNR derived from CYGNSS data is computed by: Partitioning strategyh 7 Simplified visualization of the multilayer perceptron As can be seen from Fig. 2, the observations fluctuate greatly along longitude but remain much stable along latitude. Therefore, the sea surface is divided along lon- gitude according to the change of CYGNSS observa- bles. The observables in a divided area generally change monotonously with the longitude information, and the longitude and latitude information can better attenuate the effect of geographical differences. The partition- ing strategy is shown in Fig. 6, where the blue dots are the normalized CYGNSS observables at different wind Page 9 of 15 (2023) 4:4 Li et al. Satellite Navigation peeds, and the red dotted lines represent the regional oundaries. Finally, the sea surface is divided into 22 ndependent areas, and the specific information of each rea is shown in Table 1. Multilayer perceptron configuration Recently, due to the capacity of artificial neural n works in learning complex relationships and gene izing the results from the training data, it has beco a popular application tool. Previous studies h CYGNSS data ECMWF data Matched dataset Train sets Test1 sets Original trained models Resampling 22 Resampled trained models Fusion models Test1 sets wind_1 Test1 sets wind_2 Test2 sets wind_2 Test2 sets wind_1 Fusion models Test2 sets wind 60% 20% 20% No resampling Areadatasets Spatially and temporallymatching Partition Data processing Model training Model validation Test2 sets Original trained models Resampled trained models Fig. 8 The experimental structure and process CYGNSS data ECMWF data Matched dataset Train sets Test1 sets Original trained models Resampling 22 Resampled trained models Fusion models Test1 sets wind_1 Test1 sets wind_2 60% 20% 20% No resampling Areadatasets Spatially and temporallymatching Partition Data processing Model training ECMWF data CYGNSS data Spatially and temporallymatching Test1 sets wind_1 Original trained models No resampling Fusion models Test1 sets wind_2 Resampled trained models Model training Test2 sets wind_2 Test2 sets wind_1 Fusion models Test2 sets wind Model validation Test2 sets Original trained models Resampled trained models Fig. 8 The experimental structure and process Test2 sets wind_1 Original trained models Test2 sets wind Resampled trained models Test2 sets wind_2 Multilayer perceptron configuration Recently, due to the capacity of artificial neural net- works in learning complex relationships and general- izing the results from the training data, it has become a popular application tool. Previous studies have Multilayer perceptron configuration speeds, and the red dotted lines represent the regional boundaries. Partitioning strategyh Finally, the sea surface is divided into 22 independent areas, and the specific information of each area is shown in Table 1. Recently, due to the capacity of artificial neural net- works in learning complex relationships and general- izing the results from the training data, it has become a popular application tool. Previous studies have Li et al. Satellite Navigation (2023) 4:4 Page 10 of 15 Li et al. Satellite Navigation 30 25 20 15 10 5 Not partitioning Partitioning Not partitioning Partitioning Not partitioning Partitioning RMSE: 8.07 R: 0.04 RMSE: 7.22 R: 0.16 RMSE: 5.06 R: 0.35 RMSE: 1.52 R: 0.86 RMSE: 1.38 R: 0.89 RMSE: 1.52 R: 0.87 RMSE: 1.64 R: 0.85 30 25 20 15 10 5 a Not resampling b Resampling c Fusion 100 90 80 70 60 50 40 30 20 10 0 ECMWF wind speed (m/s) 30 20 10 0 30 20 Data density (%) 100 90 80 70 60 50 40 30 20 10 0 Data density (%) 100 90 80 70 60 50 40 30 20 10 0 Data density (%) 10 0 30 25 20 15 10 5 30 25 20 15 10 5 ECMWF wind speed (m/s) 30 20 10 0 30 20 10 0 30 25 20 15 10 5 30 25 20 15 10 5 ECMWF wind speed (m/s) 30 20 CYGNSS wind speed (m/s) CYGNSS wind speed (m/s) 10 0 30 20 10 0 RMSE: 4.18 R: 0.37 RMSE: 7.40 R: 0.09 RMSE: 4.06 R: 0.45 RMSE: 1.49 R: 0.87 RMSE: 1.35 R: 0.90 Fig. Implementation processh The experimental structure and process are displayed in Fig. 8. The CYGNSS observables data is first spatiotem- porally matched with the ECMWF wind. Then, the sea surface is divided into 22 areas according to the par- titioning strategy. In each area, 60% of the data is used as the training set, and the rest is for the Test1 and the Test2, 20% each. In order to weaken the underestimation of high wind speed data due to low incidence, the train- ing set is resampled for each area to increase the propor- tion of high wind speed data in the training set. Training sets before and after resampling are trained separately to obtain the original trained model and the resampled trained model. The two models are used to retrieve the test sets, respectively. The results show that the original model has excellent performance at low wind speeds, while the resampled model has better performance at medium and high wind speeds. To achieve the best per- formance of the retrieval results in the whole wind speed range, the two retrieval wind speeds are fused. For the Test1 set, the wind speeds obtained by the two mod- els are fed into the multilayer perceptron for training to obtain a fusion model that can accommodate both high and low wind speeds, and the model is validated using the Test2 set. It should be noted that when the trained mod- els remain unchanged, the performance characteristic of the retrieved wind speeds is also basically unchanged. Thus, the obtained fusion model will be universal. On the other hand, if the trained model changes, the fusion model needs to be retrained. Finally, the obtained model is evaluated. For the independent Test2 set, the original Fig. 10 Wind speed retrieval error curves of different methods successfully used artificial neural networks for GNSS- R wind speed retrieval (Reynolds et al., 2020; Li et al., 2021). Because Multi-Layer Perceptron (MLP) is one of the most common artificial neural networks, it is used for wind speed retrieval in this paper. MLP can automatically extract or explore some “rea- sonable solution rules” hidden in the data by learning from a large amount of sample data. MLP is a gradi- ent descent procedure that computes the value of the derivative in a very efficient way and modifies the weights according to a parameter known as “learn- ing rate” (Mart et al., 2006). Partitioning strategyh 9 Density scatterplots of ECMWF wind speed versus retrieved wind speed 30 25 20 15 10 5 Not partitioning RMSE: 8.07 R: 0.04 RMSE: 1.52 R: 0.86 a Not ECMWF wind speed (m/s) 30 20 10 0 30 25 20 15 10 5 Not partitioning Partitioning RMSE: 8.07 R: 0.04 RMSE: 5.06 R: 0.35 RMSE: 1.52 R: 0.86 RMSE: 1.38 R: 0.89 30 25 20 15 10 5 a Not resampling 100 90 80 70 60 50 40 30 20 10 0 ECMWF wind speed (m/s) 30 20 10 0 30 20 Data density (%) 10 0 Partitioning RMSE: 5.06 R: 0.35 RMSE: 1.38 R: 0.89 30 25 20 15 10 5 esampling 100 90 80 70 60 50 40 30 20 10 0 30 20 Data density (%) 10 0 Not partitioning RMSE: 7.22 R: 0.16 RMSE: 1.64 R: 0.85 b R 30 25 20 15 10 5 ECMWF wind speed (m/s) 30 20 10 0 Partitioning RMSE: 1.52 R: 0.87 ampling 100 90 80 70 60 50 40 30 20 10 0 Data density (%) 30 25 20 15 10 5 30 20 10 0 RMSE: 4.18 R: 0.37 Not partitioning Partitioning RMSE: 7.22 R: 0.16 RMSE: 1.52 R: 0.87 RMSE: 1.64 R: 0.85 b Resampling 100 90 80 70 60 50 40 30 20 10 0 Data density (%) 30 25 20 15 10 5 30 25 20 15 10 5 ECMWF wind speed (m/s) 30 20 10 0 30 20 10 0 RMSE: 4.18 R: 0.37 Partitioning 100 90 80 70 60 50 40 30 20 10 0 Data density (%) 30 25 20 15 10 5 CYGNSS wind speed (m/s) 30 20 10 0 RMSE: 4.06 R: 0.45 RMSE: 1.35 R: 0.90 Not partitioning 30 25 20 15 10 5 ECMWF wind speed (m/s) 30 20 CYGNSS wind speed (m/s) 10 0 RMSE: 7.40 R: 0.09 RMSE: 1.49 R: 0.87 Fig. 9 Density scatterplots of ECMWF wind speed versus retrieved wind speed (2023) 4:4 Page 11 of 15 Page 11 of 15 Li et al. Partitioning strategyh Satellite Navigation 5 10 15 20 ≥25 0 3 6 9 12 RMSE (m/s) Not partitioning Partitioning 2 4 6 8 10 12 1.0 1.5 2.0 5 10 15 20 ≥25 0 3 6 9 12 RMSE (m/s) 2 4 6 8 10 12 1 2 5 10 15 20 ≥25 Wind speed (m/s) 0 3 6 9 12 RMSE (m/s) 2 4 6 8 10 12 1.0 1.5 2.0 Not resamping Resamping Fusion Fig. 10 Wind speed retrieval error curves of different methods Not partitioning Each hidden node input ( dj net ) is then transformed through the non-linear transfer function to produce a hidden node output, Yj (Rani et al., 2014). In this study, the transfer function tanh is used and expressed as follows: Partitioning (6) Yj = f dj net = edj net −e−dj net edj net + e−dj net (6) The structure of the multilayer perceptron is shown in Fig. 7. Generally, the MLP wind speed retrieval perfor- mance of multi- hidden layer is always better than that of a single hidden layer, but adding more hidden MLP lay- ers does not necessarily improve model performance. On the contrary, too many hidden layers will greatly increase the training time and even cause over-fitting and reduce model performance. After testing, the trained neural network structure with four hidden layers (the first and second layers each contain 16 neurons, and the third and fourth layers each contain 32 neurons) can achieve the best performance. Implementation processh Structurally, MLP is usu- ally composed of three layers: input layer, hidden layer, and output layer. There are nodes in each layer, and the nodes in adjacent layers are connected by weights, but the nodes in each layer are independent of each other. The principle of the forward propagation of multilayer perceptron can be summarized as: (5) dj net = N i=0 WijXi (5) where X0 and W0j are the bias ( X0 = 1) and its weight, respectively. N represents the number of input nodes. Li et al. Satellite Navigation (2023) 4:4 Page 12 of 15 Li et al. Satellite Navigation 0.80 0.83 0.86 0.89 0.92 0.95 Correlation coefficient Conventional method Improved method Area-1 Area-2 Area-3 Area-4 Area-5 Area-6 Area-7 Area-8 Area-9 Area-10 Area-11 Area-12 Area-13 Area-14 Area-15 Area-16 Area-17 Area-18 Area-19 Area-20 Area-21 Area-22 Area number 0.8 1.1 1.4 1.7 2.0 2.3 RMSE (m/s) Conventional method Improved method Fig. 11 The retrieval performance comparison between the conventional method and the improved method for each area Area number Area number ig. 11 The retrieval performance comparison between the conventional method and the improved method for each area trained model and resampled trained model are used to obtain two retrieval wind speeds, and the fusion model is used to fuse the two retrieval wind speeds. Figure 10 shows the error curves of the retrieval results obtained with different methods. The retrieval results for the wind speeds above 25 m/s are merged together due to very few samples. It can be seen from the figure that the retrieval errors of the methods with partitioning are smaller than those of the methods without partitioning. In the range of 0–10 m/s, the retrieval errors of the unresampled methods are smaller than those of the resampled. When the wind speed is larger than 10 m/s, the retrieval errors of the resampled are smaller than those of the unresampled. After fusion, the retrieval errors are at a low level in the whole wind speed range. Evaluation Figure 9 shows the wind speed inversion results with dif- ferent methods, where (a), (b), and (c) are the non-resa- mpling method, resampling method, and fusion method, respectively. The results using the methods without par- titioning are shown on the left, and with partitioning on the right. It can be seen from the figure that the perfor- mances of the methods with partitioning are all better than that of the methods without partitioning. Compared with the unresampled method, the resampled method has better performance in the range of 20–30 m/s, with the RMSE reducing from 5.06 to 4.18 m/s and the R increasing from 0.35 to 0.37. However, in the range of 0–20 m/s, the RMSE increases from 1.38 to 1.52 m/s and the R decreases from 0.89 to 0.87. In contrast, the fused method has better performance at both high and low wind speeds. When the wind speed is less than 20 m/s, the RMSE is 1.34 m/s and the R is 0.90. When the wind speed is greater than 20 m/s, the RMSE is 4.06 m/s and the R is 0.45. Figure 11 shows the retrieval performances in dif- ferent areas using the conventional method and the improved method. The conventional method represents the unresampled method without partitioning, and the improved method represents the fusion method with partitioning. Compared with the conventional method, the improved method shows better performance in all areas, with the RMSEs decreasing from (1.26–2.14 m/s) to (1.09–1.75 m/s) and Rs increasing from (0.82–0.88) to (0.88–0.92). The specific retrieval performance in each area is given in Table 2. Page 13 of 15 Page 13 of 15 Li et al. Satellite Navigation (2023) 4:4 Li et al. Conclusions h In this paper, the impact of the geographical differences on CYGNSS observables is investigated. When the wind speeds are the same, it is found that the observables vary with the location. The factors that cause this difference are diverse and complex. Although latitude and longi- tude information is included in the conventional method, it cannot effectively reduce the errors caused by geo- graphic differences due to the non-monotonic changes of observables with respect to the location along latitude and longitude. Fortunately, the geographical differences of CYGNSS observables are regular and stable. In small areas, the fluctuations of CYGNSS observables become smooth and small. Therefore, an improved GNSS-R wind speed retrieval method is proposed, which divides the sea surface into several independent areas according to the geographical differences of CYGNSS observables. In each area, CYGNSS observables vary monotonically with longitude, and wind speeds are retrieved indepen- dently. In addition, to correct the error caused by the low occurrence of the high wind samples, the random train- ing samples and resampled training samples are used for wind speed retrieval in each area, respectively. Although this resampling method can improve the accuracy of high wind speeds retrieval, the accuracy of low wind speeds retrieval is reduced. To balance between the retrieval accuracies for high and low wind speeds, the results with the random training samples and the resampling samples are fused. Since the ANN model was trained and validated with the wind speeds from ECMWF, it is necessary to evaluate the wind speed retrieval performance with an independent source of wind speed measurements. Therefore, the wind speed measurements collected by the National Data Buoy Center (NDBC) buoys were also used as a reference. In order to reduce the impact of reflected signals from land contamination, the sites within 25 km from the coastline were excluded, and five buoy sites at different distances from the coast- line (34 km, 50 km, 75 km, 200 km and 380 km) were selected for comparison experiments. Evaluation Satellite Navigation Table 2 The wind speed retrieval performance in each area Area Conventional method Improved method RMSE (m/s) Correlation coefficient RMSE (m/s) Correlation coefficient Area-1 1.44 0.88 1.32 0.90 Area-2 1.42 0.87 1.28 0.89 Area-3 1.54 0.85 1.35 0.89 Area-4 1.46 0.87 1.28 0.90 Area-5 1.42 0.87 1.26 0.90 Area-6 1.39 0.85 1.17 0.90 Area-7 1.26 0.86 1.09 0.90 Area-8 1.46 0.86 1.27 0.90 Area-9 1.42 0.87 1.30 0.89 Area-10 1.50 0.84 1.31 0.88 Area-11 1.53 0.85 1.26 0.90 Area-12 1.57 0.87 1.25 0.91 Area-13 1.77 0.87 1.42 0.92 Area-14 2.14 0.82 1.75 0.88 Area-15 1.81 0.83 1.53 0.89 Area-16 1.69 0.86 1.44 0.90 Area-17 1.75 0.84 1.50 0.89 Area-18 1.63 0.86 1.38 0.90 Area-19 1.57 0.85 1.34 0.90 Area-20 1.73 0.85 1.53 0.88 Area-21 1.67 0.84 1.45 0.89 Area-22 1.49 0.85 1.33 0.89 retrieval of CYGNSS observables matched with buoy data is performed using the conventional model and the improved model, respectively. The results show that the wind speed retrieval performance of the improved model is better than that of the conventional model, with the RMSEs decreasing from (0.81–1.35 m/s) to (0.76–1.31 m/s) and Rs increasing from (0.67–0.90) to (0.68–0.91). Brief information on these buoys and their specific wind speed retrieval performance are given in Table 3. References Arroyo, A. A., Camps, A., Aguasca, A., Forte, G. F., & Onrubia, R. (2014). Dual- polarization GNSS-R interference pattern technique for soil moisture mapping. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 7(5), 1533–1544. https://doi.org/10.1109/JSTARS.2014. 2320792 Asgarimehr, M., Arnold, C., Weigel, T., Ruf, C., & Wickert, J. (2022). GNSS reflec- tometry global ocean wind speed using deep learning: Development and assessment of CYGNSSnet. Remote Sensing of Environment, 269, 112801. https://doi.org/10.1016/j.rse.2021.112801 112801. https://doi.org/10.1016/j.rse.2021.112801 Carrenoluengo, H., Luzi, G., & Crosetto, M. (2020). Above-ground biomass retrieval over tropical forests: A novel GNSS-R approach with CYGNSS. Remote Sensing, 12(9), 1368. https://doi.org/10.3390/rs12091368 Chen, F., Guo, F., Liu, L., & Nan, Y. (2021). An improved method for pan-tropical above-ground biomass and canopy height retrieval using CYGNSS. Remote Sensing, 13(13), 2491. https://doi.org/10.3390/rs13132491 Chen, F., Zhang, X., Guo, F., Zheng, J., Nan, Y., & Freeshahd, M. (2022). TDS-1 GNSS reflectometry wind geophysical model function response to GPS block types. Geo-Spatial Information Science. https://doi.org/10.1080/ 10095020.2021.1997076 As we observed, CYGNSS observables are moder- ately correlated with marine gravity in some regions. However, many unknown factors such as the physi- cal properties of seawater and the performance of the instrument may also affect the observables as shown in previous studies (Chen et al., 2022; Wang et al., 2021). Exploring these unknown factors can not only improve the performance of ocean wind speed retrieval but also provide a broader application prospect for GNSS-R. In addition, although the results obtained above show that the observables can only reflect the large-scale grav- ity variation in some regions, it can provide the possi- bility of using GNSS-R for marine gravity retrieval in the future. Of course, to obtain a better marine gravity retrieval using GNSS-R, it is necessary to exclude the influence of other factors. It is quite challenging to use GNSS-R for marine gravity retrieval and requires fur- ther research both experimentally and theoretically. Chen-Zhang, D. D., Ruf, C. S., Ardhuin, F., & Park, J. (2016). GNSS-R nonlocal sea state dependencies: Model and empirical verification. Journal of Geophysical Research Oceans, 12(11), 8379–8394. https://doi.org/10.1002/ 2016JC012308 Clarizia, M. P., & Ruf, C. S. (2016). Wind speed retrieval algorithm for the cyclone global navigation satellite system (CYGNSS) mission. IEEE Transactions on Geoscience and Remote Sensing, 54(8), 4419–4432. https://doi.org/10. 1109/TGRS.2016.2541343 Clarizia, M. P., Ruf, C. S., Jales, P., & Gommenginger, C. (2014). Spaceborne GNSS-R minimum variance wind speed estimator. Acknowledgements Garrison, J. L., & Katzberg, S. J. (2000). The application of reflected GPS signals to ocean remote sensing. Remote Sensing of Environment, 73(2), 175–187. https://doi.org/10.1016/S0034-4257(00)00092-4 g We thank NASA for providing the CYGNSS observation data and ECMWF for providing the wind speed products. Garrison, J. L., Komjathy, A., Zavorotny, V. U., & Katzberg, S. J. (2002). Wind speed measurement using forward scattered GPS signals. IEEE Transactions on Geoscience and Remote Sensing, 40(1), 50–65. https://doi.org/10.1109/36. 981349 Conclusions h The wind speed Table 3 Brief information about buoys and wind speed retrieval performance Table 3 Brief information about buoys and wind speed retrieval performance Buoy ID Lon (°E) Lat (°N) DIST (km) Number of matchups Conventional model Improved model RMSE (m/s) R RMSE (m/s) R 41009 − 81.18 28.50 34 986 1.35 0.82 1.31 0.83 41013 − 77.76 33.44 50 1602 1.35 0.90 1.27 0.91 42040 − 88.23 29.20 75 657 1.21 0.67 1.20 0.68 41010 − 78.48 28.87 200 422 0.81 0.89 0.76 0.90 51000 − 153.79 23.52 380 705 1.18 0.88 1.10 0.87 Page 14 of 15 Li et al. Satellite Navigation (2023) 4:4 Li et al. Satellite Navigation (2023) Compared with the conventional method, the improved method shows better performance at both low wind speeds and high wind speeds. In the range of 0–20 m/s, the RMSEs and Rs of the conventional method and the improved method are 1.52 m/s, 1.34 m/s and 0.86, 0.90, respectively In the range of 20–30 m/s, the RMSEs and Rs of the conventional method and the improved method are 8.07 m/s, 4.06 m/s and 0.04, 0.45, respectively. In all independent areas, the improved method shows better retrieval performance, with RMSEs decreasing from (1.26–2.14 m/s) to (1.09–1.75 m/s) and Rs increasing from (0.82–0.88) to (0.88–0.92). In conclu- sion, the results demonstrate that the improved method has a better performance for wind speed retrieval. Availability of data and materials The variational CYGNSS wind retrievals are available from the authors. Declarations p g Guo, W., Du, H., Guo, C., Southwell, B. C., Cheong, J. W., & Dempster, A. W. (2022). Information fusion for GNSS-R wind speed retrieval using statistically modified convolutional neural network. Remote Sensing of Environment. https://doi.org/10.1016/j.rse.2022.112934 Author contributions Conceptualization and Methodology: FG and ZL; Writing original draft: ZL and FG; Editing: FG, ZL, FC, ZZ, and XZ; Review: FG, ZL, FC, ZZ, and XZ; All authors read and approved the final manuscript. Gleason, S. (2006). Remote sensing of ocean, ice and land surfaces using bistati- cally scattered GNSS signals from low earth orbit. Ph.D. Dissertation, Univer- sity of Surrey, Guildford, UK. c o edge e ts We thank NASA for providing the CYGNSS observation data and ECMWF for providing the wind speed products. References IEEE Transactions on Geoscience and Remote Sensing, 52(11), 6829–6843. https://doi.org/10. 1109/TGRS.2014.2303831 Foti, G., Gommenginger, C., Jales, P., Unwin, M., & Rosello, J. (2015). Spaceborne GNSS reflectometry for ocean winds: first results from the UK Techdemo- sat-1 mission. Geophysical Research Letters. https://doi.org/10.1002/2015G L064204 Foti, G., Gommenginger, C., & Srokosz, M. (2017). First spaceborne GNSS- Reflectometry observations of hurricanes from the UK Techdemosat-1 mission. Geophysical Research Letters, 44(12), 12358–12366. https://doi. org/10.1002/2017GL076166 g Garcia, E. S., Sandwell, D. T., & Smith, W. (2014). Retracking CryoSat-2, Envisat and Jason-1 radar altimetry waveforms for improved gravity field recov- ery. Geophysical Journal International, 196(3), 1402–1422. https://doi.org/ 10.1093/gji/ggt469 Funding Thi k This work was supported by the National Natural Science Foundation of China (Grant No. 42074029), the Fund for Creative Research Groups of China (Grant No. 41721003), and the Natural Science Foundation of Hubei Province for Distinguished Young Scholars (Grant No. 2021CFA039). Gleason, S., Johnson, J., Ruf, C., & Bussy-Virat, C. (2020). Characterizing back- ground signals and noise in spaceborne GNSS reflection ocean observa- tions. IEEE Geoscience and Remote Sensing Letters, 17(4), 587–591. https:// doi.org/10.1109/LGRS.2019.2926695 Guo, W., Du, H., Cheong, J. W., Southwell, B. J., & Dempster, A. G. (2021). GNSS-R wind speed retrieval of sea surface based on particle swarm optimization algorithm. IEEE Transactions on Geoscience and Remote Sensing, 99, 1–14. https://doi.org/10.1109/TGRS.2021.3082916 Availability of data and materials Competing interests The authors declare that they have no competing interests. Hammond, M. L., Foti, G., Gommenginger, C., & Srokosz, M. (2020). Temporal variability of GNSS-Reflectometry ocean wind speed retrieval perfor- mance during the UK Techdemosat-1 mission. Remote Sensing of Environ- ment, 242, 111744. https://doi.org/10.1016/j.rse.2020.111744 Received: 12 September 2022 Accepted: 27 December 2022 Received: 12 September 2022 Accepted: 27 December 2022 Page 15 of 15 Li et al. Satellite Navigation (2023) 4:4 Li et al. Satellite Navigation (2023) Li, X., Yang, D., Yang, J., Zheng, G., & Li, W. (2021). Analysis of coastal wind speed retrieval from CYGNSS mission using artificial neural network. Remote Sensing of Environment, 260, 112454. https://doi.org/10.1016/j.rse.2021. 112454 Liu, B., et al. (2021). First assessment of CYGNSS-incorporated SMAP sea surface salinity retrieval over pan-Tropical Ocean. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 14, 12163–12173. https:// doi.org/10.1109/JSTARS.2021.3128553 Mart, R., El-Fallahi, A., & Lasdon, L. (2006). Path relinking and GRG for artificial neural networks. European Journal of Operational Research, 169(2), 508–519. https://doi.org/10.1016/j.ejor.2004.08.012 Martin-Neira, M. (1993). A passive reflectometry and interferometry system (PARIS): Application to ocean altimetry. ESA Journal, 17(4), 331–355. Morris, M., & Ruf, C. S. (2017). Determining tropical cyclone surface wind speed structure and intensity with the CYGNSS satellite constellation. Journal of Applied Meteorology and Climatology, 56(7), 1847–1865. https://doi.org/10. 1175/JAMC-D-16-0375.1 Pan, Y., Ren, C., Liang, Y., Zhang, Z., & Shi, Y. (2020). Inversion of surface vegeta- tion water content based on GNSS-IR and MODIS data fusion. Satellite Navigation, 1(1), 21. https://doi.org/10.1186/s43020-020-00021-z Rani, B., Srinivas, K., & Govardhan, A. (2014). Rainfall prediction with TLBO opti- mized ANN. Journal of Scientific and Industrial Research, 73, 643–647. Reynolds, J., Clarizia, M.P., & Santi, E. (2020). Wind speed estimation from CYGNSS using artificial neuralnetworks. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 13, 708–716. https://doi. org/10.1109/JSTARS.2020.2968156 Roggenbuck, O., Reinking, J., & Lambertus, T. (2019). Determination of signifi- cant wave heights using damping coefficients of attenuated GNSS SNR data from static and kinematic observations. Remote Sensing, 11(4), 409. https://doi.org/10.3390/rs11040409 Ruf, C., Posselt, D., Majumdar, S., Gleason, S., & Morris, M. (2016). CYGNSS hand- book. Michigan Publishing Services. Ruf, C. S., et al. (2016). New ocean winds satellite mission to probe hurricanes and tropical convection. Bulletin of the American Meteorological Society. https://doi.org/10.1175/BAMS-D-14-00218.1 Ruf, C., & Balasubramaniam, R. (2019). Development of the CYGNSS geophysi- cal model function for wind speed. Competing interests IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(1), 66–77. https://doi. org/10.1109/JSTARS.2018.2833075 Ruf, C., Gleason, S., Jelenak, Z., Katzberg, S., Ridley, A., Rose, R., et al. (2013). The NASA EV-2 Cyclone Global Navigation Satellite System (CYGNSS) mission. IEEE Aerospace Conference. https://doi.org/10.1109/AERO.2013.6497202 Voronovich, A. G., & Zavorotny, V. U. (2017). Bistatic radar equation for signals of opportunity revisited. IEEE Transactions on Geoscience and Remote Sensing, 56(4), 1959–1968. https://doi.org/10.1109/TGRS.2017.2771253 Wang, T., Ruf, C. S., Gleason, S., O’Brien, A. J., & Russel, A. (2021). Dynamic cali- bration of GPS effective isotropic radiated power for GNSS-reflectometry earth remote sensing. IEEE Transactions on Geoscience and Remote Sensing, 99, 1–12. https://doi.org/10.1109/TGRS.2021.3070238 Yan, Q., & Huang, W. (2016). Spaceborne GNSS-R sea ice detection using delay- doppler maps: First results from the UK Techdemosat-1 mission. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 9(10), 4795–4801. https://doi.org/10.1109/JSTARS.2016.2582690 Zavorotny, V. U., & Voronovich, A. G. (2000). Scattering of GPS signals from the ocean with wind remote sensing application. IEEE Transactions on Geoscience and Remote Sensing, 38(2), 951–964. https://doi.org/10.1109/ 36.841977 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations.
https://openalex.org/W2900969499	https://europepmc.org/articles/pmc6265410?pdf=render	English	null	Therapeutic Potential of Extracellular Vesicles for Demyelinating Diseases; Challenges and Opportunities	Frontiers in molecular neuroscience	2,018	cc-by	7,160	Challenges and Opportunities Iñaki Osorio-Querejeta1,2, Ainhoa Alberro1, Maider Muñoz-Culla1,2, Imre Mäger3,4 and David Otaegui1,2* Iñaki Osorio-Querejeta1,2, Ainhoa Alberro1, Maider Muñoz-Culla1,2, Imre Mäger3,4 and David Otaegui1,2* 1 Multiple Sclerosis Unit, Biodonostia Health Research Institute, San Sebastian, Spain, 2 Spanish Network of Multiple Sclerosis, Barcelona, Spain, 3 Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford, United Kingdom, 4 Institute of Technology, University of Tartu, Tartu, Estonia Multiple Sclerosis is a demyelinating disease of the central nervous system for which no remyelination therapy is available and alternative strategies are being tested. Extracellular vesicles (EVs) have emerged as players in physiological and pathological processes and are being proposed as therapeutic targets and mediators. More concretely, EVs have shown to be involved in myelination related processes such as axon-oligodendrocyte communication or oligodendrocyte precursor cell migration. In addition, EVs have been shown to carry genetic material and small compounds, and to be able to cross the Blood Brain Barrier. This scenario led scientists to test the ability of EVs as myelin regeneration promoters in demyelinating diseases. In this review we will address the use of EVs as remyelination promoters and the challenges and opportunities of this therapy will be discussed. Edited by: David Blum, INSERM U1172 Centre de Recherche Jean Pierre Aubert, France Keywords: remyelination, exosomes, myelin, multiple sclerosis, microRNAs, EAE, oligodendrocyte MINI REVIEW MINI REVIEW published: 23 November 2018 doi: 10.3389/fnmol.2018.00434 INTRODUCTION Reviewed by: Thierry Burnouf, Taipei Medical University, Taiwan Pablo Villoslada, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Spain Reviewed by: Thierry Burnouf, Taipei Medical University, Taiwan Pablo Villoslada, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Spain *Correspondence: David Otaegui david.otaegui@biodonostia.org Myelin is a membranous sheath produced by oligodendrocytes (Ols) in the central nervous system (CNS) that surrounds axons allowing the saltatory nerve impulse transmission. Moreover, myelin protects axons and contributes to the maintenance of its homeostasis. Myelin can be damaged in a physiological context, due to the normal aging process, but it can also be caused by pathological mechanisms. The latter scenario occurs in a wide variety of pathological situations, such as traumatic demyelination, leukodystrophies or multiple sclerosis (MS), being the last one the most common demyelinating disease. Although no speciﬁc auto-antigen has been identiﬁed yet, MS is considered a chronic autoimmune CNS disease that includes the breakdown of the Blood- Brain Barrier, inﬂammation, demyelination, oligodendrocyte loss, gliosis and axonal degeneration (Baecher-Allan et al., 2018). It is accepted that the activation of peripheral autoreactive eﬀector CD4+ T cells that migrate into the CNS attacking the myelin sheath is the main cause of MS appearance. Once in the CNS a second reactivation occurs in which other cells types such as B and CD8+ T cells of the adaptive immune response, together with natural killers and microglia cells of the innate immune system contribute to the disease causing oligodendrocyte destruction, myelin loss, and an imbalance of the homeostasis of axons (reviewed in Baecher-Allan et al., 2018). This imbalance causes axon damage and an ineﬃcient nerve impulse transmission. In the ﬁrst stages of the disease, myelin can be restored, recovering normal electrical signal transmission. This is a complex process named remyelination in which a dynamic combination of diﬀerent signaling Received: 08 June 2018 Accepted: 08 November 2018 Published: 23 November 2018 November 2018 \| Volume 11 \| Article 434 Citation: Osorio-Querejeta I, Alberro A, Muñoz-Culla M, Mäger I and Otaegui D (2018) Therapeutic Potential of Extracellular Vesicles for Demyelinating Diseases; Challenges and Opportunities. Front. Mol. Neurosci. 11:434. doi: 10.3389/fnmol.2018.00434 November 2018 \| Volume 11 \| Article 434 1 Frontiers in Molecular Neuroscience \| www.frontiersin.org Extracellular Vesicles and Remyelination Osorio-Querejeta et al. pathways and molecules such as growth factors, cytokines and chemokines are tightly regulated (Kuhlmann et al., 2008). To achieve remyelination, oligodendrocyte precursor cells (OPCs) need to (1) proliferate, (2) migrate into the lesions, and (3) diﬀerentiate to myelinating Ols that will generate new myelin (Miron et al., 2011). Nevertheless, with the progression of the disease this process tends to fail. It is not completely understood why remyelination capacity decreases with time, but a lack of OPCs, a poor migration of these cells or their impossibility to diﬀerentiate to Ols have been proposed (Franklin, 2002). In addition, it is increasingly recognized that age is not only a risk factor for neurodegeneration but also adversely inﬂuences regenerative processes and remyelination (Hampton et al., 2012). Moreover, some factors such as genetic background and diet are also involved in the reduction of the remyelination capacity (revised in Adamo, 2014). membrane. Microvesicles vary greatly in size, ranging generally from 0.3 to 1 µm in diameter; however, it must be noted that in many scenarios it can be diﬃcult to separate exosomes from microvesicles purely based on their size (Willms et al., 2018). Another type of membrane vesicles are apoptotic bodies, which are 1–5 µm in size and were described many years ago and have diﬀerent features to those derived from living cells (György et al., 2011). Currently, the generic term EV is used to refer to the complete set of secreted vesicles (Gould and Raposo, 2013). EVs play an essential role in indirect intercellular communication as their membrane, cytosolic proteins, lipids, metabolites and genetic material can be transferred between cells (Théry et al., 2001; Valadi et al., 2007). They can follow two diﬀerent ways of integration: by direct fusion with the plasma membrane or by endocytosis (Morelli et al., 2004; Montecalvo et al., 2012). Most cell types release EVs being secreted both in physiologic and pathogenic conditions. They can be isolated from many body ﬂuids, including plasma and cerebrospinal ﬂuid (CSF). Citation: EVs are involved in many biological processes, their capacity to regulate immune response and cell diﬀerentiation being the two most important processes in the context of this review (Robbins and Morelli, 2014). Moreover, EVs take part in the transmission of information across the CNS (Frühbeis et al., 2013b) and have been found to play a role in the regulation of synaptic activity (Fauré et al., 2006) and myelin sheath biogenesis (Marzesco et al., 2005; Bakhti et al., 2011), as well as in the repair of damaged neurons (Court et al., 2011). To avoid neurodegeneration and promote neuroprotection, as well as the restoration of the fast saltatory conduction, the generation of new myelin is of extreme importance. The promotion of remyelination might protect axons avoiding their degeneration and probably improving patients’ prognosis. Therefore, pharmaceutical companies and researchers that work in the ﬁeld are trying their best to develop new remyelination therapies. To this end, the replacement of the endogenous OPC population and the stimulation of endogenous OPCs to regenerate myelin are being analyzed, being diﬀerentiation a key point in this process (Hartley et al., 2014). Finally, targeting the immune system has been also pointed out as a therapeutic strategy to induce remyelination (Dombrowski et al., 2017; El Behi et al., 2017). The pathogenesis of several diseases has been shown to be linked to EVs, including cancer (Robbins and Morelli, 2014), neurodegenerative diseases (Basso and Bonetto, 2016; Thompson et al., 2016) and, of particular interest to this work, MS (Verderio et al., 2012; Sáenz-Cuesta et al., 2014a,b; Selmaj et al., 2017). The implication of EVs and their ability to carry messages from one cell to another suggests that the use of EVs as a drug delivery system or as a treatment, might be an interesting way of targeting and modulating the course of the disease. Moreover, the fact that EVs are able to cross the BBB makes them strong candidates for CNS disease therapy (Jan et al., 2017). An ideal therapy should be able to cross the BBB and reach the CNS, target OPCs and not other cell types and should have minimal side eﬀects. Owing to their natural capacity to aﬀect cell proliferation and diﬀerentiation, and their potential to cross BBB, extracellular vesicles (EVs) have emerged as highly promising candidates for the treatment of demyelinating diseases, as discussed in detail in the following sections. Frontiers in Molecular Neuroscience \| www.frontiersin.org WHY EXTRACELLULAR VESICLES? Intercellular communication is a key factor for the functioning and regulation of all biological processes. Apart from the two classical mechanisms – direct cell-to-cell communication and transfer of secreted soluble molecules –, in the last years extracellular vesicles (EVs) have been found to play a central role in intercellular communication. Several works have been published demonstrating the therapeutic potential of EVs. These works will be discussed in the following paragraphs and have been summarized in Table 1. In some demyelinating pathologies, such as MS, the immune system is responsible for the damage caused to myelin. In fact, all the available treatments for MS are immunomodulatory or immunoregulatory drugs that prevent autoimmune attacks on the myelin sheath. In this way, the ability of exosomes isolated from pregnant mice serum or human periodontal ligament stem cells-derived exosomes to reduce the clinical score of the Experimental Autoimmune Encephalomyelitis (EAE), an animal model of MS, has been addressed by inhibiting the immune response, and more concretely by dampening Th1 response (Williams et al., 2013; Rajan et al., 2016). In addition, Extracellular vesicles are membrane-bound particles secreted by cells. There are diﬀerent types of EVs and the most common classiﬁcation is based on their size and biogenesis (Raposo and Stoorvogel, 2013). EVs formed inside multivesicular bodies and released upon fusion of these bodies with the plasma membrane are called exosomes. Their main characteristic is to have a uniform size of between 30 and 150 nm, thus being the smallest EVs. On the other hand, those known as microvesicles, come from the evagination and direct budding from the plasma November 2018 \| Volume 11 \| Article 434 2 Extracellular Vesicles and Remyelination Osorio-Querejeta et al. TABLE 1 \| Summary of therapeutic potential of EVs for demyelinating diseases. Reference EVs type EVs Source Isolation method Principal experiment Route of administration Result Williams et al., 2013 Exosomes Virgin and pregnant mice serum Ultracentrifugation EAE Intravenous Stablished EAE supression. Rajan et al., 2016 Exosomes HPLSC culture supernatant ExoQuick TC EAE Intravenous Inmmunomodulation of EAE. Zhuang et al., 2011 Exosomes Glioblastoma culture supernatant Sequential centrifugation steps EAE Intranasal EAE inhibition. Frühbeis et al., 2013a Exosomes Oli-Neu cultures supernatant Sequential centrifugation steps Oligodendrocyte- neuron co-culture N/A Exosomes mediated communication. Krämer-Albers et al., 2007 Exosomes Primary oligodendrocytes culture supernatant Ultracentrifugation Oligodendrocyte culture N/A Exosomes contain PLP, MBP, MOG and CNP. WHY EXTRACELLULAR VESICLES? Bakhti et al., 2011 Exosomes Primary oligodendrocytes culture supernatant Sequential centrifugation steps Oligodendrocyte culture N/A Oligodendrocytes derived exosomes inhibit OPC differentiation. Kurachi et al., 2016 Extracellular vesicles MVECs culture supernatant ExoQuick TC Oligodendrocyte Precursor cell culture N/A OPCs survival, proliferation and motility. Otero-Ortega et al., 2017 Exosomes MSC culture supernatant miRCURY Exosomes Isolation Kit Subcortical ischemic stroke Intravenous Promotion of olifodendrocyte formation and remyelination. Pusic and Kraig, 2014 Exosomes Youth and Environmental Enriched rat serum ExoQuick TC Old rats Intranasal Enhanced myelin content. Pusic et al., 2016 Exosomes Environmental Enriched rat serum ExoQuick TC Demyeliantion hipocampal slice culture N/A Myelination increased and oxidative stress reduced. Doeppner et al., 2015 Extracellular vesicles MSC culture supernatant PEG precipitation method Ischemic stroke Intravenous Neuroprotection and neuroregeneration. Drommelschmidt et al., 2017 Extracellular vesicles MSC culture supernatant PEG precipitation method Perinatal brain induced inﬂammation Intraperitoneal Immunomodulation and reduction of micro- and astrogliosis. TABLE 1 \| Summary of therapeutic potential of EVs for demyelinating diseases. Even though the authors did not demonstrate the mechanism by which Ols regulate OPCs in an inhibitory way, these results reinforce the implication of exosomes in OPC diﬀerentiation, an essential step for myelination and remyelination. In a diﬀerent work, the ability of pregnant mice serum-derived exosomes to promote the traﬃcking of OPCs into lesions from EAE mice after intravenous administration was shown (Williams et al., 2013) emphasizing the implication of exosomes in myelination related processes. Even though the authors did not demonstrate the mechanism by which Ols regulate OPCs in an inhibitory way, these results reinforce the implication of exosomes in OPC diﬀerentiation, an essential step for myelination and remyelination. In a diﬀerent work, the ability of pregnant mice serum-derived exosomes to promote the traﬃcking of OPCs into lesions from EAE mice after intravenous administration was shown (Williams et al., 2013) emphasizing the implication of exosomes in myelination related processes. the intranasal administration of curcumin-loaded glioblastoma- derived exosomes to EAE animals ameliorated the clinical symptoms of the model. Although the mechanism of action is not clear, the induction of immune tolerance and the apoptosis of activated immune cells are postulated to be behind this process. This data demonstrate that exosomes could work as anti-inﬂammatory drug delivery vehicles (Zhuang et al., 2011). As was mentioned in the introduction, Ols are responsible for generating myelin that enwraps axons. November 2018 \| Volume 11 \| Article 434 Frontiers in Molecular Neuroscience \| www.frontiersin.org WHY EXTRACELLULAR VESICLES? The communication between Ols and axons is essential for the survival and functional maintenance of both. Interestingly, this communication between Ols and axons has been shown to be mediated by exosomes and, in addition, the interactions between Ols and axons might aﬀect the cargo of exosomes (Frühbeis et al., 2013a). Moreover, when the cargo of exosomes released by Ols was analyzed, researchers found that those vesicles contained high levels of myelin related proteins; more concretely PLP, MBP, MOG, and CNP (Krämer- Albers et al., 2007). This data was the ﬁrst evidence of the possible role that exosomes could be playing in myelination. In a more recent work, it was suggested that Ol-derived exosomes were able to inhibit the diﬀerentiation of OPCs (Bakhti et al., 2011). To analyze the role that EVs play in pathological systems, several models have been used. In a model of white matter infarction in rats, researchers demonstrated that EVs derived from microvascular endothelial cells (MVECs) were taken up by OPCs, inhibiting the apoptosis of OPCs and promoting survival, proliferation and motility of the cells. The authors demonstrate that those EVs contained microRNAs and adhesion molecules which were responsible for the shown eﬀects (Kurachi et al., 2016). Moreover, Mesenchymal Stem Cell-derived exosomes (MSC-Exs) have been shown to promote oligodendrocyte formation and remyelination in a model of subcortical ischemic stroke. After intravenous administration of MSC-Exs, authors November 2018 \| Volume 11 \| Article 434 Frontiers in Molecular Neuroscience \| www.frontiersin.org Frontiers in Molecular Neuroscience \| www.frontiersin.org 3 Extracellular Vesicles and Remyelination Osorio-Querejeta et al. were able to detect higher levels of MOG protein and more myelinated axons. Interestingly, the 2416 proteins detected in the exosomes and described to be involved in brain repair functions were suggested by the authors as mediators of the eﬀect (Otero- Ortega et al., 2017). However, it appears that EV loading with exogenous cargoes prior to intranasal administration is not always essential for therapeutic eﬀects in the CNS, as recently demonstrated in a status epilepticus mouse model. Unmodiﬁed human bone marrow derived MSC-Exs reduced neuron loss and inﬂammation in the hippocampus of treated mice, which more importantly led to preservation of memory function (Long et al., 2017). These properties of unmodiﬁed MSC-Exs for treating CNS disease are particularly interesting and promising for MS. WHY EXTRACELLULAR VESICLES? Given the trend toward replacing certain MSC cell therapies with EV based therapies, and the fact that a number of MSC cell therapies have been tested in Phase I/II clinical trials for treating MS as well (Heldring et al., 2015), it is likely that MSC EVs will gain further focus in the short term for targeting MS pathology as well. exosomes and described to be involved in brain repair functions were suggested by the authors as mediators of the eﬀect (Otero- Ortega et al., 2017). Furthermore, a work published in 2014 demonstrated that exosomes from young and environmentally enriched rats signiﬁcantly increased the myelin content, oligodendrocyte precursor and neuronal stem cell levels and reduced oxidative stress and astrogliosis in demyelinated hippocampal slice cultures (Pusic and Kraig, 2014; Pusic et al., 2016). They also tested the eﬀect of these exosomes in vivo by intranasal administration in aged rats, showing positive results in myelin generation. The authors related the exosome-derived pro-remyelination eﬀect to their cargo, suggesting that the presence of miR-219 was responsible for promoting remyelination (Pusic and Kraig, 2014). Another aspect of demyelinating diseases is that the lack of myelin wrapping axons might, if remyelination does not take place, induce the disruption of the axons and, therefore, neurodegeneration. Neuroprotection is a key factor which might improve patients’ outcome and increase their life quality. Regarding to this, mesenchymal stem cells derived EVs were shown to be eﬀective peripheral immunomodulators in models of traumatic brain injury after both intravenous or intraperitoneal administration, decreasing inﬂammation and increasing neuroprotection, angiogenesis and neurological function, opening therapeutic possibilities in which neuroprotection can be reinforced (Doeppner et al., 2015; Drommelschmidt et al., 2017). Furthermore, a work published in 2014 demonstrated that exosomes from young and environmentally enriched rats signiﬁcantly increased the myelin content, oligodendrocyte precursor and neuronal stem cell levels and reduced oxidative stress and astrogliosis in demyelinated hippocampal slice cultures (Pusic and Kraig, 2014; Pusic et al., 2016). They also tested the eﬀect of these exosomes in vivo by intranasal administration in aged rats, showing positive results in myelin generation. The authors related the exosome-derived pro-remyelination eﬀect to their cargo, suggesting that the presence of miR-219 was responsible for promoting remyelination (Pusic and Kraig, 2014). g g p gy In addition to the intranasal administration route, as described above, other local delivery options have shown eﬃcacy for EV based CNS therapies as well. WHY EXTRACELLULAR VESICLES? Unilateral direct brain infusion of glioblastoma derived exosomes, pre-loaded with hydrophobic siRNA, led to exosome-dependent bilateral Huntington mRNA silencing in the brain of treated mice (Didiot et al., 2016). Other therapeutic strategies not directly relying on drug delivery can be eﬃcient as well. Intracerebral neuroblastoma exosome administration to an Alzheimer disease mouse model reduced amyloid-β levels in the brain and lowered the associated synaptotoxicity, tapping thus into natural EV-mediated Aβ clearance pathways (Yuyama et al., 2014). Similar eﬀects were observed also when using primary neuron exosomes, the eﬀect being cell type speciﬁc as glial exosomes were less eﬃcient in the capture of amyloid-β (Yuyama et al., 2015). This is not surprising as the transport of exosomes to brain parenchyma can be speciﬁcally related to the presence of speciﬁc surface molecules such as folate receptor α (Grapp et al., 2013) as well as other EV related signatures that can, for example, mediate periphery-brain signaling in inﬂammation (Balusu et al., 2016). Another aspect of demyelinating diseases is that the lack of myelin wrapping axons might, if remyelination does not take place, induce the disruption of the axons and, therefore, neurodegeneration. Neuroprotection is a key factor which might improve patients’ outcome and increase their life quality. Regarding to this, mesenchymal stem cells derived EVs were shown to be eﬀective peripheral immunomodulators in models of traumatic brain injury after both intravenous or intraperitoneal administration, decreasing inﬂammation and increasing neuroprotection, angiogenesis and neurological function, opening therapeutic possibilities in which neuroprotection can be reinforced (Doeppner et al., 2015; Drommelschmidt et al., 2017). DELIVERY INTO THE CENTRAL NERVOUS SYSTEM To be able to use EVs as therapeutic biopharmaceuticals for treating MS, it is imperative to ensure that EVs will reach their target cells in the CNS. That can be achieved, for example, by delivering EVs directly to the brain, by using systemic injections, or by administering vesicles via intranasal route. The intranasal route can be eﬃcient for diﬀerent cell type derived EVs, including T-cell, ﬁbroblast and tumor derived exosomes (Zhuang et al., 2011). This delivery route not only leads to increased brain accumulation of exosomes, but more importantly, it also results in reduced inﬂammation in EAE animals if exosomes are loaded with therapeutic anti-inﬂammatory molecules, as was previously mentioned (Zhuang et al., 2011). The latter clearly underlines the potential of EVs for treating MS via the intranasal route, which is further supported by successful experiments conducted in the context of other CNS diseases such as Parkinson’s disease (PD). In a mouse model of PD, catalase-loaded macrophage exosomes reached the brain and provided antioxidant-mediated neuroprotection (Haney et al., 2015). Neuroprotection was also induced by curcumin loaded embryonic stem cell exosomes in an ischemia-reperfusion injury model (Kalani et al., 2016). Repeated treatments with curcumin loaded exosomes led to a reduction of inﬂammation and improved neurological score and restored the expression of several BBB proteins. In many cases, however, systemic rather than local therapeutic EV administration would be preferred for various reasons, including the safety of the treatment administration. Despite the fact that BBB is virtually impermeable to most molecules there is some evidence that unmodiﬁed exosomes can enter the brain to some extent (Yang et al., 2015), but brain exposure is signiﬁcantly increased when using certain brain targeting ligands such as the rabies glycoprotein derived RVG peptide (Wiklander et al., 2015). The brain targeting RVG peptide, even though the precise targeting mechanism has not been fully elucidated, led to increased brain delivery of siRNA when decorated on dendritic cell exosomes (Alvarez-Erviti et al., 2011). Using that strategy, it was possible to lower the levels of Bace1 on both mRNA and protein levels in the brains of wild type mice (Alvarez-Erviti et al., 2011), and in reduced level of α-synuclein mRNA in S129D α-Syn transgenic mice (Cooper et al., 2014). Frontiers in Molecular Neuroscience \| www.frontiersin.org CHALLENGES AND OPPORTUNITIES Two characteristic aspects of MS are inﬂammation and neurodegeneration. The inhibition of inﬂammation and the promotion of remyelination are postulated as two therapeutic November 2018 \| Volume 11 \| Article 434 Frontiers in Molecular Neuroscience \| www.frontiersin.org Frontiers in Molecular Neuroscience \| www.frontiersin.org 4 Extracellular Vesicles and Remyelination Osorio-Querejeta et al. FIGURE 1 \| Summary of proposed therapeutic approaches for demyelinating diseases. Established cell lines, donor- or patient-derived cells are isolated and grown. EVs can be loaded with drugs/small compounds, miRNAs/siRNAs and/or surface antibodies, which provide new options in remyelination therapy. The loading can be performed during the cell culture (endogenous loading) or once EVs are isolated (exogenous loading\reviewed in Vader et al., 2016). This might depend on the strategy and purpose of the therapy (Sutaria et al., 2017). Therapeutic EVs can be isolated by ultracentrifugation, differential centrifugation, immunoafﬁnity or size-exclusion chromatography (Lener et al., 2015). Finally, EVs could be administered to the patient intranasally, intraperitoneally or intravenously and again, this is something that will vary according to the therapeutic strategy. FIGURE 1 \| Summary of proposed therapeutic approaches for demyelinating diseases. Established cell lines, donor- or patient-derived cells are isolated and grown. EVs can be loaded with drugs/small compounds, miRNAs/siRNAs and/or surface antibodies, which provide new options in remyelination therapy. The loading can be performed during the cell culture (endogenous loading) or once EVs are isolated (exogenous loading\reviewed in Vader et al., 2016). This might depend on the strategy and purpose of the therapy (Sutaria et al., 2017). Therapeutic EVs can be isolated by ultracentrifugation, differential centrifugation, immunoafﬁnity or size-exclusion chromatography (Lener et al., 2015). Finally, EVs could be administered to the patient intranasally, intraperitoneally or intravenously and again, this is something that will vary according to the therapeutic strategy. ways to improve patients’ outcome. As it has been widely shown, EVs can play a role in both immunomodulation and remyelination. But, what is the future going to be like with EVs mediated MS therapy? (Figure 1). We consider that it might be a feasible treatment for MS acting as immunomodulatory agents and tissue repair mediator. In addition to the source, the isolation method is also a relevant aspect to be mentioned, as several methods can lead to diﬀerent EVs types. Although several eﬀectors has been made in order to standardize isolation techniques, there is still controversy (Gardiner et al., 2016). The First Thing to Consider The ﬁrst thing to consider is the source where EVs are isolated from. In this sense established cells lines or cell isolated from the patient or a compatible donor can be used to isolate EVs. Bioﬂuids such as plasma or urine are also an alternative. It is not clear which source is the most eﬃcient and each one has got detractors. In this sense and as explained above, several are the sources that have been used with promising results, opening a wide range of EVs origins to be used. However, we consider that EVs isolated from cell culture might be more reproducible and “easy to manage.” In this way, cell therapies derived EVs are suggested as strong candidates as disease treatment. The use of cell-free stem cell-based therapy decreases the risk of cell therapy maintaining the beneﬁcial eﬀect of those cells. As an example, Mesenchymal Stem Cell derived vesicles have been widely studied as therapeutic mediators for several diseases (Review in Börger et al., 2017; Phinney and Pittenger, 2017). Frontiers in Molecular Neuroscience \| www.frontiersin.org REFERENCES Didiot, M. C., Hall, L. M., Coles, A. H., Haraszti, R. A., Godinho, B. M. D. C., Chase, K., et al. (2016). Exosome-mediated delivery of hydrophobically modiﬁed siRNA for huntingtin mRNA silencing. Mol. Ther. 24, 1836–1847. doi: 10.1038/mt.2016.126 Adamo, A. M. (2014). Nutritional factors and aging in demyelinating diseases. Genes Nutr. 9:360. doi: 10.1007/s12263-013-0360-8 Alvarez-Erviti, L., Seow, Y., Yin, H., Betts, C., Lakhal, S., and Wood, M. J. A. (2011). Delivery of siRNA to the mouse brain by systemic injection of targeted exosomes. Nat. Biotechnol. 29, 341–345. doi: 10.1038/nbt.1807 Doeppner, T. R., Herz, J., Orgnes, A., Jana, S., Ludwig, A.-K., Radtke, S., et al. (2015). Extracellular vesicles improve post-stroke neuroregeneration and prevent postischemic immunosuppression. Stem Cells Tranl. Med. 4, 927–931. doi: 10.5966/sctm.2013-0118 Baecher-Allan, C., Kaskow, B. J., and Weiner, H. L. (2018). Multiple sclerosis: mechanisms and immunotherapy. Neuron 97, 742–768. doi: 10.1016/j.neuron. 2018.01.021 Dombrowski, Y., O’Hagan, T., Dittmer, M., Penalva, R., Mayoral, S. R., Bankhead, P., et al. (2017). Regulatory T cells promote myelin regeneration in the central nervous system. Nat. Neurosci. 20, 674–680. doi: 10.1038/nn.4528 Bakhti, M., Winter, C., and Simons, M. (2011). Inhibition of myelin membrane sheath formation by oligodendrocyte-derived exosome-like vesicles. J. Biol. Chem. 286, 787–796. doi: 10.1074/jbc.M110.190009 Drommelschmidt, K., Serdar, M., Bendix, I., Herz, J., Bertling, F., Prager, S., et al. (2017). Mesenchymal stem cell-derived extracellular vesicles ameliorate inﬂammation-induced preterm brain injury. Brain. Behav. Immun. 60, 220–232. doi: 10.1016/j.bbi.2016.11.011 Balusu, S., Van Wonterghem, E., De Rycke, R., Raemdonck, K., Stremersch, S., Gevaert, K., et al. (2016). Identiﬁcation of a novel mechanism of blood–brain communication during peripheral inﬂammation via choroid plexus-derived extracellular vesicles. EMBO Mol. Med. 8, 1162–1183. doi: 10.15252/emmm. 201606271 Dugas, J. C., Cuellar, T. L., Scholze, A., Ason, B., Ibrahim, A., Emery, B., et al. (2010). Dicer1 and miR-219 are required for normal oligodendrocyte diﬀerentiation and myelination. Neuron 65, 597–611. doi: 10.1016/j.neuron. 2010.01.027 Basso, M., and Bonetto, V. (2016). Extracellular vesicles and a novel form of communication in the brain. Front. Neurosci. 10:127. doi: 10.3389/fnins.2016. 00127 El Behi, M., Sanson, C., Bachelin, C., Guillot-Noël, L., Fransson, J., Stankoﬀ, B., et al. (2017). Adaptive human immunity drives remyelination in a mouse model of demyelination. Brain 140, 967–980. doi: 10.1093/brain/awx008 Börger, V., Bremer, M., Ferrer-Tur, R., Gockeln, L., Stambouli, O., Becic, A., et al. (2017). Mesenchymal stem/stromal cell-derived extracellular vesicles and their potential as novel immunomodulatory therapeutic agents. Int. J. Mol. Sci. 18:E1450. A Second Consideration A second consideration can be the use of non-modiﬁed or bioengineered vesicles. The use of non-modiﬁed EVs has shown promising results (Pusic and Kraig, 2014; Pusic et al., 2014). In fact, there are several clinical trials recruiting patients in which the ability of allogenic mesenchymal stem cell derived exosomes in acute ischemic stroke or the eﬀect of plasma derived exosomes on cutaneous wound healing will be addressed (NCT02565264, 2015; NCT03384433, 2017). Nevertheless, the modiﬁcation of the cargo of EVs by bioengineering techniques is an interesting and promising ﬁeld in EV-mediated therapies and we consider that it might be a more eﬀective treatment method. It has been proved that cells which are genetically modiﬁed to overexpress a concrete microRNA, release EVs enriched in that November 2018 \| Volume 11 \| Article 434 Frontiers in Molecular Neuroscience \| www.frontiersin.org Frontiers in Molecular Neuroscience \| www.frontiersin.org 5 Extracellular Vesicles and Remyelination Osorio-Querejeta et al. microRNA (Squadrito et al., 2014). In this sense, microRNAs have shown to be involved in the diﬀerentiation of OPCs; more concretely miR-138, miR-219 and miR-338 (Dugas et al., 2010; de Faria et al., 2012; Wang et al., 2017). The enrichment of those microRNAs in the cargo of EVs might induce OPC diﬀerentiation and therefore remyelination after demyelination. Vesicles can also be loaded with small compounds and drugs with anti- inﬂammatory eﬀects. In this sense, curcumin loaded exosomes demonstrated to induce neuroprotection (Kalani et al., 2016). We also propose that nowadays immunomodulatory drugs could also be loaded in exosomes in order to obtain a controlled and direct administration into the CNS. This therapeutic approach is of interest due to the immunological component of MS. Finally, EVs can be modiﬁed to express membrane receptors of the target cell, in this way increasing the uptake by the cell and decreasing non-speciﬁc bindings (Alvarez-Erviti et al., 2011). patients’ outcome. However, even if the implication of EVs in remyelination related processes has been addressed in several works, our knowledge about the therapeutic potential of EVs is just beginning and an exciting future is awaiting us. FUNDING Extracellular vesicles have demonstrated that they are key players in myelin regeneration and the applications that EVs could have in the stimulation of remyelination in pathological states are many. As we have mentioned previously, treatment to induce remyelination is still not available and the use of EVs is becoming a promising and feasible method to immunomodulate, induce myelin restoration, and in this way decreasing neurodegeneration and therefore, increasing IO-Q and AA are supported by a grant of the Department of Education of the Basque Government (PRE_2016_2_0042 and PRE_2017_2_0002, respectively). IO-Q was funded with an EMBO Short Term Fellowship (#7109). MM-C is supported by the Gipuzkoa Regional Council (DFG15/006) and the Basque Government (ELKARTEK 16/014). IO-Q and AA are supported by a grant of the Department of Education of the Basque Government (PRE_2016_2_0042 and PRE_2017_2_0002, respectively). IO-Q was funded with an EMBO Short Term Fellowship (#7109). MM-C is supported by the Gipuzkoa Regional Council (DFG15/006) and the Basque Government (ELKARTEK 16/014). AUTHOR CONTRIBUTIONS IO-Q wrote the sections “Introduction and Therapeutic Potential of EVs for Demyelinating Diseases” and had produced the table and the ﬁgure. AA wrote the section “Why Extracellular Vesicles?” IM wrote the section “Delivery into the Central Nervous System.” MM-C and DO supervised the work. All authors contributed to the section “Challenges and Opportunities.” REFERENCES Bethesda, MD: ClinicalTrials.gov. Available at: https: //clinicaltrials.gov/ct2/show/NCT02565264?te Gould, S. J., and Raposo, G. (2013). As we wait: coping with an imperfect nomenclature for extracellular vesicles. J. Extracell. Vesicles 2, 3–5. doi: 10.3402/ jev.v2i0.20389 NCT03384433 (2017). Allogenic Mesenchymal Stem Cell Derived Exosome in Patients With Acute Ischemic Stroke (NCT03384433). Bethesda, MD: ClinicalTrials.gov. Available at: https://clinicaltrials.gov/ct2/show/ NCT03384433 Grapp, M., Wrede, A., Schweizer, M., Hüwel, S., Galla, H. J., Snaidero, N., et al. (2013). Choroid plexus transcytosis and exosome shuttling deliver folate into brain parenchyma. Nat. Commun. 4:2123. doi: 10.1038/ncomms3123 Otero-Ortega, L., Laso-García, F., Gómez-de Frutos, M. D. C., Rodríguez- Frutos, B., Pascual-Guerra, J., Fuentes, B., et al. (2017). White matter repair after extracellular vesicles administration in an experimental animal model of subcortical stroke. Sci. Rep. 7:44433. doi: 10.1038/srep 44433 György, B., Szabó, T. G., Pásztói, M., Pál, Z., Misják, P., Aradi, B., et al. (2011). Membrane vesicles, current state-of-the-art: emerging role of extracellular vesicles. Cell. Mol. Life Sci. 68, 2667–2688. doi: 10.1007/s00018-011-0689-3 Hampton, D. W., Innes, N., Merkler, D., Zhao, C., Franklin, R. J. M., and Chandran, S. (2012). Focal immune-mediated white matter demyelination reveals an age-associated increase in axonal vulnerability and decreased remyelination eﬃciency. Am. J. Pathol. 180, 1897–1905. doi: 10.1016/j.ajpath. 2012.01.018 Phinney, D. G., and Pittenger, M. F. (2017). Concise review: MSC-derived exosomes for cell-free therapy. Stem Cells 35, 851–858. doi: 10.1002/stem.2575 Pusic, A. D., and Kraig, R. P. (2014). Youth and environmental enrichment generate serum exosomes containing miR-219 that promote CNS myelination. Glia 62, 284–299. doi: 10.1002/glia.22606 Haney, M. J., Klyachko, N. L., Zhao, Y., Gupta, R., Plotnikova, E. G., He, Z., et al. (2015). Exosomes as drug delivery vehicles for Parkinson’s disease therapy. J. Control. Release 207, 18–30. doi: 10.1016/j.jconrel.2015.03.033 Pusic, A. D., Pusic, K. M., Clayton, B. L. L., and Kraig, R. P. (2014). IFNγ-stimulated dendritic cell exosomes as a potential therapeutic for remyelination. J. Neuroimmunol. 266, 12–23. doi: 10.1016/j.jneuroim.2013.10. 014 Hartley, M. D., Altowaijri, G., and Bourdette, D. (2014). Remyelination and multiple sclerosis: therapeutic approaches and challenges. Curr. Neurol. Neurosci. Rep. 14:485. doi: 10.1007/s11910-014-0485-1 Pusic, K. M., Pusic, A. D., and Kraig, R. P. (2016). Environmental enrichment stimulates immune cell secretion of exosomes that promote CNS myelination and may regulate inﬂammation. Cell. Mol. Neurobiol. 36, 313–325. doi: 10.1007/ s10571-015-0269-4 Heldring, N., Mäger, I., Wood, M. J. A., Le Blanc, K., and Andaloussi, S. E. L. (2015). REFERENCES doi: 10.1016/j.jneuroim.2017.02.002 Lener, T., Gimona, M., Aigner, L., Börger, V., Buzas, E., Camussi, G., et al. (2015). Applying extracellular vesicles based therapeutics in clinical trials – An ISEV position paper. J. Extracell. Vesicles 4, 1–31. doi: 10.3402/jev.v4.30087 Squadrito, M. L., Baer, C., Burdet, F., Maderna, C., Gilﬁllan, G. D., Lyle, R., et al. (2014). Endogenous RNAs modulate MicroRNA sorting to exosomes and transfer to acceptor cells. Cell Rep. 8, 1432–1446. doi: 10.1016/j.celrep.2014.07. 035 Long, Q., Upadhya, D., Hattiangady, B., Kim, D.-K., An, S. Y., Shuai, B., et al. (2017). Intranasal MSC-derived A1-exosomes ease inﬂammation, and prevent abnormal neurogenesis and memory dysfunction after status epilepticus. Proc. Natl. Acad. Sci. U.S.A. 114, E3536–E3545. doi: 10.1073/pnas.170392 0114 Sutaria, D. S., Badawi, M., Phelps, M. A., and Schmittgen, T. D. (2017). Achieving the promise of therapeutic extracellular vesicles: the devil is in details of therapeutic loading. Pharm. Res. 34, 1053–1066. doi: 10.1007/s11095-017- 2123-5 Marzesco, A.-M., Janich, P., Wilsch-Bräuninger, M., Dubreuil, V., Langenfeld, K., Corbeil, D., et al. (2005). Release of extracellular membrane particles carrying the stem cell marker prominin-1 (CD133) from neural progenitors and other epithelial cells. J. Cell Sci. 118, 2849–2858. doi: 10.1242/jcs.02439 Théry, C., Boussac, M., Véron, P., Raposo, G., Garin, J., and Amigorena, S. (2001). Proteomic analysis of dendritic cell-derived exosomes: a secreted subcellular compartment distinct from apoptotic vesicles. J. Immunol. 166, 7309–7318. doi: 10.4049/jimmunol.166.12.7309 Miron, V. E., Kuhlmann, T., and Antel, J. P. (2011). Cells of the oligodendroglial lineage, myelination, and remyelination. Biochim. Biophys. Acta 1812, 184–193. doi: 10.1016/j.bbadis.2010.09.010 Thompson, A. G., Gray, E., Heman-Ackah, S. M., Mäger, I., Talbot, K., Andaloussi, S., et al. (2016). Extracellular vesicles in neurodegenerative disease – Pathogenesis to biomarkers. Nat. Rev. Neurol. 12, 346–357. doi: 10.1038/ nrneurol.2016.68 Montecalvo, A., Larregina, A. T., Shufesky, W. J., Stolz, D. B., Sullivan, M. L. G., Karlsson, J. M., et al. (2012). Mechanism of transfer of functional microRNAs between mouse dendritic cells via exosomes. Blood 119, 756–766. doi: 10.1182/ blood-2011-02-338004 Vader, P., Mol, E. A., Pasterkamp, G., and Schiﬀelers, R. M. (2016). Extracellular vesicles for drug delivery. Adv. Drug Deliv. Rev. 106, 148–156. doi: 10.1016/j. addr.2016.02.006 Valadi, H., Ekström, K., Bossios, A., Sjöstrand, M., Lee, J. J., and Lötvall, J. O. (2007). Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat. Cell Biol. 9, 654–659. doi: 10.1038/ ncb1596 Morelli, A. E., Larregina, A. T., Shufesky, W. J., Sullivan, M. L. REFERENCES Therapeutic potential of multipotent mesenchymal stromal cells and their extracellular vesicles. Hum. Gene Ther. 26, 506–517. doi: 10.1089/hum.2015.072 Rajan, T. S., Giacoppo, S., Diomede, F., Ballerini, P., Paolantonio, M., Marchisio, M., et al. (2016). The secretome of periodontal ligament stem cells from MS patients protects against EAE. Sci. Rep. 6:38743. doi: 10.1038/ srep38743 Jan, A. T., Malik, M. A., Rahman, S., Yeo, H. R., Lee, E. J., Abdullah, T. S., et al. (2017). Perspective insights of exosomes in neurodegenerative diseases: a critical appraisal. Front. Aging Neurosci. 9:17. doi: 10.3389/fnagi.2017.00317 Kalani, A., Chaturvedi, P., Kamat, P. K., Maldonado, C., Bauer, P., Joshua, I. G., et al. (2016). Curcumin-loaded embryonic stem cell exosomes restored neurovascular unit following ischemia-reperfusion injury. Int. J. Biochem. Cell Biol. 36, 1011–1014. doi: 10.1002/jmri.23741.Proton Raposo, G., and Stoorvogel, W. (2013). Extracellular vesicles: exosomes, microvesicles, and friends. J. Cell Biol. 200, 373–383. doi: 10.1083/jcb.201211138 Raposo, G., and Stoorvogel, W. (2013). Extracellular vesicles: exosomes, microvesicles, and friends. J. Cell Biol. 200, 373–383. doi: 10.1083/jcb.201211138 Robbins, P. D., and Morelli, A. E. (2014). Regulation of immune responses by extracellular vesicles. Nat. Rev. Immunol. 14, 195–208. doi: 10.1038/ nri3622 Robbins, P. D., and Morelli, A. E. (2014). Regulation of immune responses by extracellular vesicles. Nat. Rev. Immunol. 14, 195–208. doi: 10.1038/ nri3622 Krämer-Albers, E.-M., Bretz, N., Tenzer, S., Winterstein, C., Möbius, W., Berger, H., et al. (2007). Oligodendrocytes secrete exosomes containing major myelin and stress-protective proteins: trophic support for axons? Proteomics Clin. Appl. 1, 1446–1461. doi: 10.1002/prca.200700522 Sáenz-Cuesta, M., Irizar, H., Castillo-Triviño, T., Muñoz-Culla, M., Osorio- Querejeta, I., Prada, A., et al. (2014a). Circulating microparticles reﬂect treatment eﬀects and clinical status in multiple sclerosis. Biomark. Med. 8, 653–661. doi: 10.2217/bmm.14.9 Kuhlmann, T., Miron, V., Cui, Q., Cuo, Q., Wegner, C., Antel, J., et al. (2008). Diﬀerentiation block of oligodendroglial progenitor cells as a cause for remyelination failure in chronic multiple sclerosis. Brain 131, 1749–1758. doi: 10.1093/brain/awn096 Sáenz-Cuesta, M., Osorio-Querejeta, I., and Otaegui, D. (2014b). Extracellular vesicles in multiple sclerosis?: what are they telling us?? Front. Cell. Neurosci. 8:100. doi: 10.3389/fncel.2014.00100 Kurachi, M., Mikuni, M., and Ishizaki, Y. (2016). Extracellular vesicles from vascular endothelial cells promote survival, proliferation and motility of oligodendrocyte precursor cells. PLoS One 11:e0159158. doi: 10.1371/journal. pone.0159158 Selmaj, I., Mycko, M. P., Raine, C. S., and Selmaj, K. W. (2017). The role of exosomes in CNS inﬂammation and their involvement in multiple sclerosis. J. Neuroimmunol. 306, 1–10. REFERENCES doi: 10.3390/ijms18071450 Fauré, J., Lachenal, G., Court, M., Hirrlinger, J., Chatellard-Causse, C., Blot, B., et al. (2006). Exosomes are released by cultured cortical neurones. Mol. Cell. Neurosci. 31, 642–648. doi: 10.1016/j.mcn.2005.12.003 Cooper, J. M., Wiklander, P. B. O., Nordin, J. Z., Al-Shawi, R., Wood, M. J., Vithlani, M., et al. (2014). Systemic exosomal siRNA delivery reduced alpha- synuclein aggregates in brains of transgenic mice. Mov. Disord. 29, 1476–1485. doi: 10.1002/mds.25978 Franklin, R. J. M. (2002). Why does remyelination fail in multiple sclerosis? Nat. Rev. Neurosci. 3, 705–714. doi: 10.1038/nrn917 Frühbeis, C., Fröhlich, D., Kuo, W. P., Amphornrat, J., Thilemann, S., Saab, A. S., et al. (2013a). Neurotransmitter-triggered transfer of exosomes mediates oligodendrocyte-neuron communication. PLoS Biol. 11:e1001604. doi: 10.1371/ journal.pbio.1001604 Court, F. A., Midha, R., Cisterna, B. A., Grochmal, J., Shakhbazau, A., Hendriks, W. T., et al. (2011). Morphological evidence for a transport of ribosomes from schwann cells to regenerating axons. Glia 59, 1529–1539. doi: 10.1002/glia. 21196 Frühbeis, C., Fröhlich, D., Kuo, W. P., and Krämer-Albers, E.-M. (2013b). ll l l d f l ll Frühbeis, C., Fröhlich, D., Kuo, W. P., and Krämer-Albers, E.-M. (2013b). Extracellular vesicles as mediators of neuron-glia communication. Front. Cell. Neurosci. 7:182. doi: 10.3389/fncel.2013.00182 Gardiner, C., Vizio, D., Di, Sahoo, S., The, C., Witwer, K. W., et al. (2016). T h i d f th i l ti d h t i ti f t ll l i l Extracellular vesicles as mediators of neuron-glia communication. Front. Cell. Neurosci. 7:182. doi: 10.3389/fncel.2013.00182 de Faria, O., Cui, Q. L., Bin, J. M., Bull, S. J., Kennedy, T. E., Bar-Or, A., et al. (2012). Regulation of miRNA 219 and miRNA clusters 338 and 17-92 in oligodendrocytes. Front. Genet. 3:46. doi: 10.3389/fgene.2012.00046 Neurosci. 7:182. doi: 10.3389/fncel.2013.00182 Gardiner, C., Vizio, D., Di, Sahoo, S., The, C., Witwer, K. W., et al. (2016). Gardiner, C., Vizio, D., Di, Sahoo, S., The, C., Witwer, K. W., et al. (2016). h d f h l d h f ll l l Techniques used for the isolation and characterization of extracellular vesicles: Frontiers in Molecular Neuroscience \| www.frontiersin.org November 2018 \| Volume 11 \| Article 434 6 Extracellular Vesicles and Remyelination Osorio-Querejeta et al. results of a worldwide survey. J. Extracell. Vesicles 1, 1–6. doi: 10.3402/jev.v5. 32945 results of a worldwide survey. J. Extracell. Vesicles 1, 1–6. doi: 10.3402/jev.v5. 32945 NCT02565264 (2015). Eﬀect of Plasma Derived Exosomes on Cutaneous Wound Healing (NCT02565264). REFERENCES G., Stolz, D. B., Papworth, G. D., et al. (2004). Endocytosis, intracellular sorting, and processing of exosomes by dendritic cells. Blood 104, 3257–3266. doi: 10.1182/blood-2004- 03-0824 November 2018 \| Volume 11 \| Article 434 Frontiers in Molecular Neuroscience \| www.frontiersin.org Extracellular Vesicles and Remyelination Osorio-Querejeta et al. Verderio, C., Muzio, L., Turola, E., Bergami, A., Novellino, L., Ruﬃni, F., et al. (2012). Myeloid microvesicles are a marker and therapeutic target for neuroinﬂammation. Ann. Neurol. 72, 610–624. doi: 10.1002/ana. 23627 Yuyama, K., Sun, H., Sakai, S., Mitsutake, S., Okada, M., Tahara, H., et al. (2014). Decreased amyloid-β pathologies by intracerebral loading of glycosphingolipid- enriched exosomes in Alzheimer model mice. J. Biol. Chem. 289, 24488–24498. doi: 10.1074/jbc.M114.577213 Wang, H., Moyano, A. L., Ma, Z. Z., Deng, Y., Lin, Y., Zhao, C., et al. (2017). miR-219 cooperates with miR-338 in myelination and promotes myelin repair in the CNS. Dev. Cell 40, 566.e5–582.e5. doi: 10.1016/j.devcel.2017. 03.001 Yuyama, K., Sun, H., Usuki, S., Sakai, S., Hanamatsu, H., Mioka, T., et al. (2015). A potential function for neuronal exosomes: sequestering intracerebral amyloid-B peptide. FEBS Lett. 589, 84–88. doi: 10.1016/j.febslet.2014.11.027 Zhuang, X., Xiang, X., Grizzle, W., Sun, D., Zhang, S., Axtell, R. C., et al. (2011). Treatment of brain inﬂammatory diseases by delivering exosome encapsulated anti-inﬂammatory drugs from the nasal region to the brain. Mol. Ther. 19, 1769–1779. doi: 10.1038/mt.2011.164 Wiklander, O. P. B., Nordin, J. Z., O’Loughlin, A., Gustafsson, Y., Corso, G., Mäger, I., et al. (2015). Extracellular vesicle in vivo biodistribution is determined by cell source, route of administration and targeting. J. Extracell. Vesicles 4:26316. doi: 10.3402/jev.v4.26316 Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Williams, J. L., Gatson, N. N., Smith, K. M., Almad, A., McTigue, D. M., and Whitacre, C. C. (2013). Serum exosomes in pregnancy-associated immune modulation and neuroprotection during CNS autoimmunity. Clin. Immunol. 149, 236–243. doi: 10.1016/j.clim.2013.04.005 Willms, E., Cabañas, C., Mäger, I., and Wood, M. J. A. (2018). Extracellular vesicle heterogeneity: subpopulations, isolation techniques, and diverse functions in cancer progression. Front. Immunol. 9:738. doi: 10.3389/ﬁmmu.2018.00738 Copyright © 2018 Osorio-Querejeta, Alberro, Muñoz-Culla, Mäger and Otaegui. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). Frontiers in Molecular Neuroscience \| www.frontiersin.org November 2018 \| Volume 11 \| Article 434 REFERENCES The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Yang, T., Martin, P., Fogarty, B., Brown, A., Schurman, K., Phipps, R., et al. (2015). Exosome delivered anticancer drugs across the blood-brain barrier for brain cancer therapy in Danio Rerio. Pharm. Res. 32, 2003–2014. doi: 10.1007/s11095- 014-1593-y November 2018 \| Volume 11 \| Article 434 Frontiers in Molecular Neuroscience \| www.frontiersin.org 8
https://openalex.org/W2776144146	https://europepmc.org/articles/pmc5842820?pdf=render	English	null	Dentists’ opinions on knowledge, attitudes and barriers in providing oral health care to older people living independently in the Netherlands and Flanders (Belgium)	BDJ open	2,017	cc-by	8,242	INTRODUCTION differences between the two can also be reported. The number of inhabitants in the Netherlands is 16.8 millions of people, Flanders has 6.4 million inhabitants.9 Among other things, it appears that in Flanders there are relatively more practicing dentists. On the other hand, in the Netherlands, also oral hygienists are involved in providing professional oral health care. So, in the Netherlands oral health care is increasingly rendered in larger practices, in which various oral health care providers are collaborating.9,10 Further- more, there are differences with regard to oral health. The DMFT index, indicating the extent of oral decay through caries, is lower in older people of 65 and over in Flanders than it is in the Netherlands.9,11–14 Also, the ﬁnancing of oral health care for older people who live at home is different in both countries. In the Netherlands, in principle, these patients have to pay for the cost of oral health care themselves, although there are reimbursement possibilities depending on speciﬁc needs of care. In Flanders, it goes that patients will be reimbursed in all cases to a minimum of 75% of the cost.9 In many western countries, the proportion of older people within the population is on the increase and it is to be expected that this tendency will further continue in the years to come.1 It is known that becoming older is linked with a higher incidence of illnesses such as heart and vascular disease, dementia, diabetes mellitus, orthopaedic problems and neurological disease.2 These illnesses may lead to care dependency. With high care dependency, problems in maintaining good oral health can also be expected.3 Knowledge on oral health care to vulnerable older people is mainly based on research conducted amongst residents of nursing homes. However, not much is known about oral health care provided to older people who live at home independently. In general, it can be said that older people make less use of professional oral health care than younger people do. Factors such as illness, being less mobile and accordingly being housebound, insufﬁcient ﬁnancial means, fear and a low self-estimated demand for care might be related.4–7 The aim of this study was to record how dentists in the Netherlands and Flanders assess their knowledge on oral health care to vulnerable older people, what their attitude is in this respect and what barriers they experience in rendering oral health care to these older people. 1BENECOMO, Flemish-Netherlands Geriatric Oral Research Group, Belgium, The Netherlands; 2Department of Social Dentistry and Behavioral Sciences, Academic Centre for Dentistry Amsterdam (ACTA), University of Amsterdam and Vrije Universiteit, Amsterdam, The Netherlands; 3Department of Research, Royal Dutch Dental Association (KNMT), Utrecht, The Netherlands; 4Department of Oral Biochemistry, Academic Centre for Dentistry Amsterdam (ACTA), University of Amsterdam and Vrije Universiteit, Amsterdam, The Netherlands; 5University Ghent, Faculty of Medicine and Health Sciences, Department of Dentistry, Community Dentistry and Oral Public Health Ghent, Belgium, The Netherlands and 6Caphri/Department of Family Medicine and Department Health Service Research, Maastricht University, Maastricht, The Netherlands. Correspondence: PC Bots-VantSpijker (pcbots-vantspijker@acta.nl) Received 20 June 2017; revised 31 August 2017; accepted 5 September 2017 www.nature.com/bdjopen ARTICLE OPEN Dentists’ opinions on knowledge, attitudes and barriers in providing oral health care to older people living independently in the Netherlands and Flanders (Belgium) PC Bots-VantSpijker1,2, JJM Bruers2,3, CP Bots4, LMJ De Visschere1,5 and JMGA Schols1,6 PC Bots-VantSpijker1,2, JJM Bruers2,3, CP Bots4, LMJ De Visschere1,5 and JMGA Schols1,6 OBJECTIVES: The aim of this study was to investigate how dentists in the Netherlands and Flanders assessed their knowledge on oral health care to older people, what their attitude was and what barriers they experienced in rendering care to older people. METHODS: The survey data was collected from a random sample of Dutch and Flemish dentists. Five hundred ninety-ﬁve dentists (37%) of the Dutch sample and 494 dentists of the Flemish sample (41%) completed the online questionnaire. Dentists were asked to respond to 15 Likert type items, representing opinions on provision of oral health care to older people and to give information about the number of older patients treated and about some profession-speciﬁc and personal characteristics. OBJECTIVES: The aim of this study was to investigate how dentists in the Netherlands and Flanders assessed their knowledge on oral health care to older people, what their attitude was and what barriers they experienced in rendering care to older people. METHODS: The survey data was collected from a random sample of Dutch and Flemish dentists. Five hundred ninety-ﬁve dentists (37%) of the Dutch sample and 494 dentists of the Flemish sample (41%) completed the online questionnaire. Dentists were asked to respond to 15 Likert type items, representing opinions on provision of oral health care to older people and to give information about the number of older patients treated and about some profession-speciﬁc and personal characteristics. RESULTS: The average number of patients treated per week was nearly twice as high in the Netherlands as in Flanders. Nevertheless, differences of opinions between dentists in the Netherlands and Flanders were relatively limited. CONCLUSIONS: This survey shows that in particular the actual number of older patients treated appears to be related with differences of opinions between Dutch and Flemish dentists about oral health care provided to (vulnerable) older people who live at home. CONCLUSIONS: This survey shows that in particular the actual number of older patients treated appears to be related with differences of opinions between Dutch and Flemish dentists about oral health care provided to (vulnerable) older people who live t h BDJOpen (2017) 3, 17020; doi:10.1038/bdjopen.2017.20; published online 22 December 2017 Construction delta scores The data representing the dentists’ opinions on the provision of oral health care to older people who still live at home showed no correlation from a statistical point of view. Therefore, the construction of one or more sum scores, based on the reactions of the dentists questioned on the 15 Likert type items, was not possible. Given the vast number of results accumulated by presenting these items separately, the data was subsequently processed in order to realise some data reduction. To this purpose, the Likert type items were ﬁrst re-coded from ﬁve into three possible answers; (very much) agree received score 1, neutral score 0 and (very much) disagree score -1. This procedure resulted in the loss of some information, but it did make it possible to express the opinion of the dentists in one ﬁgure. The mean value per item of this so-called delta score (Δ) was between −100 and +100 and represented the difference between the proportion of dentists who ‘agree’ minus the proportion of dentists who ‘disagree’. Here it holds that the closer the value is to +1,00 the more positive the dentists were inclined towards the item concerned. But the various scores could only be interpreted relatively. A mean delta (Δ) of 0.80 is more positive than a mean delta of 0.55 but this difference is not to be interpreted as an absolute difference of 0.25 because of the speciﬁc processing. Therefore, the delta score has an ordinal character, not assuming the magnitude of difference between each category, but representing the tendency to agree or disagree upon an opinion. In this way delta scores were determined for all the Likert type items. The questionnaire, tested in a pilot by a dozen dentists, consisted of three parts. In the ﬁrst part of the questionnaire some profession-speciﬁc and personal characteristics were recorded. In the second part, dentists were asked to give some information about the number of older patients they treated who still live at home. An inventory was made of how many patients per week they treat in their practice, how many of these are 75 years old or over and what percentage of these older patients they consider to be vulnerable. Vulnerable people were deﬁned as people who experienced problems in physical, psychological and social areas (includ- ing ﬁnancial problems). Situation per country After the analysis of some general and profession-speciﬁc characteristics of the dentists questioned, the division between the numbers of vulnerable older people that dentists said themselves they were treating was looked into. In doing so, the estimated percentage of vulnerable persons among the older patients who were treated was also determined and based on the distribution of these estimated percentages divided into quartiles. Subsequently, it was studied what the opinions of dentists in both countries are with regard to oral health care for these speciﬁc groups of patients and whether or not dentists in the Netherlands and Flanders differ from each other in this respect. Finally, it was assessed whether dentists who treat comparatively many (vulnerable) older people who live at home hold different views from dentists who proportionally treat fewer of these patients. In doing so, not only the differences within the Netherlands and Flanders were studied, but also the differences between the two countries. In these comparisons, the extremities were chosen as the starting points, i.e. only the dentists in the ﬁrst and the fourth quartiles, respectively, were compared to each other. In this way, it was possible to show more clearly any correlation between treating relatively more or relatively less vulnerable older people and the forming of opinions. For these ‘ﬁrst and fourth quartile’ dentists also the percentage of female dentists, the average age of dentists, the mean number of patients treated Furthermore, Table 2 shows that dentists in Flanders had in their practices a signiﬁcantly greater proportion of vulnerable older people who live at home independently (22.2% vs 18.8%). From the delta scores (Δ) presented in Table 3 it becomes clear that dentists in the Netherlands and Flanders only slightly differ in the way they look upon oral health care for vulnerable older people who live at home. In general, it can be concluded that reaction from Flemish dentists were slightly more positive with regard to treating vulnerable older people than those from their Dutch counterparts. In the Netherlands, dentists are more conﬁdent that they have sufﬁcient knowledge of the (adverse) effects of medication used by older people (0.42 vs 0.30). Flemish dentists more often believe that dental schools should pay more attention to provide students with adequate knowledge and skills with respect to the provision of oral health care to vulnerable older people (0.68 vs 0.60). Statistical analysis Since there were no differences between both study groups as regards to the questions asked and the data collection, the data from both countries was combined in a database and used for further statistical analysis. General characteristics In the end, 595 dentists (37%) of the Dutch sample completed the questionnaire. In Flanders 494 dentists (41%) did the same. As for gender, age, geographical location, graduation year and the university of qualiﬁcation the total group of respondents proved to be fairly representative for the population of dentists in the Netherlands and Flanders.8 In Flanders, relatively more participating dentists were women, whereas the average number of patients treated per week was nearly twice as high in the Netherlands as in Flanders (Table 1). Construction delta scores The age limit of 75 years was chosen, because it appeared that vulnerability increasingly occurs from that age.14 In part three of the questionnaire, dentists were asked to respond to 15 Likert type items representing opinions on the provision of oral health care to vulnerable older people who still live at home independently. The web based questionnaire drawn up for this study was randomly sent by e-mail to the dentists in the samples. By using a personalized login and password, which was communicated in the e-mail, dentists obtained access to the online questionnaire via a secure website. In this way, conﬁdential data collection was guaranteed. Dentists consented to participate in this study by sending their answers to the questions. All the dentists, in the Netherlands as well as in Flanders, who had not responded within two weeks and within ﬁve weeks consecutively, were sent a reminder by e-mail. The online survey period terminated six weeks after the start of the survey. Both the distribution of the questionnaires and the data management of the returned questionnaires were done by a research institute, independently from the authors, to ensure conﬁdentiality. Data collection The data were collected from a random sample of dentists in the Netherlands and Flanders by means of a web survey. In the Netherlands and Belgium, for survey research not involving patients, no ethical committee approval is required. In December 2010, a sample of 1592 of the ~ 8,000 dentists in the Netherlands, aged 64 years or younger, were invited to participate in this study. The data collection was described earlier in the article of Bots- VantSpijker et al.8 In September 2011, a sample of 1,200 dentists In Flanders (26% of the total population of Flemish dentists) were invited to take part in the survey. These dentists were randomly selected out of the 3,500 members of the Verbond van Vlaamse Tandartsen (VVT), a dental association representing 76% of all practising 4,615 Flemish dentists. Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al 2 INTRODUCTION The second aim was to consider whether there differences exist in these aspects between dentists in the Netherlands and in Flanders. For readability, Flanders will be mentioned as a country although it is in fact the Dutch speaking part of Belgium. Apart from the fact that older people may experience barriers in going to the dentist, dentists can also experience barriers to render oral health care to the elderly. Among other things this also becomes apparent from a survey by Bots-VantSpijker et al.8, conducted among dentists in the Netherlands. In line with this, it is interesting to investigate what the situation is in Flanders (Belgium). As for language, the level of prosperity and living conditions the region of Flanders in the neighbouring country of Belgium can be compared with the Netherlands. But distinctive MATERIAL AND METHODS Data collection per week, and the mean number of patients treated per week aged 75 and over were also determined. By means of ANOVA and the Mann Whitney U test for ordinal data it was assessed whether the outcome of the ‘ﬁrst and fourth quartile’ dentists showed any signiﬁcant differences. The analyses were done using SPSS for Windows (Version 15, IBM SPSS, Armonk, NY, USA). Situation per country They expressed more willingness to visit frail housebound older people for a regular dental check-up than dentists in the Netherlands (0.08 vs −0.08). Dutch dentists deny more often that treating vulnerable older people is not very challenging (−0.36 vs −0.27). BDJOpen (2017) 17020 BDJOpen (2017) 17020 Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al 3 Table 1. Personal and professional characteristics of dentists, per (part of the) country The Netherlands Flanders Total Proportion female (in %)a 24 48 34b Age in January 2011 (in %)c * 39 or younger 20 19 20 40–54 44 52 47 55–65 36 29 33 Mean age (s.d.) 49.4 (9.9) 48.4 (10.2) 49 (10) Year of dental graduation (in %) 1979 or earlier 25 31 28 1980–1989 43 39 41 1990–1999 15 18 16 2000 or later 17 12 15 Mean year (s.d.) 1986.8 (10) 1985.7 (10) 1986.3 (10) Proportion practice owners (in %)a 84 90 87b Mean (sd) number of patients treatedd 105.3 (58) 58.2 (24.8) 84.7 (52)b Mean (sd) number of patients treated, aged 75 years and overd 8.0 (8.7) 6.3 (5.3) 7.2 (7.5)b N 589–595 422–458 1.017–1.055 aDummy variable (1/0). bF-test: Po0.05. cChi Square test: Po0.05 / Cramer’s Vo0.15. dIn the previous full working week in the practice by the dentist himself. Table 1. Personal and professional characteristics of dentists, per (part of the) country Table 1. Personal and professional characteristics of dentists, per (part of the) country Flanders, group IV showed more often than group II willingness to visit housebound frail older people for a regular dental check-up and stated more frequently that their practice is easily accessible to vulnerable older people and less frequently that providing oral health care to vulnerable older people is difﬁcult due to its complexity and practical barriers. Flanders, group IV showed more often than group II willingness to visit housebound frail older people for a regular dental check-up and stated more frequently that their practice is easily accessible to vulnerable older people and less frequently that providing oral health care to vulnerable older people is difﬁcult due to its complexity and practical barriers. Table 2. Situation per country Percentage of ‘vulnerable’ patients within the number of patients, aged 75 years and over, treated by the dentist in the previous full working week, per (part of the) country The Netherlands Flanders Total 0% of the patients (1st quartile) 24% 26% 25% 1–9% of the patients (2nd quartile) 27% 18% 23% 10–29% of the patients (3rd quartile) 24% 24% 24% 30–100% of the patients (4th quartile) 25% 32% 28% Mean 18.8 22.2 20.3a Median 8 10 10 Modus 0 0 0 s.d. 24.8 26.2 25.5 Maximum 100 100 100 Minimum 0 0 0 N 590 450 140 aF-test: Po05. Table 2. Percentage of ‘vulnerable’ patients within the number of patients, aged 75 years and over, treated by the dentist in the previous full working week, per (part of the) country In the comparison between Dutch and Flemish dentists who treat relatively few vulnerable older people (groups I and II), it is striking that they only differ signiﬁcantly in one speciﬁc statement: The dentists in Flanders more often than the dentists in the Netherlands agree with the statement that daily attention for oral hygiene is a prerequisite for preventing oral health problems in dentate vulnerable older people. In the comparison between dentists in the Netherlands and Flanders who treat relatively many vulnerable older people (groups III and IV), it is striking that Flemish dentists are more often prepared to visit housebound frail older people than Dutch dentists. In addition, they see fewer technical barriers. The dentists in the Netherlands who treat relatively many vulnerable older people indicate that to them, reimbursement forms a barrier less frequently than it does to dentists from Flanders. They also more often than their Flemish counterparts, endorse the statement that they have sufﬁcient knowledge of the (adverse) effects of medication used by older people. But the dentist in Flanders sees fewer technical barriers compared to the dentist in the Netherlands. (0.52 vs 0.68). Dutch dentists deny more often that poor reimbursement for oral health care provisions to vulnerable older people forms a barrier to professional dedication towards this special patient group (−0.41 vs −0.27). But the dentist in Flanders sees fewer technical barriers compared to the dentist in the Netherlands. (0.52 vs 0.68). Dutch dentists deny more often that poor reimbursement for oral health care provisions to vulnerable older people forms a barrier to professional dedication towards this special patient group (−0.41 vs −0.27). DISCUSSION AND CONCLUSIONS From this study, it appears that there exist some differences of opinions between Dutch and Flemish dentists about oral health care provided to (vulnerable) older people who live at home on the subjects of knowledge, attitude and barriers. Table 5 offers a summary of the results. K1–K5—opinions on knowledge. A1–A4—opinions on attitude. B1–B6—opinions on barriers Agree—proportion dentists who agree with an opinion. Disagree—proportion dentists who disagree with an opinion. Δ—mean difference score (Likert type items were recoded from ﬁve into three possible answers: (very much) agree (score 1) neutral (score 0) and (very much) disagree (score −1). Following this, a delta score (Δ) was calculated by taking the mean of the scores per statement, which represents the difference between proportion dentists who ‘agree’ minus proportion dentists who ‘disagree’). aMann-Whitney U-test: Po05. The Netherlands versus Flanders 0.95 (0.96–0.01) 0.97 (0.97–0.01) 0.96 (0.97–0.01) 0.24 A1 Every dentist is responsible for providing proper oral health care to housebound frail older people who used to visit his clinic regularly. 0.27 (0.49–0.21) 0.24 (0.47–0.23) 0.26 (0.48–0.22) 0.48 A2 I am willing to visit housebound frail older people for a regular dental check-up. -0.08 (0.38–0.46) 0.08 (0.44–0.37) -0.01 (0.40–0.42) 0.01a A3 I have experienced several times over that, at a certain moment, (frail) older people stopped coming to the practice regularly. 0.62 (0.73–0.11) 0.70 (0.77–0.08) 0.65 (0.75–0.10) 0.08 A4 From a dentist’s point of view, treating vulnerable older people is not very challenging. -0.36 (0.19–0.55) -0.27 (0.21–0.48) -0.32 (0.20–0.52) 0.04a B1 Opportunities to refer vulnerable older people with complex oral health problems to a colleague with speciﬁc knowledge and skills are limited. 0.67 (0.74–0.07) 0.67 (0.75–0.08) 0.67 (0.74–0.08) 0.98 B2 Providing oral health care to vulnerable older people is difﬁcult due to its complexity and practical barriers. 15.7 (45.5–29.8) 22.0 (59.9–27.9) 18.5 (47.4–29.0) 0.24 B3 The reimbursement of oral health care provision to vulnerable older people is poor. 0.16 (0.46–0.30) 0.22 (0.60–0.28) 0.19 (0.47–0.29) 0.84 B4 My practice is easily accessible for vulnerable older people, without major obstacles. 0.72 (0.81–0.10) 0.71 (0.79–0.08) 0.71 (0.80–0.09) 0.48 B5 Usually, oral health care for vulnerable older people implies restraints with regard to technical facilities. 0.68 (0.76–0.08) 0.52 (0.63–0.11) 0.61 (0.70–0.09) 0.00a B6 I regard the poor reimbursement of oral health care provision to vulnerable older people as a barrier to professional dedication to this special patient group. −0.41 (0.15–0.56) −0.27 (0.19–0.46) −0.35 (0.17–0.52) 0.00a n = 517–553 n = 403–425 n = 920–978 K1 K5 opinions on knowledge pants’ opinions on oral health care provision to vulnerable older people who live at home, per (part of the) country The Netherlands (NL) Flanders (FL) Total P-value Δ (agree–disagree) Δ (agree–disagree) Δ (agree–disagree) NL versus FL Physical, psychological, and social aspects have an impact on oral health care decision making 0.92 (0.93–0.01) 0.90 (0.92–0.02) 0.91 (0.93–0.02) 0.55 I have sufﬁcient knowledge of the (adverse) effects of medication used by older people. 0.42 (0.56–0.15) 0.30 (0.52–0.21) 0.37 (0.54–0.18) 0.04a I am capable of providing oral health care to cognitively impaired vulnerable older people. 0.39(0.53–0.14) 0.46 (0.58–0.12) 0.42 (0.55–0.13) 0.14 Dental schools should pay more attention to providing students with adequate knowledge and skills with respect to oral health care provision to vulnerable older people. The Netherlands versus Flanders The Netherlands versus Flanders Table 4 shows the opinions on oral health care to vulnerable older people, distinguished into dentists in the Netherlands (group I) and Flanders (group II) who treat relatively few vulnerable older people and dentists who treat relatively many vulnerable older people (group III the Netherlands, group IV Flanders). It appears that in Flanders, the differences between these groups of dentists cover a somewhat wider range of opinions than in the Nether- lands. In Flanders, as in the Netherlands, groups III and IV stated more often than groups I and II that they are capable of providing oral health care to cognitively impaired vulnerable older people and that they have experienced several times over that frail older people stopped coming to their clinics regularly. In addition, in Interpretation of differences per country It is remarkable that the Dutch dentists state more frequently that they have sufﬁcient knowledge about the (adverse) effects of medication used by older people. Accordingly, the Flemish dentists stated more often that dental schools should pay more attention to furnishing students with adequate knowledge and skills with respect to providing oral health care to vulnerable older people. These differences may be explained by a diverse interpretation of the competences in geriatric dentistry in the Netherlands and Flanders. BDJOpen (2017) 17020 Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al Table 3. Participants’ opinions on oral health care provision to vulnerable older people who live at home, per (part of the) country The Netherlands (NL) Flanders (FL) Total P-value Δ (agree–disagree) Δ (agree–disagree) Δ (agree–disagree) NL versus FL K1 Physical, psychological, and social aspects have an impact on oral health care decision making 0.92 (0.93–0.01) 0.90 (0.92–0.02) 0.91 (0.93–0.02) 0.55 K2 I have sufﬁcient knowledge of the (adverse) effects of medication used by older people. 0.42 (0.56–0.15) 0.30 (0.52–0.21) 0.37 (0.54–0.18) 0.04a K3 I am capable of providing oral health care to cognitively impaired vulnerable older people. 0.39(0.53–0.14) 0.46 (0.58–0.12) 0.42 (0.55–0.13) 0.14 K4 Dental schools should pay more attention to providing students with adequate knowledge and skills with respect to oral health care provision to vulnerable older people. 0.60 (0.67–0.07) 0.68 (0.73–0.06) 0.64 (0.70–0.06) 0.04a K5 Daily attention for oral hygiene care is a prerequisite for preventing oral health problems in dentate vulnerable older people. The Netherlands versus Flanders 0.60 (0.67–0.07) 0.68 (0.73–0.06) 0.64 (0.70–0.06) 0.04a Daily attention for oral hygiene care is a prerequisite for preventing oral health problems in dentate vulnerable older people. 0.95 (0.96–0.01) 0.97 (0.97–0.01) 0.96 (0.97–0.01) 0.24 Every dentist is responsible for providing proper oral health care to housebound frail older people who used to visit his clinic regularly. 0.27 (0.49–0.21) 0.24 (0.47–0.23) 0.26 (0.48–0.22) 0.48 I am willing to visit housebound frail older people for a regular dental check-up. -0.08 (0.38–0.46) 0.08 (0.44–0.37) -0.01 (0.40–0.42) 0.01a I have experienced several times over that, at a certain moment, (frail) older people stopped coming to the practice regularly. 0.62 (0.73–0.11) 0.70 (0.77–0.08) 0.65 (0.75–0.10) 0.08 From a dentist’s point of view, treating vulnerable older people is not very challenging. -0.36 (0.19–0.55) -0.27 (0.21–0.48) -0.32 (0.20–0.52) 0.04a Opportunities to refer vulnerable older people with complex oral health problems to a colleague with speciﬁc knowledge and skills are limited. 0.67 (0.74–0.07) 0.67 (0.75–0.08) 0.67 (0.74–0.08) 0.98 Providing oral health care to vulnerable older people is difﬁcult due to its complexity and practical barriers. 15.7 (45.5–29.8) 22.0 (59.9–27.9) 18.5 (47.4–29.0) 0.24 The reimbursement of oral health care provision to vulnerable older people is poor. 0.16 (0.46–0.30) 0.22 (0.60–0.28) 0.19 (0.47–0.29) 0.84 My practice is easily accessible for vulnerable older people, without major obstacles. 0.72 (0.81–0.10) 0.71 (0.79–0.08) 0.71 (0.80–0.09) 0.48 Usually, oral health care for vulnerable older people implies restraints with regard to technical facilities. 0.68 (0.76–0.08) 0.52 (0.63–0.11) 0.61 (0.70–0.09) 0.00a I regard the poor reimbursement of oral health care provision to vulnerable older people as a barrier to professional dedication to this special patient group. −0.41 (0.15–0.56) −0.27 (0.19–0.46) −0.35 (0.17–0.52) 0.00a n = 517–553 n = 403–425 n = 920–978 k l d I have experienced several times over that, at a certain moment, (frail) older people stopped coming to the practice regularly. B6 BDJOpen (2017) 17020 Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al 5 Table 4. The Netherlands versus Flanders Dentist’ characteristics and professional attitudes on oral health care to vulnerable elderly, per (part of the) country The Netherlands (NL) Flanders (FL) Comparisons Less 1st quartile (groupI) High 4th quartile (group III) Less 1st quartile (group II) More 4th quartile (group IV) P-values NL-less versus NL-more (I–III) P-values FL-less versus FL-more (II–IV) P-values NL- less versus FL- less (I–II) P-values NL- more versus FL- more (III–IV) NL-less NL-more FL-less FL-more Female dentists (in %) 32.6 24.7 61.8 49.6 0.13 0.06 0.00a 0.00a Mean age of dentists 49.0 48.4 46.7 47.5 0.62 0.55 0.09 0.42 Mean number of patients treated per week 98.3 106.0 51.3 56.4 0.30 0.07 0.00a 0.00a Mean number of patients treated per week, aged 75 years and over 2.6 8.8 3.0 6.5 0.00a 0.00a 0.32 0.02a Δ K1 0.92 0.96 0.90 0.90 0.21 0.89 0.86 0.11 Δ K2 0.37 0.45 0.17 0.24 0.31 0.51 0.07 0.03b Δ K3 0.18 0.41 0.28 0.50 0.01b 0.02b 0.27 0.31 Δ K4 0.70 0.63 0.69 0.71 0.34 0.83 0.72 0.25 Δ K5 0.92 0.96 0.98 0.96 0.10 0.17 0.03b 0.46 Δ A1 0.28 0.32 0.23 0.24 0.66 0.99 0.70 0.37 Δ A2 -0.17 -0.19 -0.07 0.23 0.88 0.01b 0.37 0.00b Δ A3 0.51 0.71 0.57 0.73 0.05b 0.04b 0.93 0.89 Δ A4 -0.31 -0.41 -0.30 -0.28 0.37 0.92 0.64 0.11 Δ B1 0.68 0.75 0.69 0.70 0.32 0.95 0.73 0.57 Δ B2 0.26 0.17 0.35 0.13 0.38 0.03b 0.34 0.67 Δ B3 0.58 0.53 0.49 0.62 0.85 0.18 0.51 0.34 Δ B4 0.62 0.76 0.57 0.74 0.10 0.01b 0.32 0.85 Δ B5 0.64 0.71 0.54 0.60 0.45 0.79 0.18 0.04b Δ B6 −0.32 -0.46 -0.40 -0.27 0.12 0.17 0.47 0.02b Positive about guideline 0.79 0.77 0.77 0.85 0.67 0.11 0.66 0.11 N 114–141 136–150 102–116 127–144 Δ—mean difference score (Likert type items were recoded from ﬁve into three possible answers: (very much) agree (score 1) neutral (score 0) and (very much) disagree (score −1). Following this, a delta score (Δ) was calculated by taking the mean of the scores per statement, which represents the difference between proportion dentists who ‘agree’ minus proportion dentists who ‘disagree’.). K1–K5—opinions on knowledge. K1—Physical, psychological, and social aspects have an impact on oral health care decision-making. K2—I have sufﬁcient knowledge of the (adverse) effects of medication used by older people. K3—I am capable of providing oral health care to cognitively impaired frail older people. The Netherlands versus Flanders K4—Dental schools should pay more attention to teaching students adequate knowledge and skills with respect to oral health care provision to vulnerable older people. K5—Daily attention for oral hygiene is a prerequisite for preventing oral health problems in dentate vulnerable older people. A1–A4 opinions on attitudes. A1—Every dentist is responsible for providing proper oral health care to housebound frail older people who used to visit his clinic regularly. A2—I am willing to visit housebound frail older people for a regular dental check-up. A3—I have experienced several times over that, at a certain moment, (frail) older people stopped coming to the clinic regularly. A4—From a dentist’s point of view, treating vulnerable older people is not very challenging. B1–B6—opinions on barriers. B1—Opportunities to refer vulnerable older people with complex oral health care problems to a colleague with speciﬁc knowledge and skills are limited. B2—Providing oral health care to vulnerable older people is difﬁcult due to its complexity and practical barriers. B3—The reimbursement of oral health care provision to vulnerable older people is poor. B4—My practice is easily accessible for vulnerable older people, without major obstacles. Table 4. The Netherlands versus Flanders Dentist’ characteristics and professional attitudes on oral health care to vulnerable elderly, per (part of the) country The Netherlands (NL) Flanders (FL) Comparisons Less 1st quartile (groupI) High 4th quartile (group III) Less 1st quartile (group II) More 4th quartile (group IV) P-values NL-less versus NL-more (I–III) P-values FL-less versus FL-more (II–IV) P-values NL- less versus FL- less (I–II) P-values NL- more versus FL- more (III–IV) NL-less NL-more FL-less FL-more Female dentists (in %) 32.6 24.7 61.8 49.6 0.13 0.06 0.00a 0.00a Mean age of dentists 49.0 48.4 46.7 47.5 0.62 0.55 0.09 0.42 Mean number of patients treated per week 98.3 106.0 51.3 56.4 0.30 0.07 0.00a 0.00a Mean number of patients treated per week, aged 75 years and over 2.6 8.8 3.0 6.5 0.00a 0.00a 0.32 0.02a Δ K1 0.92 0.96 0.90 0.90 0.21 0.89 0.86 0.11 Δ K2 0.37 0.45 0.17 0.24 0.31 0.51 0.07 0.03b Δ K3 0.18 0.41 0.28 0.50 0.01b 0.02b 0.27 0.31 Δ K4 0.70 0.63 0.69 0.71 0.34 0.83 0.72 0.25 Δ K5 0.92 0.96 0.98 0.96 0.10 0.17 0.03b 0.46 Δ A1 0.28 0.32 0.23 0.24 0.66 0.99 0.70 0.37 Δ A2 -0.17 -0.19 -0.07 0.23 0.88 0.01b 0.37 0.00b Δ A3 0.51 0.71 0.57 0.73 0.05b 0.04b 0.93 0.89 Δ A4 -0.31 -0.41 -0.30 -0.28 0.37 0.92 0.64 0.11 Δ B1 0.68 0.75 0.69 0.70 0.32 0.95 0.73 0.57 Δ B2 0.26 0.17 0.35 0.13 0.38 0.03b 0.34 0.67 Δ B3 0.58 0.53 0.49 0.62 0.85 0.18 0.51 0.34 Δ B4 0.62 0.76 0.57 0.74 0.10 0.01b 0.32 0.85 Δ B5 0.64 0.71 0.54 0.60 0.45 0.79 0.18 0.04b Δ B6 −0.32 -0.46 -0.40 -0.27 0.12 0.17 0.47 0.02b Positive about guideline 0.79 0.77 0.77 0.85 0.67 0.11 0.66 0.11 N 114–141 136–150 102–116 127–144 Δ diff (Lik i d d f ﬁ i h ibl ( h) ( ) l ( ) d ( h) Table 4. Dentist’ characteristics and professional attitudes on oral health care to vulnerable elderly, per (part of the) country B3—The reimbursement of oral health care provision to vulnerable older people is poor. p p p p B4—My practice is easily accessible for vulnerable older people, without major obstacles. —Usually, oral health care to vulnerable older people implies restraints with regard to technical facilities. —Poor reimbursement of oral health care provision to vulnerable older people is a barrier for my professional dedication t st: Po0.05. Limitations of the study The response rate of the Dutch dentists (37%) and of the Flemish dentists (41%) is, compared to other web-based surveys, satisfactory.19,20 The number of dentists in this study is in accordance with 7.5% of the total number of dentists in the Netherlands and 10.7% of the total number of dentists in Flanders. In addition, it holds that the Dutch dentists involved in this study are representative of the population of dentists in that country, be it with a slight over representation of older dentists. A number of the Dutch dentists in this study belong to a group who regularly participate in surveys.21 This may have resulted in a positive bias, because these dentists are maybe motivated above average. The group of dentists from Flanders, all member of at least one organisation, has been approached for this study once only, resulting in a possibly lesser positive bias. In addition, it does not seem plausible that the timeline difference between both researches has inﬂuenced the outcome of the questionnaire. A1—Every dentist is responsible for providing proper oral health care to housebound frail older people who used to visit his clinic regularly. A2—I am willing to visit housebound frail older people for a regular dental check-up. A3—I have experienced several times over that, at a certain moment, (frail) older people stopped coming to the clinic regularly. A4—From a dentist’s point of view, treating vulnerable older people is not very challenging. B1–B6—opinions on barriers. B1—Opportunities to refer vulnerable older people with complex oral health care problems to a colleague with speciﬁc knowledge and skills are limited. B2—Providing oral health care to vulnerable older people is difﬁcult due to its complexity and practical barriers. B3–The reimbursement of oral health care provision to vulnerable older people is poor. Another limitation of this study is that the participants estimated the average numbers of patients treated per week aged 75 years or over, and those of vulnerable patients aged 75 years or over. It remains uncertain whether these estimations reﬂect the reality. To identify the exact average numbers of patients treated per week, dentists should have gathered information from their patient administrations. It seems unlikely that every dentist in this study did so. forms a barrier in providing oral health care to vulnerable older patients. This indicates that differences in the ﬁnancing of the dental services effect the providing of oral health care to frail older people. The Netherlands versus Flanders bMan–Whitney U-test: Po0.05. The fact that the Flemish dentists saw fewer technical barriers in the rendering of oral health care to the elderly than their Dutch colleagues did, is in accordance with the fact that they were more prepared to make house calls. Based on clinical experience, this could have something to do with the smaller sized practices in Flanders: the smaller the practice, the more time to call on patients at home. It is also conceivable that in group practices there is more opportunity to free up time for making home visits. It is not clear how this difference could be explained. From a survey among German GPs about their attitude towards house calls, the main objection against making house calls was the reimbursement it yielded. Perhaps this argument also plays a part in the difference between the two countries in this study.16 In the Netherlands, the cost of oral health care are in more cases entirely for the patient himself, although it is possible to take out additional insurance. In Flanders, older people will get a 75% reimbursement for oral health care. Nevertheless, the Flemish dentists were more often convinced that inadequate ﬁnancing BDJOpen (2017) 17020 Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al 6 Table 5. Summary of the results of Tables 3 and 4 Statement Table 3 Table 4 l–h Neth l–h Fl l-l Neth/Fl h-h Neth/Fl K1 0 0 0 0 0 K2 −(NL4FL) 0 0 0 −(NL4FL) K3 0 −(H4L) −(H4L) 0 0 K4 + (FL4NL) 0 0 0 0 K5 0 0 0 + (FL4NL) 0 A1 0 0 0 0 0 A2 + (FL4NL) 0 −(H4L) 0 + (FL4NL) A3 0 −(H4L) −(H4L) 0 0 A4 +(FL4NL) 0 0 0 0 B1 0 0 0 0 0 B2 0 0 + (L4H) 0 0 B3 0 0 0 0 0 B4 0 0 −(H4L) 0 0 B5 −(NL4FL) 0 0 0 −(NL4FL) B6 + (FL4NL) 0 0 0 + (FL4NL) et al., who described that the availability of dental services, the organization of the dental health care delivery system, and price subsidy for dental treatment are signiﬁcant mitigating factors that may inﬂuence the use of dental services among older people.17 Bots-VantSpijker et al.8 already demonstrated that the ﬁnancial reimbursement was one of the barriers, and in addition some factors related to attitude and knowledge. The Netherlands versus Flanders Besides, the funding of oral health care in the Netherlands and Flanders is quite complex and the dentists in this study were not asked about this topic. That dentists in Flanders more often than the dentists in the Netherlands agreed with the statement that daily attention for oral hygiene is a prerequisite for preventing oral health problems in dentate vulnerable older people, might be caused by the fact that the DMFT of people, aged 65–75 years old is lower in Flanders than in the Netherlands.11,12 Interpretation of differences regarding number vulnerable older patients treated Interpretation of differences regarding number vulnerable older patients treated In the comparison between dentists who treat relatively few and dentists who treat relatively many vulnerable older people, it comes to the latter felt more capable to do so. This can be explained and is in accordance with the ﬁndings of Ettinger,18 who described the importance of students spending sufﬁcient clinical time with frail and medically compromised patients. Having sufﬁcient clinical time, students may learn to develop rational treatment planning, and perform sufﬁcient treatment procedures in order to feel clinically conﬁdent and comfortable towards the treatment of vulnerable and medically compromised patients. Although this study was conducted among students it is to be expected that the more vulnerable older patients one treats, the more capable one feels and will therefore also notice more often that patients no longer come to the practice. It is also remarkable, for that matter, that in the comparison between dentists who treat relatively few and relatively many vulnerable older patients, reimbursement did not appear to play a signiﬁcant role, neither in the Netherlands, nor in Flanders. Score 0, no signiﬁcant difference between the Netherlands and Flanders or between low and high number of vulnerable elderly treated. Score −, the outcome measure of Flanders is smaller than that of the Netherlands or the outcome measure for a low number of patients treated is smaller than for a high number of patients treated. Score+, the outcome measure of Flanders is bigger than that of the Netherlands or the outcome measure for a low number of patients treated is bigger than for a high number of patients treated. K1—Physical, psychological, and social aspects have an impact on oral health care decision-making. K2—I have sufﬁcient knowledge of the (adverse) effects of medication used by older people. K3—I am capable of providing oral health care to cognitively impaired frail older people. K4—Dental schools should pay more attention to teaching students adequate knowledge and skills with respect to oral health care provision to vulnerable older people. p p K5—Daily attention for oral hygiene is a prerequisite for preventing oral health problems in dentate vulnerable older people. DECLARATIONS 5 Borreani E, Jones K, Scambler S, Gallagher JE. Informing the debate on oral health care for older people: a qualitative study of older people's views on oral health and oral health care. Gerodontology 2010; 27: 11–18. Ethics: In the Netherlands and Belgium, for survey research not involving patients, no ethical committee approval is required. They survey includedquestions to dentists about characteristics of the practice they work in, as well as about their professional opinion and behaviour on non-sensitive matters. On avolunta- rily basis they could decide to respond or not to respond to the request to participate in the survey. Dentists consented to participate in this study by sending in their answers to the questions. The survey was carried out by an independent third party research institute (theInstitute for Applied Social Sciences (ITS), which was linked to the Radboud University in Nijmegen, the Netherlands (www.ru.nl/its/) until May 2016) commissioned by the Royal Dutch Dental Association (Koninklijke Nederlandse Maatschappijtot bevordering der Tandheelkunde, KNMT). The independent third party research institute conﬁdentially col- lects and manages all research data in such a way that researchers cannot trace any of the data to anindividual dentist or practice. The institute carried out the distribution of the questionnaires, the data management of the returned ques- tionnaires, the interlinking of the research data with basic information of dentists and the anonymization of the data (necessary to make the research database available for the actual researchers). 6 Ornstein KA, DeCherrie L, Gluzman R, Scott ES, Kansal J, Shah T, Katz R, Soriano TA. Signiﬁcant Unmet Oral Health Needs of Homebound Elderly Adults. J Am Geriatr Soc. 2015; 63: 151–157. 7 De Visschere L. The approaches of oral health problems of institutionalized older people in an European context. In: De Visschere L, The development and appli- cation of an oral health care model for institutionalized older people. Gent: Academia Press, Scientiﬁc Publishers, 2010; 27. Online information available at http://hdl.handle.net/1854/LU-1085832. (Accessed November 2017). 8 Bots-VantSpijker PC, Bruers JJ, Bots CP, Vanobbergen JN, De Visschere LM, de Baat C, Schols JM. Opinions of dentists on the barriers in providing oral health care to community-dwelling frail older people: a questionnaire survey. Gerodontology 2016; 33: 268–274. 9 EU Manual of dental practice 2014, edition 5, data registered 2007-2014. Online information available at http://www.eudental.eu/library/eu-manual.html. (Acces- sed November 2017). 10 Online information available at http://statbel.fgov.be/nl/statistieken/cijfers/bev olking/structuur/woonplaats/, 2014 (Accessed November 2017). 11 TNO, 2013. CONTRIBUTIONS PC: main author, responsible for construction of the questionnaire, design, statistical analysis and intellectual set-up of the article. JB: participated in the design of the study and performed the statistical analysis and contributed to the draft of the manuscript. CB: conceived of the study, and participated in its design and contributed to draft the discussion of the manuscript. LV: main responsible person for conducting the part of this study which involved Belgium and contributed to the manuscript with respect to the Belgium situation. JS: contributed to the introduction and design and provided substantial comment to the main manuscript. All authors read and approved the ﬁnal manuscript. ACKNOWLEDGEMENTS We thank T. Adamson of the Royal Dutch Dental Association as a language editor, for her translation of the manuscript. Although both in the Netherlands and Flanders Dutch language is spoken, it is conceivable that in the two countries dentists have differently interpreted some statements. For example, in the Netherlands the dental consultation (with or without treatment) is seen as a procedure, resulting from a charging system, which is based on procedures. The Flemish dentist has possibly regarded the description of treatment more as an actual preventive or curative procedure. Conclusion The differences in opinions between dentists in the Netherlands and Flanders, also taking into account the number of vulnerable older people they themselves are treating in their practice, appear to be relatively limited. More additional qualitative research in the dental practice on mainly attitude and barriers experienced is desirable.22 This will allow for the differences ascertained in this study to be better circumscribed. In addition, more insight is needed in the actual demand for oral health care to vulnerable older people who still live at home independently. What are the oral health problems with which these older patients confront dentists in the general practice and what care do dentists offer to this speciﬁc group of patients? Oral health care to vulnerable older people in the dental practice can only be further improved if more scientiﬁc and professional knowl- edge comes available. The authors declare no conﬂict of interest. The authors declare no conﬂict of interest. Limitations of the study This is in accordance with the study of Holm-Pedersen In addition, the vulnerability of the group of older patients had to be estimated by dentists. Given the nature of this survey study it was not possible to apply a known vulnerability index, such as from Tilburg Frailty Indicator.15 BDJOpen (2017) 17020 Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al 7 of the third party are: KNMT: j.bruers@knmt.nl (research coordinator KNMT). of the third party are: KNMT: j.bruers@knmt.nl (research coordinator KNMT). The hypothesis that the season, in which the questionnaire was administered potentially, might have inﬂuenced the estimated numbers of patients treated per week aged 75 or over could not be excluded. Furthermore, the terms ‘complexity’ and ‘practical barriers’ were open to free interpretation and not deﬁned in detail. Therefore, the lack of a detailed explanation may also have inﬂuenced the results. REFERENCES 1 European Commision. “Confronting demographic change: a new solidarity between the generations”. Green, Paper, 2005. Online information available at http://eur-lex.europa.eu/legal-content/EN/TXT/?uri = LEGISSUM:c10128. (Accessed November 2017). 2 Branca S, Bennati E, Ferlito L, Spallina G, Cardillo E, Malaguarnera M, Motta M. IMUSCE. The health-care in the extreme longevity. Arch Gerontol Geriatr. 2009; 49: 32–34. 3 Van der Putten GJ. de Baat C, De Visschere L, Schols J. Poor oral health, a potential new geriatric syndrome. Gerodontology 2014; 3: 17–24. 4 Fiske J, Gelbier S, Watson RM. Barriers to dental care in an elderly population resident in an inner city area. J Dent. 1990; 18: 236–242. DECLARATIONS Online information available at http://repository.tudelft.nl/view/tno/ uuid%3Ac776997e-93f8-4fd8-a317-15fade31925b/ (Accessed November 2017). 12 Dataregistratie en evaluatie van de mondgezondheid van de Belgische bevolking 2008-2010. Interuniversitaire Cel Epidemiologie, 2011. Online information avail- able at www.ice.ugent.be/ﬁles/rapportNL.pdf. (Accessed November 2017). 13 Koninklijke Nederlandse Maatschappij tot bevordering der Tandheelkunde (KNMT). Online information available at http://www.staatvandemondzorg.nl Nieuwegein: KNMT, 2015. (Accessed November 2017). 14 Kalsbeek H, Truin GJ. Epidemiologie. In Loveren C Van, Weijden GA van der, (ed) Preventieve tandheelkunde. pp 10-16. Houten/Diegem: Bohn Staﬂeu van Loghum, 2000. 17 Holm-Pedersen P, Vigild M, Nitschke I, Berkey DB. Dental care for aging popula- tions in Denmark, Sweden, Norway, United Kingdom, and Germany. J Dent Educ 2005; 69: 987–997. 18 Ettinger RL. A 30-year review of a geriatric dentistry teaching programme. Ger- odontology 2012; 29: e1252–1260. Data availability Commissioned by the Royal Dutch Dental Association (KNMT), the data are stored and managed by an independent research institute, which operates as a third party and is afﬁliated with the Radboud University Nijmegen. If others would like to use the data, a request must be made directly to KNMT to provide an anonymous version of the research database. The contact details 15 Gobbens RJ, van Assen MA, Luijkx KG, Schols JM. The predictive validity of the Tilburg Frailty Indicator: disability, health care utilization, and quality of life in a population at risk. Gerontologist. 2012; 52: 619–631. 16 Theile G, Kruschinski C, Buck M, Müller CA, Hummers-Pradier E. Home visits - central to primary care, tradition or an obligation? A qualitative study. BMC. Fam Pract. 2011; 12: 24. BDJOpen (2017) 17020 Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al 22 Dawes M, Summerskill W, Glasziou P, Cartabellotta A, Martin J, Hopayian K, Porzsolt F, Burls A, Osborne J. Second International Conference of Evidence-Based 19 Cook C, Heath F, Thompson R. A meta-analysis of response rates in web- or internet-based surveys. Educ Psychol Measur 2000; 60: 821–836. Dentists’ opinions on providing oral health care to elderly PC Bots-VantSpijker et al 8 Health Care Teachers and Developers. Sicily statement on evidence-based practice. BMC Med Educ. 2005; 5: 1. 17 Holm-Pedersen P, Vigild M, Nitschke I, Berkey DB. Dental care for aging popula- tions in Denmark, Sweden, Norway, United Kingdom, and Germany. J Dent Educ 2005; 69: 987–997. 18 Ettinger RL. A 30-year review of a geriatric dentistry teaching programme. Ger- odontology 2012; 29: e1252–1260. This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/ by/4.0/ 19 Cook C, Heath F, Thompson R. A meta-analysis of response rates in web- or internet-based surveys. Educ Psychol Measur 2000; 60: 821–836. 20 Nulty DD. The adequacy of response rates to online and paper surveys: what can be done? Assess Eval Higher Educ 2008; 33: 301–314. 21 Bruers JJ, Zeegers GL. The Data Stations Project of the Dutch Dental Association. Adv Dent Res 2005; 18: 50–52. 22 Dawes M, Summerskill W, Glasziou P, Cartabellotta A, Martin J, Hopayian K, Porzsolt F, Burls A, Osborne J. Second International Conference of Evidence-Based © The Author(s) 2017 BDJOpen (2017) 17020
https://openalex.org/W2317628667	https://journals.iucr.org/e/issues/2010/12/00/hg2753/hg2753.pdf	English	null	2-[(5,7-Dibromoquinolin-8-yl)oxy]-<i>N</i>-(2-methoxyphenyl)acetamide	Acta crystallographica. Section E	2,010	cc-by	4,056	Table 1 Hydrogen-bond geometry (A˚ , ). In the title compound, C18H14Br2N2O3, an intramolecular N— H N hydrogen bond forms an eight-membered ring. The dihedral angle between the planes of the quinoline system and the benzene ring is 41.69 (1). The crystal packing is stabilized by intermolecular C—H O hydrogen bonds and short Br O interactions [3.0079 (19) A˚ ]. D—H A D—H H A D A D—H A N2—H2A N1 0.90 (1) 2.24 (1) 3.065 (3) 153 (1) C18—H18C O2i 0.96 2.53 3.342 (3) 142 Symmetry code: (i) x; y 1; z. Data collection: SMART (Bruker, 2001); cell reﬁnement: SAINT (Bruker, 2001); data reduction: SAINT; program(s) used to solve structure: SHELXTL (Sheldrick, 2008); program(s) used to reﬁne structure: SHELXTL; molecular graphics: SHELXTL; software used to prepare material for publication: SHELXTL. Related literature The structure of N,N-dicyclohexyl-2-(5,7-dibromoquinolin-8- yloxy)acetamide has been reported by Liu et al. (2007). For bond-length data, see: Allen et al. (1987). For applications of 8-hydroxyquinoline and its derivatives, see: Bratzel et al. (1972). Some 8-hydroxyquinoline derivatives and their trans- ition metal complexes exhibit antibacterial activity, see: Patel & Patel (1999). This work was supported by the National Natural Science Foundation of China (No. 20971076), the Outstanding Adult– young Scientiﬁc Research Encouraging Foundation of Shan- dong Province, China (No. 2008BS0901). The authors acknowledge X.-F. Tang’s help with this paper. Experimental Crystal data C18H14Br2N2O3 Mr = 466.13 Monoclinic, P21=n a = 8.7570 (18) A˚ Supplementary data and ﬁgures for this paper are available from the IUCr electronic archives (Reference: HG2753). Yong-Hong Wen,* Hong-Qing Qin and Hui-Ling Wen College of Chemistry and Molecular Engineering, Qingdao University of Science and Technology, Qingdao 266042, People’s Republic of China Correspondence e-mail: wenyyhh@126.com H atoms treated by a mixture of independent and constrained reﬁnement max = 0.83 e A˚ 3 min = 0.71 e A˚ 3 Received 15 November 2010; accepted 19 November 2010 Key indicators: single-crystal X-ray study; T = 293 K; mean (C–C) = 0.004 A˚; R factor = 0.044; wR factor = 0.093; data-to-parameter ratio = 17.4. Table 1 Hydrogen-bond geometry (A˚ , ). D—H A D—H H A D A D—H A N2—H2A N1 0.90 (1) 2.24 (1) 3.065 (3) 153 (1) C18—H18C O2i 0.96 2.53 3.342 (3) 142 Symmetry code: (i) x; y 1; z. organic compounds b = 8.7279 (17) A˚ c = 22.372 (5) A˚ = 98.04 (3) V = 1693.1 (6) A˚ 3 Z = 4 Mo K radiation = 4.81 mm1 T = 293 K 0.06 0.02 0.02 mm Data collection Bruker SMART CCD area-detector diffractometer Absorption correction: multi-scan (SADABS; Sheldrick, 1996) Tmin = 0.761, Tmax = 0.910 12864 measured reﬂections 4027 independent reﬂections 3316 reﬂections with I > 2(I) Rint = 0.057 Reﬁnement R[F 2 > 2(F 2)] = 0.044 wR(F 2) = 0.093 S = 1.06 4027 reﬂections 231 parameters 1 restraint H atoms treated by a mixture of independent and constrained reﬁnement max = 0.83 e A˚ 3 min = 0.71 e A˚ 3 Acta Crystallographica Section E Structure Reports Online ISSN 1600-5368 Acta Crystallographica Section E Structure Reports Online ISSN 1600-5368 Mo K radiation = 4.81 mm1 T = 293 K 0.06 0.02 0.02 mm V = 1693.1 (6) A Z = 4 Data collection Bruker SMART CCD area-detector diffractometer Absorption correction: multi-scan (SADABS; Sheldrick, 1996) Tmin = 0.761, Tmax = 0.910 Reﬁnement R[F 2 > 2(F 2)] = 0.044 wR(F 2) = 0 093 Sheldrick, G. M. (1996). SADABS. University of Go¨ttingen, Germany. Sheldrick, G. M. (2008). Acta Cryst. A64, 112–122. Patel, A. K. & Patel, V. M. (1999). Synth. React. Inorg. Met. Org. Chem. 29, 193–197. S1. Comment 8-Hydroxyquinoline and its derivatives have found extensive application as analytical reagents, e.g. in absorption spectrophotometry, fluorimetry, solvent extraction and partition chromatography (Bratzel et al., 1972). Some 8-hydroxy- quinoline derivatives and their complexes with transition metals demonstrate antibacterial activity (Patel & Patel,1999). Recently, the structure of 5,7-dibromosubstituted 8-hydroxyquinolinate amide-type compound, namely N,N-dicyclo- hexyl-2-(5,7-dibromoquinolin-8-yloxy)acetamide, (II), has been reported (Liu et al., 2007). Here, we have synthesized and carried out the structure determination of the title compound, (I), (Fig. 1). All bond lengths in (I) are within normal ranges (Allen et al., 1987) and comparable with those in the related compound (II). The sum of the angles around atoms N2 and C11 are 359.9° and 360.0°, respectively, implying a planar configuration. There is one intramolecular hydrogen bond, viz. N2—H2···N1 (Table 1), forming one larger eight- membered ring. The dihedral angle between the planes of the quinoline system and the benzene ring is 41.69 (1)°. The crystal packing is stabilized by intermolecular C18—H18C···O2 hydrogen bond (Table 1) and Br···O short-contact interactions. S2. Experimental To a solution of 5,7-dibromo-8-hydroxyquinoline (3.02 g, 10 mmol) in acetone (60 ml) were added 2-chloro-N-(4-meth- oxyphenyl)acetamide (2.0 g,10 mmol), K2CO3 (1.52 g, 11 mmol) and KI (0.5 g), and the resulting mixture was stirred at 333 K for 5 h. After cooling to room temperature, the mixture was washed three times with water and filtered. Colourless single crystals of (I) suitable for X-ray diffraction study were obtained by slow evaporation of an acetone solution over a period of 6 d. References Allen, F. H., Kennard, O., Watson, D. G., Brammer, L., Orpen, A. G. & Taylor, R. (1987). J. Chem. Soc. Perkin Trans. 2, pp. S1–19. Bratzel, M. P., Aaron, J. J., Winefordner, J. D., Schulman, S. G. & Gershon, H. (1972). Anal. Chem. 44, 1240–1245. k (2001) S A d SA k A S di i i Allen, F. H., Kennard, O., Watson, D. G., Brammer, L., Orpen, A. G. & Taylor, R. (1987). J. Chem. Soc. Perkin Trans. 2, pp. S1–19. Bratzel, M. P., Aaron, J. J., Winefordner, J. D., Schulman, S. G. & Gershon, H. (1972). Anal. Chem. 44, 1240–1245. Bruker (2001). SMART and SAINT. Bruker AXS Inc., Madison, Wisconsin, USA. u, J.-F., Tang, X.-F. & Wen, Y.-H. (2007). Acta Cryst. E63, o44 Experimental Patel, A. K. & Patel, V. M. (1999). Synth. React. Inorg. Met. Org. Chem. 29, 193–197. Crystal data C18H14Br2N2O3 Mr = 466.13 Crystal data Sheldrick, G. M. (1996). SADABS. University of Go¨ttingen, Germany. Sheldrick, G. M. (2008). Acta Cryst. A64, 112–122. Crystal data C18H14Br2N2O3 Mr = 466.13 Monoclinic, P21=n a = 8.7570 (18) A˚ o3294 Wen et al. Acta Cryst. (2010). E66, o3294 doi:10.1107/S1600536810048312 supporting information Acta Cryst. (2010). E66, o3294 supporting information Acta Cryst. (2010). E66, o3294 [https://doi.org/10.1107/S1600536810048312] 2-[(5,7-Dibromoquinolin-8-yl)oxy]-N-(2-methoxyphenyl)acetamide Yong-Hong Wen, Hong-Qing Qin and Hui-Ling Wen Acta Cryst. (2010). E66, o3294 [https://doi.org/10.1107/S1600536810048312] S1. Comment S3. Refinement E66, o3294 supporting information 2-[(5,7-Dibromoquinolin-8-yl)oxy]-N-(2-m Crystal data C18H14Br2N2O3 Mr = 466.13 Monoclinic, P21/n Hall symbol: -P 2yn a = 8.7570 (18) Å b = 8.7279 (17) Å c = 22.372 (5) Å β = 98.04 (3)° V = 1693.1 (6) Å3 Z = 4 Data collection Bruker SMART CCD area-detector diffractometer Radiation source: fine-focus sealed tube Graphite monochromator phi and ω scans Absorption correction: multi-scan (SADABS; Sheldrick, 1996) Tmin = 0.761, Tmax = 0.910 Refinement Refinement on F2 Least-squares matrix: full R[F2 > 2σ(F2)] = 0.044 wR(F2) = 0.093 S = 1.06 4027 reflections 231 parameters 1 restraint Primary atom site location: structure-invarian direct methods 2-[(5,7-Dibromoquinolin-8-yl)oxy]-N-(2-methoxyphenyl)acetamide Crystal data C18H14Br2N2O3 Mr = 466.13 Monoclinic, P21/n Hall symbol: -P 2yn a = 8.7570 (18) Å b = 8.7279 (17) Å c = 22.372 (5) Å β = 98.04 (3)° V = 1693.1 (6) Å3 Z = 4 F(000) = 920 Dx = 1.829 Mg m−3 Mo Kα radiation, λ = 0.71073 Å Cell parameters from 4165 reflections θ = 1.8–27.9° µ = 4.81 mm−1 T = 293 K Column, colourless 0.06 × 0.02 × 0.02 mm F(000) = 920 Dx = 1.829 Mg m−3 Mo Kα radiation, λ = 0.71073 Å Cell parameters from 4165 reflections θ = 1.8–27.9° µ = 4.81 mm−1 T = 293 K Column, colourless 0.06 × 0.02 × 0.02 mm F(000) = 920 Dx = 1.829 Mg m−3 Mo Kα radiation, λ = 0.71073 Å Cell parameters from 4165 reflections θ = 1.8–27.9° µ = 4.81 mm−1 T = 293 K Column, colourless 0.06 × 0.02 × 0.02 mm 12864 measured reflections 4027 independent reflections 3316 reflections with I > 2σ(I) Rint = 0.057 θmax = 27.9°, θmin = 1.8° h = −11→11 k = −11→8 l = −29→28 12864 measured reflections 4027 independent reflections 3316 reflections with I > 2σ(I) Rint = 0.057 θmax = 27.9°, θmin = 1.8° h = −11→11 k = −11→8 l = −29→28 Secondary atom site location: difference Fourier map Hydrogen site location: inferred from neighbouring sites H atoms treated by a mixture of independent and constrained refinement w = 1/[σ2(Fo2) + (0.0337P)2 + 1.4595P] where P = (Fo2 + 2Fc2)/3 (Δ/σ)max = 0.005 Δρmax = 0.83 e Å−3 Δρmin = −0.71 e Å−3 Secondary atom site location: difference Fourier map Hydrogen site location: inferred from neighbouring sites H atoms treated by a mixture of independent and constrained refinement Refinement on F2 Least-squares matrix: full R[F2 > 2σ(F2)] = 0.044 wR(F2) = 0.093 S = 1.06 4027 reflections 231 parameters 1 restraint Hydrogen site location: inferred from neighbouring sites H atoms treated by a mixture of independent and constrained refinement w = 1/[σ2(Fo2) + (0.0337P)2 + 1.4595P] where P = (Fo2 + 2Fc2)/3 (Δ/σ)max = 0.005 Δρmax = 0.83 e Å−3 Δρmin = −0.71 e Å−3 Primary atom site location: structure-invariant direct methods Acta Cryst. (2010). E66, o3294 S3. Refinement H atoms were positioned geometrically, with C—H = 0.95–0.99 Å, respectively, and constrained to ride on their parent atoms, with Uiso(H) = 1.2 Ueq(C). The amide proton was refined freely, giving a N—H bond distance of 0.898 (9) Å. sup-1 Acta Cryst. (2010). E66, o3294 sup-1 supporting information supporting information Figure 1 The molecular structure of (I), with atom labels and 30% probability displacement ellipsoids. Figure 2 The packing diagram of (I), viewed down the c axis, showing the intermolecular hydrogen bonds (dashed lines). Figure 1 Figure 1 Figure 1 g The molecular structure of (I), with atom labels and 30% probability displacement ellipsoids. The molecular structure of (I), with atom labels and 30% probability displacement ellipsoids. Figure 2 The packing diagram of (I), viewed down the c axis, showing the intermolecular hydrogen bonds (dashed lines). Figure 2 g The packing diagram of (I), viewed down the c axis, showing the intermolecular hydrogen bonds (dashed lines). g The packing diagram of (I), viewed down the c axis, showing the intermolecular hydrogen bonds (dashed lines). The packing diagram of (I), viewed down the c axis, showing the intermolecular hydrogen bonds (dashed lines). sup-2 Acta Cryst. (2010). Special details Geometry. All e.s.d.'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.'s are taken into account individually in the estimation of e.s.d.'s in distances, angles and torsion angles; correlations between e.s.d.'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.'s is used for estimating e.s.d.'s involving l.s. planes. Refinement. Refinement of F2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F2, conventional R-factors R are based on F, with F set to zero for negative F2. The threshold expression of F2 > σ(F2) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F2 are statistically about twice as large as those based on F, and R- factors based on ALL data will be even larger. sup-3 Acta Cryst. (2010). E66, o3294 Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å2) x y z Uiso/Ueq Br1 0.77254 (3) 0.41089 (3) 1.064695 (11) 0.01979 (6) Br2 0.16180 (3) 0.18905 (3) 1.014627 (12) 0.02468 (7) O1 0.81081 (18) 0.22709 (18) 0.94999 (7) 0.0173 (4) l atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å2) Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å2) x y z Uiso/Ueq sup-3 Acta Cryst. (2010). Special details E66, o3294 supporting information O2 1.0293 (2) 0.39129 (19) 0.84558 (8) 0.0241 (4) O3 0.8399 (2) −0.13107 (19) 0.82222 (8) 0.0262 (5) N2 0.9123 (2) 0.1556 (2) 0.84366 (9) 0.0171 (5) N1 0.5971 (2) 0.0662 (2) 0.87776 (9) 0.0190 (5) C1 0.6236 (3) 0.2953 (3) 1.01398 (11) 0.0171 (5) C2 0.4727 (3) 0.2851 (3) 1.02927 (11) 0.0180 (5) H2 0.4474 0.3365 1.0630 0.022* C3 0.3646 (3) 0.1998 (3) 0.99438 (11) 0.0189 (6) C4 0.3988 (3) 0.1219 (3) 0.94234 (11) 0.0168 (5) C5 0.2943 (3) 0.0290 (3) 0.90434 (11) 0.0203 (6) H5 0.1941 0.0150 0.9127 0.024* C6 0.3416 (3) −0.0401 (3) 0.85523 (12) 0.0222 (6) H6 0.2736 −0.1005 0.8297 0.027* C7 0.4956 (3) −0.0190 (3) 0.84358 (11) 0.0185 (6) H7 0.5262 −0.0676 0.8102 0.022* C8 0.5505 (3) 0.1348 (3) 0.92702 (11) 0.0167 (5) C9 0.6633 (3) 0.2237 (3) 0.96423 (10) 0.0144 (5) C10 0.8431 (3) 0.3518 (3) 0.91172 (11) 0.0199 (6) H10A 0.8976 0.4317 0.9362 0.024* H10B 0.7467 0.3946 0.8920 0.024* C11 0.9396 (3) 0.3002 (3) 0.86413 (11) 0.0167 (5) C12 0.9771 (3) 0.0828 (3) 0.79679 (11) 0.0171 (6) C13 1.0774 (3) 0.1524 (3) 0.76220 (11) 0.0219 (6) H13 1.1071 0.2536 0.7697 0.026* C14 1.1340 (3) 0.0712 (3) 0.71627 (12) 0.0267 (7) H14 1.2012 0.1188 0.6934 0.032* C15 1.0912 (3) −0.0787 (3) 0.70449 (12) 0.0270 (7) H15 1.1279 −0.1317 0.6733 0.032* C16 0.9923 (3) −0.1508 (3) 0.73961 (12) 0.0244 (6) H16 0.9645 −0.2526 0.7322 0.029* C17 0.9358 (3) −0.0713 (3) 0.78527 (11) 0.0194 (6) C18 0.8150 (3) −0.2923 (3) 0.82018 (13) 0.0284 (7) H18A 0.7705 −0.3216 0.7801 0.043* H18B 0.7460 −0.3198 0.8482 0.043* H18C 0.9116 −0.3443 0.8308 0.043* H2A 0.8405 (13) 0.106 (2) 0.8612 (8) 0.032 (8)* C1 0.6236 (3) 0.2953 (3) 1.01398 (11) 0.0171 (5) C2 0.4727 (3) 0.2851 (3) 1.02927 (11) 0.0180 (5) H2 0.4474 0.3365 1.0630 0.022* C3 0.3646 (3) 0.1998 (3) 0.99438 (11) 0.0189 (6) C4 0.3988 (3) 0.1219 (3) 0.94234 (11) 0.0168 (5) C5 0.2943 (3) 0.0290 (3) 0.90434 (11) 0.0203 (6) H5 0.1941 0.0150 0.9127 0.024* C6 0.3416 (3) −0.0401 (3) 0.85523 (12) 0.0222 (6) H6 0.2736 −0.1005 0.8297 0.027* C7 0.4956 (3) −0.0190 (3) 0.84358 (11) 0.0185 (6) H7 0.5262 −0.0676 0.8102 0.022* C8 0.5505 (3) 0.1348 (3) 0.92702 (11) 0.0167 (5) C9 0.6633 (3) 0.2237 (3) 0.96423 (10) 0.0144 (5) C10 0.8431 (3) 0.3518 (3) 0.91172 (11) 0.0199 (6) H10A 0.8976 0.4317 0.9362 0.024* H10B 0.7467 0.3946 0.8920 0.024* C11 0.9396 (3) 0.3002 (3) 0.86413 (11) 0.0167 (5) C12 0.9771 (3) 0.0828 (3) 0.79679 (11) 0.0171 (6) C13 1.0774 (3) 0.1524 (3) 0.76220 (11) 0.0219 (6) H13 1.1071 0.2536 0.7697 0.026* C14 1.1340 (3) 0.0712 (3) 0.71627 (12) 0.0267 (7) H14 1.2012 0.1188 0.6934 0.032* C15 1.0912 (3) −0.0787 (3) 0.70449 (12) 0.0270 (7) H15 1.1279 −0.1317 0.6733 0.032* C16 0.9923 (3) −0.1508 (3) 0.73961 (12) 0.0244 (6) H16 0.9645 −0.2526 0.7322 0.029* C17 0.9358 (3) −0.0713 (3) 0.78527 (11) 0.0194 (6) C18 0.8150 (3) −0.2923 (3) 0.82018 (13) 0.0284 (7) H18A 0.7705 −0.3216 0.7801 0.043* H18B 0.7460 −0.3198 0.8482 0.043* H18C 0.9116 −0.3443 0.8308 0.043* H2A 0.8405 (13) 0.106 (2) 0.8612 (8) 0.032 (8)* Atomic displacement parameters (Å2) U11 U22 U33 U12 U13 U23 Br1 0.02090 (11) 0.02052 (12) 0.01835 (12) −0.00516 (10) 0.00418 (10) −0.00258 (10) Br2 0.01560 (11) 0.02843 (13) 0.03130 (14) −0.00014 (10) 0.00784 (10) 0.00388 (11) O1 0.0136 (7) 0.0182 (8) 0.0208 (8) −0.0006 (7) 0.0048 (7) 0.0022 (7) O2 0.0284 (9) 0.0165 (8) 0.0304 (10) −0.0091 (7) 0.0149 (8) −0.0046 (7) O3 0.0305 (9) 0.0132 (8) 0.0377 (10) −0.0036 (7) 0.0139 (8) −0.0035 (8) N2 0.0194 (9) 0.0132 (9) 0.0200 (10) 0.0007 (8) 0.0076 (8) 0.0011 (8) N1 0.0197 (10) 0.0200 (10) 0.0168 (10) 0.0022 (8) 0.0012 (9) 0.0018 (8) C1 0.0177 (11) 0.0161 (11) 0.0175 (12) 0.0000 (9) 0.0019 (9) 0.0037 (9) Atomic displacement parameters (Å2) U11 U22 U33 U12 U13 U23 Br1 0.02090 (11) 0.02052 (12) 0.01835 (12) −0.00516 (10) 0.00418 (10) −0.00258 (10) Br2 0.01560 (11) 0.02843 (13) 0.03130 (14) −0.00014 (10) 0.00784 (10) 0.00388 (11) O1 0.0136 (7) 0.0182 (8) 0.0208 (8) −0.0006 (7) 0.0048 (7) 0.0022 (7) O2 0.0284 (9) 0.0165 (8) 0.0304 (10) −0.0091 (7) 0.0149 (8) −0.0046 (7) O3 0.0305 (9) 0.0132 (8) 0.0377 (10) −0.0036 (7) 0.0139 (8) −0.0035 (8) N2 0.0194 (9) 0.0132 (9) 0.0200 (10) 0.0007 (8) 0.0076 (8) 0.0011 (8) N1 0.0197 (10) 0.0200 (10) 0.0168 (10) 0.0022 (8) 0.0012 (9) 0.0018 (8) C1 0.0177 (11) 0.0161 (11) 0.0175 (12) 0.0000 (9) 0.0019 (9) 0.0037 (9) Acta Cryst. Acta Cryst. (2010). E66, o3294 Special details (2010). Special details E66, o3294 sup-4 supporting information supporting information sup 5 Acta Cryst (2010) E66 o3294 C2 0.0178 (11) 0.0173 (11) 0.0209 (12) 0.0003 (9) 0.0096 (10) 0.0026 (10) C3 0.0159 (11) 0.0180 (11) 0.0232 (12) 0.0030 (9) 0.0044 (10) 0.0098 (10) C4 0.0151 (10) 0.0173 (11) 0.0172 (12) 0.0031 (9) −0.0006 (9) 0.0085 (10) C5 0.0160 (11) 0.0179 (11) 0.0261 (13) −0.0019 (10) 0.0000 (10) 0.0044 (10) C6 0.0220 (12) 0.0176 (12) 0.0250 (13) −0.0035 (10) −0.0038 (11) 0.0029 (10) C7 0.0252 (12) 0.0149 (11) 0.0143 (11) −0.0027 (10) −0.0008 (10) −0.0010 (10) C8 0.0202 (11) 0.0138 (10) 0.0161 (11) 0.0007 (9) 0.0025 (10) 0.0048 (9) C9 0.0117 (10) 0.0141 (11) 0.0175 (11) 0.0006 (9) 0.0030 (9) 0.0035 (9) C10 0.0242 (12) 0.0127 (11) 0.0247 (12) −0.0028 (10) 0.0099 (10) −0.0012 (10) C11 0.0152 (10) 0.0158 (11) 0.0191 (12) 0.0012 (9) 0.0024 (9) 0.0012 (10) C12 0.0200 (11) 0.0160 (11) 0.0138 (11) 0.0025 (9) −0.0023 (10) 0.0001 (9) C13 0.0245 (12) 0.0184 (12) 0.0230 (13) 0.0010 (10) 0.0046 (11) 0.0030 (10) C14 0.0328 (14) 0.0299 (14) 0.0193 (13) 0.0003 (12) 0.0105 (11) 0.0010 (11) C15 0.0327 (14) 0.0300 (14) 0.0184 (13) 0.0084 (12) 0.0040 (11) −0.0057 (11) C16 0.0288 (13) 0.0183 (12) 0.0261 (13) 0.0028 (11) 0.0031 (11) −0.0082 (11) C17 0.0188 (11) 0.0197 (12) 0.0188 (12) 0.0018 (10) −0.0009 (10) −0.0010 (10) C18 0.0332 (14) 0.0164 (12) 0.0357 (15) −0.0043 (11) 0.0055 (13) −0.0017 (11) Geometric parameters (Å, º) Br1—C1 1.895 (2) C6—C7 1.420 (4) Br2—C3 1.896 (2) C6—H6 0.9300 O1—C9 1.374 (3) C7—H7 0.9300 O1—C10 1.437 (3) C8—C9 1.429 (3) O2—C11 1.230 (3) C10—C11 1.518 (3) O3—C17 1.362 (3) C10—H10A 0.9700 O3—C18 1.424 (3) C10—H10B 0.9700 N2—C11 1.352 (3) C12—C13 1.389 (4) N2—C12 1.411 (3) C12—C17 1.407 (3) N2—H2A 0.898 (9) C13—C14 1.395 (4) N1—C7 1.318 (3) C13—H13 0.9300 N1—C8 1.366 (3) C14—C15 1.376 (4) C1—C9 1.363 (3) C14—H14 0.9300 C1—C2 1.414 (3) C15—C16 1.398 (4) C2—C3 1.361 (3) C15—H15 0.9300 C2—H2 0.9300 C16—C17 1.383 (4) C3—C4 1.416 (3) C16—H16 0.9300 C4—C5 1.415 (3) C18—H18A 0.9600 C4—C8 1.421 (3) C18—H18B 0.9600 C5—C6 1.368 (4) C18—H18C 0.9600 C5—H5 0.9300 C9—O1—C10 115.11 (17) O1—C10—C11 111.60 (19) C17—O3—C18 117.6 (2) O1—C10—H10A 109.3 C11—N2—C12 127.0 (2) C11—C10—H10A 109.3 C11—N2—H2A 113.8 (14) O1—C10—H10B 109.3 C12—N2—H2A 119.1 (14) C11—C10—H10B 109.3 C7—N1—C8 117.5 (2) H10A—C10—H10B 108.0 C2 0.0178 (11) 0.0173 (11) 0.0209 (12) 0.0003 (9) 0.0096 (10) 0.0026 (10) C3 0.0159 (11) 0.0180 (11) 0.0232 (12) 0.0030 (9) 0.0044 (10) 0.0098 (10) C4 0.0151 (10) 0.0173 (11) 0.0172 (12) 0.0031 (9) −0.0006 (9) 0.0085 (10) C5 0.0160 (11) 0.0179 (11) 0.0261 (13) −0.0019 (10) 0.0000 (10) 0.0044 (10) C6 0.0220 (12) 0.0176 (12) 0.0250 (13) −0.0035 (10) −0.0038 (11) 0.0029 (10) C7 0.0252 (12) 0.0149 (11) 0.0143 (11) −0.0027 (10) −0.0008 (10) −0.0010 (10) C8 0.0202 (11) 0.0138 (10) 0.0161 (11) 0.0007 (9) 0.0025 (10) 0.0048 (9) C9 0.0117 (10) 0.0141 (11) 0.0175 (11) 0.0006 (9) 0.0030 (9) 0.0035 (9) C10 0.0242 (12) 0.0127 (11) 0.0247 (12) −0.0028 (10) 0.0099 (10) −0.0012 (10) C11 0.0152 (10) 0.0158 (11) 0.0191 (12) 0.0012 (9) 0.0024 (9) 0.0012 (10) C12 0.0200 (11) 0.0160 (11) 0.0138 (11) 0.0025 (9) −0.0023 (10) 0.0001 (9) C13 0.0245 (12) 0.0184 (12) 0.0230 (13) 0.0010 (10) 0.0046 (11) 0.0030 (10) C14 0.0328 (14) 0.0299 (14) 0.0193 (13) 0.0003 (12) 0.0105 (11) 0.0010 (11) C15 0.0327 (14) 0.0300 (14) 0.0184 (13) 0.0084 (12) 0.0040 (11) −0.0057 (11) C16 0.0288 (13) 0.0183 (12) 0.0261 (13) 0.0028 (11) 0.0031 (11) −0.0082 (11) C17 0.0188 (11) 0.0197 (12) 0.0188 (12) 0.0018 (10) −0.0009 (10) −0.0010 (10) C18 0.0332 (14) 0.0164 (12) 0.0357 (15) −0.0043 (11) 0.0055 (13) −0.0017 (11) sup-5 Acta Cryst. Special details E66, o3294 supporting information 121.5 (2) O2—C11—N2 125.5 (2) 120.03 (18) O2—C11—C10 119.3 (2) 118.47 (18) N2—C11—C10 115.2 (2) 119.6 (2) C13—C12—C17 118.8 (2) 120.2 C13—C12—N2 124.7 (2) 120.2 C17—C12—N2 116.5 (2) 121.6 (2) C12—C13—C14 120.4 (2) 119.32 (19) C12—C13—H13 119.8 119.06 (17) C14—C13—H13 119.8 125.1 (2) C15—C14—C13 120.6 (3) 116.7 (2) C15—C14—H14 119.7 118.2 (2) C13—C14—H14 119.7 119.5 (2) C14—C15—C16 119.6 (3) 120.3 C14—C15—H15 120.2 120.3 C16—C15—H15 120.2 119.5 (2) C17—C16—C15 120.2 (2) 120.3 C17—C16—H16 119.9 120.3 C15—C16—H16 119.9 123.2 (2) O3—C17—C16 124.8 (2) 118.4 O3—C17—C12 114.8 (2) 118.4 C16—C17—C12 120.4 (2) 123.7 (2) O3—C18—H18A 109.5 116.6 (2) O3—C18—H18B 109.5 119.7 (2) H18A—C18—H18B 109.5 122.3 (2) O3—C18—H18C 109.5 119.4 (2) H18A—C18—H18C 109.5 118.2 (2) H18B—C18—H18C 109.5 3 1.2 (4) N1—C8—C9—C1 179.8 (2) C3 −178.45 (18) C4—C8—C9—C1 −0.2 (3) 4 −0.8 (4) N1—C8—C9—O1 −3.4 (3) r2 −179.18 (17) C4—C8—C9—O1 176.6 (2) 5 178.9 (2) C9—O1—C10—C11 −138.9 (2) C5 −2.8 (3) C12—N2—C11—O2 −1.6 (4) 8 0.0 (3) C12—N2—C11—C10 175.2 (2) C8 178.33 (17) O1—C10—C11—O2 −149.5 (2) 6 179.9 (2) O1—C10—C11—N2 33.5 (3) 6 −1.2 (3) C11—N2—C12—C13 −2.4 (4) 7 0.7 (4) C11—N2—C12—C17 177.4 (2) 6 0.9 (3) C17—C12—C13—C14 1.0 (4) 1 −0.6 (4) N2—C12—C13—C14 −179.2 (2) 4 −1.4 (3) C12—C13—C14—C15 0.1 (4) 9 178.6 (2) C13—C14—C15—C16 −1.1 (4) 1 1.6 (3) C14—C15—C16—C17 1.1 (4) 1 −179.4 (2) C18—O3—C17—C16 10.3 (3) 9 −178.5 (2) C18—O3—C17—C12 −169.3 (2) 9 0.5 (3) C15—C16—C17—O3 −179.5 (2) 1 −177.3 (2) C15—C16—C17—C12 0.0 (4) C9—C1—C2 121.5 (2) O2—C11—N2 125.5 (2) C9—C1—Br1 120.03 (18) O2—C11—C10 119.3 (2) C2—C1—Br1 118.47 (18) N2—C11—C10 115.2 (2) C3—C2—C1 119.6 (2) C13—C12—C17 118.8 (2) C3—C2—H2 120.2 C13—C12—N2 124.7 (2) C1—C2—H2 120.2 C17—C12—N2 116.5 (2) C2—C3—C4 121.6 (2) C12—C13—C14 120.4 (2) C2—C3—Br2 119.32 (19) C12—C13—H13 119.8 C4—C3—Br2 119.06 (17) C14—C13—H13 119.8 C5—C4—C3 125.1 (2) C15—C14—C13 120.6 (3) C5—C4—C8 116.7 (2) C15—C14—H14 119.7 C3—C4—C8 118.2 (2) C13—C14—H14 119.7 C6—C5—C4 119.5 (2) C14—C15—C16 119.6 (3) C6—C5—H5 120.3 C14—C15—H15 120.2 C4—C5—H5 120.3 C16—C15—H15 120.2 C5—C6—C7 119.5 (2) C17—C16—C15 120.2 (2) C5—C6—H6 120.3 C17—C16—H16 119.9 C7—C6—H6 120.3 C15—C16—H16 119.9 N1—C7—C6 123.2 (2) O3—C17—C16 124.8 (2) N1—C7—H7 118.4 O3—C17—C12 114.8 (2) C6—C7—H7 118.4 C16—C17—C12 120.4 (2) N1—C8—C4 123.7 (2) O3—C18—H18A 109.5 N1—C8—C9 116.6 (2) O3—C18—H18B 109.5 C4—C8—C9 119.7 (2) H18A—C18—H18B 109.5 C1—C9—O1 122.3 (2) O3—C18—H18C 109.5 C1—C9—C8 119.4 (2) H18A—C18—H18C 109.5 O1—C9—C8 118.2 (2) H18B—C18—H18C 109.5 C9—C1—C2—C3 1.2 (4) N1—C8—C9—C1 179.8 (2) Br1—C1—C2—C3 −178.45 (18) C4—C8—C9—C1 −0.2 (3) C1—C2—C3—C4 −0.8 (4) N1—C8—C9—O1 −3.4 (3) C1—C2—C3—Br2 −179.18 (17) C4—C8—C9—O1 176.6 (2) C2—C3—C4—C5 178.9 (2) C9—O1—C10—C11 −138.9 (2) Br2—C3—C4—C5 −2.8 (3) C12—N2—C11—O2 −1.6 (4) C2—C3—C4—C8 0.0 (3) C12—N2—C11—C10 175.2 (2) Br2—C3—C4—C8 178.33 (17) O1—C10—C11—O2 −149.5 (2) C3—C4—C5—C6 179.9 (2) O1—C10—C11—N2 33.5 (3) C8—C4—C5—C6 −1.2 (3) C11—N2—C12—C13 −2.4 (4) C4—C5—C6—C7 0.7 (4) C11—N2—C12—C17 177.4 (2) C8—N1—C7—C6 0.9 (3) C17—C12—C13—C14 1.0 (4) C5—C6—C7—N1 −0.6 (4) N2—C12—C13—C14 −179.2 (2) C7—N1—C8—C4 −1.4 (3) C12—C13—C14—C15 0.1 (4) C7—N1—C8—C9 178.6 (2) C13—C14—C15—C16 −1.1 (4) C5—C4—C8—N1 1.6 (3) C14—C15—C16—C17 1.1 (4) C3—C4—C8—N1 −179.4 (2) C18—O3—C17—C16 10.3 (3) C5—C4—C8—C9 −178.5 (2) C18—O3—C17—C12 −169.3 (2) C3—C4—C8—C9 0.5 (3) C15—C16—C17—O3 −179.5 (2) C2—C1—C9—O1 −177.3 (2) C15—C16—C17—C12 0.0 (4) Acta Cryst. Special details (2010). Acta Cryst. (2010). E66, o3294 Special details (2010). E66, o3294 sup-6 supporting information Br1—C1—C9—O1 2.3 (3) C13—C12—C17—O3 178.5 (2) C2—C1—C9—C8 −0.7 (3) N2—C12—C17—O3 −1.3 (3) Br1—C1—C9—C8 178.94 (17) C13—C12—C17—C16 −1.1 (4) C10—O1—C9—C1 −90.4 (3) N2—C12—C17—C16 179.1 (2) C10—O1—C9—C8 92.9 (2) Hydrogen-bond geometry (Å, º) D—H···A D—H H···A D···A D—H···A N2—H2A···N1 0.90 (1) 2.24 (1) 3.065 (3) 153 (1) C18—H18C···O2i 0.96 2.53 3.342 (3) 142 Symmetry code: (i) x, y−1, z. supporting information pp g Br1—C1—C9—O1 2.3 (3) C13—C12—C17—O3 178.5 (2) C2—C1—C9—C8 −0.7 (3) N2—C12—C17—O3 −1.3 (3) Br1—C1—C9—C8 178.94 (17) C13—C12—C17—C16 −1.1 (4) C10—O1—C9—C1 −90.4 (3) N2—C12—C17—C16 179.1 (2) C10—O1—C9—C8 92.9 (2) Hydrogen-bond geometry (Å, º) D—H···A D—H H···A D···A D—H···A N2—H2A···N1 0.90 (1) 2.24 (1) 3.065 (3) 153 (1) Br1—C1—C9—O1 2.3 (3) C13—C12—C17—O3 178.5 (2) C2—C1—C9—C8 −0.7 (3) N2—C12—C17—O3 −1.3 (3) Br1—C1—C9—C8 178.94 (17) C13—C12—C17—C16 −1.1 (4) C10—O1—C9—C1 −90.4 (3) N2—C12—C17—C16 179.1 (2) C10—O1—C9—C8 92.9 (2) Hydrogen-bond geometry (Å, º) D—H···A D—H H···A D···A D—H···A N2—H2A···N1 0.90 (1) 2.24 (1) 3.065 (3) 153 (1) C18—H18C···O2i 0.96 2.53 3.342 (3) 142 Symmetry code: (i) x, y−1, z. Hydrogen-bond geometry (Å, º) sup-7 Acta Cryst. (2010). E66, o3294
https://openalex.org/W2968186403	https://www.frontiersin.org/articles/10.3389/fmicb.2019.02007/pdf	English	null	Regulation of Early Host Immune Responses Shapes the Pathogenicity of Avian Influenza A Virus	Frontiers in microbiology	2,019	cc-by	10,394	Regulation of Early Host Immune Responses Shapes the Pathogenicity of Avian Inﬂuenza A Virus Jiya Sun1,2†, Jingfeng Wang1,2†, Xuye Yuan1,2, Xiangwei Wu1,2, Tianqi Sui1,2, Aiping Wu1,2, Genhong Cheng1,2,3 and Taijiao Jiang1,2* Jiya Sun1,2†, Jingfeng Wang1,2†, Xuye Yuan1,2, Xiangwei Wu1,2, Tianqi Sui1,2, Aiping Wu1,2, Genhong Cheng1,2,3 and Taijiao Jiang1,2* 1 Suzhou Institute of Systems Medicine, Suzhou, China, 2 Center for Systems Medicine, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 3 Department of Microbiology, Immunology and Molecular Genetics, University of California, Los Angeles, Los Angeles, CA, United States Avian inﬂuenza A viruses (IAV) can cross the species barrier and cause disease in humans. Understanding the pathogenesis of avian IAV remains a challenge. Interferon- mediated antiviral responses and multiple cytokines production are important host cellular antiviral immunity against IAV infection. To elucidate the pathogenicity of avian IAV, a system approach was adopted to investigate dysregulation of the two host cellular antiviral immune responses in contrast with human IAV. As a result, we revealed that avian IAV not only disrupted normal early host cellular interferon-mediated antiviral responses, but also caused abnormal cytokines production through different pathways. For avian IAV infection, dysregulation of STAT2 was mainly responsible for abnormal cellular interferon-mediated antiviral responses, and IRF5 and NFKB1 played crucial roles in unusual cytokines production. In contrast, for human IAV infection, IRF1, IRF7, and STAT1 contributed to cellular cytokines production. Furthermore, differential activation of pattern recognition receptors (PRRs) likely led to avian IAV-related abnormal early host cellular antiviral immunity, where TLR7 and RIG-I were activated by avian and human IAV, respectively. Finally, a pathogenesis model was proposed that combined of early host cellular interferon-mediated antiviral responses with cytokines production could partly explain the pathogenicity of avian IAV. In conclusion, our study provides a new perspective of the pathogenesis of avian IAV, which will be helpful in preventing their infections in the future. Edited by: Denis Kainov, Norwegian University of Science and Technology, Norway Reviewed by: Yuebang Yin, Nankai University, China Qiwei Qin, South China Sea Institute of Oceanology (CAS), China Correspondence: Aiping Wu wap@ism.cams.cn Genhong Cheng gcheng@mednet.ucla.edu Taijiao Jiang taijiao@ibms.pumc.edu.cn †These authors have contributed equally to this work Reviewed by: Yuebang Yin, Nankai University, China Qiwei Qin, South China Sea Institute of Oceanology (CAS), China Correspondence: Aiping Wu wap@ism.cams.cn Genhong Cheng gcheng@mednet.ucla.edu Taijiao Jiang taijiao@ibms.pumc.edu.cn Specialty section: This article was submitted to Virology, a section of the journal Frontiers in Microbiology Received: 14 June 2019 Accepted: 15 August 2019 Published: 11 September 2019 Keywords: inﬂuenza, virus–host interaction, early immune response, gene network, interferon ORIGINAL RESEARCH published: 11 September 2019 doi: 10.3389/fmicb.2019.02007 ORIGINAL RESEARCH published: 11 September 2019 doi: 10.3389/fmicb.2019.02007 ORIGINAL RESEARCH published: 11 September 2019 doi: 10.3389/fmicb.2019.02007 MATERIALS AND METHODS In order to improve the ability to control avian IAV, there is an urgent need for a deep understanding of their pathogenicities. In order to improve the ability to control avian IAV, there is an urgent need for a deep understanding of their pathogenicities. g p g p g Computationally, the pathogenicity of avian IAV is often explored in two ways: identiﬁcation of viral genome mutations and characterizing host cellular responses by using in vitro cell lines (Li et al., 2011; Josset et al., 2014; Simon et al., 2015) or in vivo mammalian models (Belser and Tumpey, 2013; Morrison et al., 2014; Su et al., 2017). To date, quite a few avian IAV speciﬁc genome mutations have been reported to confer binding to the human-type receptor (Auewarakul et al., 2007), increase replication eﬃciency in mammalian cells (Czudai-Matwich et al., 2014) and antagonize interferon production (Li et al., 2006). Many studies have taken a systematic approach to investigate virus-induced host cellular transcriptomes (Li et al., 2011; Josset et al., 2014; Simon et al., 2015; Chasman et al., 2016) for elucidation of avian IAV pathogenesis. For instance, Li et al. (2011) performed a co-expression network analysis of transcriptomes under H5N1 infection and identiﬁed that keratinization process was a potential novel regulator of its pathogenesis. Josset et al. (2014) revealed that H7N9 speciﬁcally elicited host cellular responses related to regulating cell cycle and gene transcription. Chasman et al. (2016) inferred pathogenicity-related gene modules by integrating cellular transcriptomes involving highly and low pathogenic IAV. Although these ﬁndings provide some clues to the pathogenicity of avian IAV in the context of the complicated virus–host interaction, dysregulation of early host cellular antiviral immune responses has not been systematically investigated. Citation: Sun J, Wang J, Yuan X, Wu X, Sui T, Wu A, Cheng G and Jiang T (2019) Regulation of Early Host Immune Responses Shapes the Pathogenicity of Avian Inﬂuenza A Virus. Front. Microbiol. 10:2007. doi: 10.3389/fmicb.2019.02007 Inﬂuenza virus is a long-term threat to global public health. In contrast to human inﬂuenza A viruses (IAV) such as H1N1 (Du et al., 2017) that usually causes seasonal epidemic every year, avian IAV such as H5N1 (Creanga et al., 2017; Peng et al., 2017) and H7N9 (Wu et al., 2013) suddenly jump from their avian hosts to human and cause a high mortality rate, about 60% for H5N1 and 38% for H7N9 (Yu et al., 2008; Gao et al., 2013), which has brought serious social panic (Su et al., 2015). September 2019 \| Volume 10 \| Article 2007 1 Frontiers in Microbiology \| www.frontiersin.org Sun et al. Early Host Responses for Flu 1http://software.broadinstitute.org/gsea/msigdb/index.jsp Data Processing and DEGs Identiﬁcation Naturally, upon infection with IAV, host cells can recognize virus entry through the RIG-I signaling pathway (Loo and Gale, 2011), which leads to cellular immune responses including expression of antiviral response genes and production of multiple cytokines. Host cellular antiviral response genes are induced via activation of the type I interferon signaling pathway (Schneider et al., 2014; McNab et al., 2015), which is leveraged by host cells to build the ﬁrst defense line against virus invasion. Cytokines are cell-to-cell signaling proteins that can activate immune cells. During lethal inﬂuenza virus infection, dysregulation of early induced cytokines is likely associated with mortality (Vogel et al., 2014). Gerlach et al. (2013) observed signiﬁcant diﬀerences in early host cellular immune responses between seasonal and pandemic human IAV. Thus, it is reasonable to focus on early host cellular antiviral immune responses to decipher the pathogenesis of highly pathogenic avian IAV. Clustering of ASGs Based on the time-series transcriptomes under treatment by 500 U/ml IFN-α (GSE70217), the ASGs were ﬁrst divided into eight clusters on the basis of their log2FC values (the breaks were −3, −2, −1, 0, 1, 2, and 3) at 6 h, and further subdivided into small clusters by hierarchical clustering, which was based on gene expression vectors with three elements comprised of log2FC at 6 h and diﬀerences of log2FC between adjacent time points. The hclust function in R was used to perform hierarchical clustering. Pathway Annotation The ﬁve pathways related to cellular survival and death were from the KEGG pathway database. In this study, we systematically compared the two host cellular antiviral immune responses including interferon- meditated antiviral responses and cytokines production between avian IAV (H5N1 and H7N9) and human IAV (H1N1). Through focusing on a set of host cellular antiviral state genes (ASGs) and multiple cytokines, we proposed a novel uniﬁed model to explain the pathogenicity of highly pathogenic avian IAV, which resulted from dysregulation of early host cellular interferon-mediated antiviral responses and cytokines production. Frontiers in Microbiology \| www.frontiersin.org Dataset Collection Raw microarray data of high-quality Calu-3 cell transcriptomes treatment by interferon-alpha (IFN-α) (GSE70217), H1N1 (GSE80697 and GSE37571), H5N1 [GSE76599 and GSE28166 (Li et al., 2011)] or H7N9 (GSE69026) were downloaded from NCBI GEO database. The used wild-type inﬂuenza strains were A/California/04/2009 (H1N1), A/Vietnam/1203/2004 (H5N1), and A/Anhui/1/2013 (H7N9), respectively. Three mutant strains were H5N1-PB2-K627E, H5N1-NS1-trunc124 and H7N9-NS1- 103F/106M. K627E mutation in PB2 as well as 103F and 106M mutations in NS1 reduce viral replication and virulence in mammalian cells (Dankar et al., 2011; Min et al., 2013). The 90-amino-acid truncation at the C-terminus of NS1 reduces the virus capacity to antagonize host cellular antiviral1 responses (Hale et al., 2008). Data Processing and DEGs Identiﬁcation Background correction and between-arrays normalization were performed using limma (Ritchie et al., 2015) package in R. Control probes and low expressed probes were removed. Based on the cutoﬀof the 95th percentile of negative control probes on each array, probes that were at least 10% brighter than the negative controls were considered as being well expressed. Probes that expressed at least 10% of all of the arrays were used. For multiple probes with the same gene annotation, the probe with maximum mean expression intensity was ﬁnally chosen. Gene expression intensity was transformed with log2 before further downstream analysis. The plotMDS method from the limma package was used to remove outlier samples. Compared with matched mock, the limma package was employed to identify DEGs [\|log2(fold change, FC)\|≥1, p-value ≤0.05]. Transcription Factor Enrichment Analysis p y Gene sets related to transcription factor binding motifs were downloaded from the Molecular Signatures Database1 (v6.1). Transcription factor enrichment analysis was based on Fisher’s exact test with 22810 human protein-coding genes as background. The ﬁsher.test function in R was used to perform Fisher’s exact test. September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 2 Sun et al. Early Host Responses for Flu FIGURE 1 \| Schema of study. Upon infection of inﬂuenza virus, viral RNAs are rapidly recognized by the RIG-I-like receptor (RLR) of host cells. On the one hand, RLR-dependent pathway induces secretion of type I interferons. The type I interferons bind to their receptors and activate the canonical Jak-Stat pathway, which further induces the expression of hundreds of interferon-stimulated genes (ISGs). The ISGs enable host cells to establish antiviral state with thousands of upregulated or downregulated host cellular antiviral state genes (ASGs). On the other hand, Recognition of virus triggers production of multiple cytokines including type I interferons through complex pathways. In this study, regulation of cellular antiviral responses and cytokines production was comprehensively compared between low and highly pathogenic IAV FIGURE 1 \| Schema of study. Upon infection of inﬂuenza virus, viral RNAs are rapidly recognized by the RIG-I-like receptor (RLR) of host cells. On the one hand, RLR-dependent pathway induces secretion of type I interferons. The type I interferons bind to their receptors and activate the canonical Jak-Stat pathway, which further induces the expression of hundreds of interferon-stimulated genes (ISGs). The ISGs enable host cells to establish antiviral state with thousands of upregulated or downregulated host cellular antiviral state genes (ASGs). On the other hand, Recognition of virus triggers production of multiple cytokines including type I interferons through complex pathways. In this study, regulation of cellular antiviral responses and cytokines production was comprehensively compared between low and highly pathogenic IAV. Overview of Study In this study, we focused on regulation of host cellular type I interferon-mediated antiviral responses (hereafter the term host cellular antiviral responses specially referred to the interferon- mediated) and multiple cytokines production to understand the pathogenicity of highly pathogenic avian IAV (Figure 1). Overall, our analyses were comprised of three parts, in which the ﬁrst was related to regulation of host cellular antiviral responses, the second for regulation of multiple cytokines Construction of Regulatory Network of ASGs database (Carrasco Pro et al., 2018). Then, the PCC between a pair of TF and cytokine gene was calculated using time- series transcriptomes, in which only diﬀerentially expressed TFs and cytokines were considered. Finally, all pairs of TF-cytokine interactions with PCCs of at least 0.7 were used to construct a virus strain-speciﬁc regulatory network of cytokines. The LASSO algorithm developed in the glmnet R package was used to infer gene regulatory network by performing regularized linear regression (LR) between each of 1658 ASGs and 1124 TFs. Human TF list was from the paper by Narang et al. (2015). Prediction accuracy was evaluated using Pearson’s correlation coeﬃcient (PCC) between real and predicted gene expression levels. To assess inﬂuences of overﬁtting on built gene regulatory models, the LR model was used to predict expression levels of ASGs based on their assigned transcription factors (TFs) by LASSO models. The cor.test function in R was used to calculate PCC. Construction of IAV Strain-Speciﬁc Regulatory Network of Cytokines First, the manually curated regulatory relationships between TFs and cytokine genes were downloaded from the CytReg September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 3 Sun et al. Early Host Responses for Flu FIGURE 2 \| Characterization of ASGs under interferon treatment. (A) The number of interferon stimulated DEGs. The red and blue colors represent up- and downregulated DEGs, respectively. All (left) means all of DEGs and TF (right) for TF DEGs. (B) The number of overlap DEGs between two separate sets of DEGs. The numbers in each square cell mean overlap gene count and the numbers in triangle cell represent total of up- or downregulated DEG at indicated times. Please ﬁnd an overlap gene count following red arrows. (C) Dynamic expression proﬁles of survival- and death-related ASGs. In the left, black color means existence of gene in corresponding pathways, and white for absence. In the right, each cell represents fold change of gene expression with log2 transformation. The abbreviations PI3K-Akt, MAPK, Apoptosis, P53 and Autophagy denote PI3K-Akt signaling pathway, MAPK signaling pathway, Apoptosis pathways, P53 signaling pathway, and regulation of autophagy, respectively. The genes with red star (∗) are TFs. (D) Expression trends of the 1819 ASGs under interferon treatment. Lines with different colors represent eight big gene groups, each of which is clustered into small gene clusters with the same color. FIGURE 2 \| Characterization of ASGs under interferon treatment. (A) The number of interferon stimulated DEGs. The red and blue colors represent up- and downregulated DEGs, respectively. All (left) means all of DEGs and TF (right) for TF DEGs. (B) The number of overlap DEGs between two separate sets of DEGs. The numbers in each square cell mean overlap gene count and the numbers in triangle cell represent total of up- or downregulated DEG at indicated times. Please ﬁnd an overlap gene count following red arrows. (C) Dynamic expression proﬁles of survival- and death-related ASGs. In the left, black color means existence of gene in corresponding pathways, and white for absence. In the right, each cell represents fold change of gene expression with log2 transformation. The abbreviations PI3K-Akt, MAPK, Apoptosis, P53 and Autophagy denote PI3K-Akt signaling pathway, MAPK signaling pathway, Apoptosis pathways, P53 signaling pathway, and regulation of autophagy, respectively. The genes with red star (∗) are TFs. (D) Expression trends of the 1819 ASGs under interferon treatment. Construction of IAV Strain-Speciﬁc Regulatory Network of Cytokines (A) Comparison of regulation patterns of the 44 ASGs during infection of human and avian IAV. High PCCs between true and predicted gene expression levels represent high prediction accuracies, which indicates that normal regulation patterns induced by interferon are kept by IAV, while low PCCs for disturbed regulation patterns. (B) Regulatory network of the 44 ASGs. For the 44 ASGs, their potential upstream regulators are given by the corresponding LASSO models with coefﬁcients at least 0.1. The wider edges represent larger coefﬁcients. The red edges are related to ﬁve key TFs with yellow color. h f h d h h d ll f h d b d h FIGURE 4 \| Aberrant regulation patterns of the 44 early upregulated ASGs upon infection of avian IAV. (A) Comparison of regulation patterns of the 44 ASGs during infection of human and avian IAV. High PCCs between true and predicted gene expression levels represent high prediction accuracies, which indicates that normal regulation patterns induced by interferon are kept by IAV, while low PCCs for disturbed regulation patterns. (B) Regulatory network of the 44 ASGs. For the 44 ASGs, their potential upstream regulators are given by the corresponding LASSO models with coefﬁcients at least 0.1. The wider edges represent larger coefﬁcients. The red edges are related to ﬁve key TFs with yellow color. FIGURE 4 \| Aberrant regulation patterns of the 44 early upregulated ASGs upon infection of avian IAV. (A) Comparison of regulation patterns of the 44 ASGs during infection of human and avian IAV. High PCCs between true and predicted gene expression levels represent high prediction accuracies, which indicates that normal regulation patterns induced by interferon are kept by IAV, while low PCCs for disturbed regulation patterns. (B) Regulatory network of the 44 ASGs. For the 44 ASGs, their potential upstream regulators are given by the corresponding LASSO models with coefﬁcients at least 0.1. The wider edges represent larger coefﬁcients. The red edges are related to ﬁve key TFs with yellow color. All of the used transcriptomes were obtained using the same microarray platform to alleviate technical noise. In order to further avoid confusing inconsistencies due to diﬀerent cell types, the Calu-3 cell line with many publicly available transcriptomes related to IAV infections were used. Construction of IAV Strain-Speciﬁc Regulatory Network of Cytokines Lines with different colors represent eight big gene groups, each of which is clustered into small gene clusters with the same color. production, and the last for generation and validation of a pathogenesis model of avian IAV. To make our results more reliable, three analysis groups related to IAV infections were designed (Supplementary Table S1). The ﬁrst was the discovery group with datasets from the same laboratory, which were used to identify avian IAV strain-speciﬁc host cellular immune response patterns. The second was validation group with independent datasets from other laboratories, which were collected to validate September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 4 Early Host Responses for Flu Sun et al. GURE 3 \| Perturbation of ASGs upon wild-type IAV infection. (A) Expression trends of the 1819 ASGs during infection of human and avian IAV. These lines have e same meanings as those of Figure 2D. For the same IAV strain, the discovery and validation datasets are from infection at the same MOI. (B) Expression rrelations between the 44 early upregulated ASGs. The top and left annotation colors for different gene groups have the same meanings as those of lines in gure 2D. In order to highlight strong correlations, highly positive correlations with PCCs ≥0.7 are set as red color, highly negative correlations with PCCs ≤–0.7 blue color, and weak correlations with PCCs > –0.7 and PCCs < 0.7 for white color. FIGURE 3 \| Perturbation of ASGs upon wild-type IAV infection. (A) Expression trends of the 1819 ASGs during infection of human and avian IAV. These lines have the same meanings as those of Figure 2D. For the same IAV strain, the discovery and validation datasets are from infection at the same MOI. (B) Expression correlations between the 44 early upregulated ASGs. The top and left annotation colors for different gene groups have the same meanings as those of lines in Figure 2D. In order to highlight strong correlations, highly positive correlations with PCCs ≥0.7 are set as red color, highly negative correlations with PCCs ≤–0.7 for blue color, and weak correlations with PCCs > –0.7 and PCCs < 0.7 for white color. September 2019 \| Volume 10 \| Article 2007 5 Frontiers in Microbiology \| www.frontiersin.org Early Host Responses for Flu Sun et al. FIGURE 4 \| Aberrant regulation patterns of the 44 early upregulated ASGs upon infection of avian IAV. Construction of IAV Strain-Speciﬁc Regulatory Network of Cytokines The Calu-3 cell is a human airway epithelial cell line from bronchial submucosal gland that is a major source of airway surface liquid, mucins, and other immunologically active substances in human lungs (Zhu et al., 2010). In addition, time points after H1N1 infection the response patterns from the discovery group. The third was the mutation group with datasets involving avian IAV mutant strains from the same laboratory as the discovery group, which were utilized to further check the response patterns under wild- type avian IAV infections. The interferon dataset under IFN-α treatment was also from the same laboratory as the discovery group. For the three groups, the human or avian IAV strains had the same multiplicity of infection (MOI), in which the MOI of H1N1 was 3, and those of H7N9 and H5N1 were 1. September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 6 Sun et al. Early Host Responses for Flu FIGURE 5 \| IAV strain-speciﬁc regulatory network of cytokines. (A–C) H1N1, H5N1 and H7N9 speciﬁc cytokine transcriptional regulation network. The different node colors represent differentially expressed cytokines speciﬁc for one strain or shared by two or more strains. The wider edges represent larger PCCs between a pair of TF and cytokine gene. FIGURE 5 \| IAV strain-speciﬁc regulatory network of cytokines. (A–C) H1N1, H5N1 and H7N9 speciﬁc cytokine transcriptional regulation network. The different node colors represent differentially expressed cytokines speciﬁc for one strain or shared by two or more strains. The wider edges represent larger PCCs between a pair of TF and cytokine gene. FIGURE 5 \| IAV strain-speciﬁc regulatory network of cytokines. (A–C) H1N1, H5N1 and H7N9 speciﬁc cytokine transcriptional regulation network. The different node colors represent differentially expressed cytokines speciﬁc for one strain or shared by two or more strains. The wider edges represent larger PCCs between a pair of TF and cytokine gene September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 7 Early Host Responses for Flu Sun et al. were no more than 48 h and those for H5N1 and H7N9 infections were within 24 h (Supplementary Table S1). delayed responses (early silent or late-response) to IFN-α treatment were applied to investigation of host cellular antiviral state changes during human and avian IAV infections. As expected, we observed virus strain-speciﬁc dynamic regulation of cellular antiviral state in discovery datasets (Figure 3A left column). Deﬁnition and Clustering of ASGs g Establishment of interferon-mediated antiviral state provides a crucial initial line of host defense against virus invasion (Levy and Garcia-Sastre, 2001). To evaluate host cellular antiviral state, a set of ASGs was ﬁrst deﬁned using time-series transcriptomes under 500 U/ml IFN-α treatment (Supplementary Table S1). As a result, we identiﬁed 1819 ASGs that exhibited signiﬁcant diﬀerential expression for at least one time point. In order to reasonably cluster these ASGs, we characterized their dynamic expressions under interferon treatment. Firstly, we found that the numbers of downregulated ASGs with the maximum at 6 h gradually decreased over the period of time, which was very diﬀerent from those of upregulated ASGs that reached the maximum at 12 h (Figure 2A). Interestingly, the numbers of ASGs and signiﬁcantly diﬀerentially expressed TFs were highly correlated (Figure 2A), which supported that transcription of ASGs was under control in a cascade manner due to interferon treatment alone. This fact was further conﬁrmed by TF enrichment analysis, where upregulated ASGs were signiﬁcantly enriched in IRF-related binding motifs at 6 and 12 h, while downregulated ASGs were dominated by other TFs such as SP1, ELK1, E4F1, ETS2, and SRF at 6 h (Supplementary Table S2). Secondly, downregulated ASGs showed time-speciﬁc expression dynamics, whereas many upregulated ASGs were still highly expressed at the next time point (Figure 2B). Compared with upregulated ASGs that rarely became downregulated at late infection stages, about 26.7% (212/788) of downregulated ASGs at 6 h changed into high expression at 12 h. Thirdly, we observed that the 66 ASGs from survival- and death-related signaling pathways exhibited dynamical expressions over times, where the majority of genes were downregulated and minor were upregulated at 6 h (Figure 2C). This was consistent with the fact that host cell can modulate cellular survival-death balance for its antiviral immunity (Upton and Chan, 2014). Construction of IAV Strain-Speciﬁc Regulatory Network of Cytokines For low pathogenic human IAV H1N1, early upregulated ASGs were highly induced from the early to late infection stages, whereas early downregulated and late-response ASGs were suppressed. In contrast, highly pathogenic avian IAV H5N1 and H7N9 showed big diﬀerences. For H5N1, early upregulated ASGs were moderately induced in the early stage and remarkably suppressed in the late stage. Notably, the changes of cellular antiviral state induced by H7N9 were completely unexpected, in which early upregulated ASGs were initially suppressed and progressively became moderately activated, while early downregulated ASGs were initially activated and gradually suppressed. Although H5N1 and H7N9 are all avian IAV, our results strongly suggested that they had big diﬀerences in viral survival strategy. Furthermore, these ﬁndings were well supported in the independent validation datasets (Figure 3A right column). IAV Avian IAV Caused Distinctive Expression Trends of Early Response ASGs Next, the above mentioned 18 gene clusters with rapid responses (early upregulated and early downregulated) or Avian IAV Caused Distinctive Expression Trends of Early Response ASGs IAV As stated above, highly pathogenic avian IAV caused distinctive time-speciﬁc gene expression trends of the ASGs. Besides, impaired gene synchronization between the ASGs should also be associated with the pathogenicity of avian IAV. To test this, we further investigated diﬀerential co-expression patterns of the ASGs between human and avian IAV. Here, gene co-expression was used to represent gene synchronization. According to the gene ordering from clustering of the ASGs (Supplementary Table S3), we visualized their pairwise PCC matrix (Supplementary Figure S2). After interferon treatment, it was obviously observed that early upregulated (from gene group 8 to 6) and early downregulated ASGs (from gene group 1 to 3) formed the most highly correlated modules, while early silent ASGs (from gene group 4 to 5) formed several moderately correlated modules (Supplementary Figure S2a). In contrast to early downregulated ASGs that most consisted of only one module, early upregulated ASGs corresponded to two clear modules, where one was from gene group 8 and 7 and the other from gene group 6. During human and avian IAV infections, the regular interferon-induced co-expression patterns of ASGs were widely perturbed but still indistinctly observed (Supplementary Figures S2b–d). Due to interferon treatment alone, the early downregulated ASGs was likely caused by the early downregulated ASGs, which was supported by TF enrichment analysis (Supplementary Table S2). Hence, we further only visualized the early upregulated group 8, 7, and 6, in which, interestingly, many ASGs showed conserved co- expression patterns between interferon treatment and H1N1 infection (Supplementary Figures S2e–h). Based on these observed time-speciﬁc expression features of the ASGs, we ﬁrst divided the 1819 ASGs into eight gene groups on basis of their expression levels at the early time point (6 h), and then used gene expression changes between adjacent time points to further cluster each gene group into small gene clusters. Within each gene cluster, similar expression trends between genes demonstrated that our clustering approach had good performances (Supplementary Figure S1). In total, we identiﬁed 18 gene clusters that were prepared for evaluating the regulation of host cellular antiviral state during IAV infection (Supplementary Figure S1 and Supplementary Table S3). Moreover, the host cellular antiviral state established by IFN-α provided a reference to compare antiviral state changes during human and avian IAV infections (Figure 2D). Avian IAV Disrupted Interferon-Induced Normal Regulation of the 44 Early Upregulated ASGs p g The ﬁnding that co-expression of the 44 early upregulated ASGs was made weaker by H5N1 and H7N9 (Figure 3B) likely resulted from impaired gene regulations caused by avian IAV. To verify this, we ﬁrst employed the LASSO algorithm (Omranian et al., 2016) to build regulatory relationships between 1124 TFs and 1658 response ASGs in interferon treated cells, and then applied these gene regulatory models to predict expression levels of response ASGs in IAV infected cells. For these built gene regulatory models, the numbers of predicted TFs for response ASGs had an approximate normal distribution with the peak 12 (Supplementary Figure S4a). To examine inﬂuences of overﬁtting on the gene regulatory models, the widely used LR algorithm was adopted to predict expression levels of response ASGs in IAV infected cells, which was based on their assigned TFs by LASSO models. Overall, the LASSO and LR models showed very similar predictive powers for most response ASGs (Supplementary Figure S4b and Supplementary Table S5), indicating that the regulatory models of these response ASGs given by the LASSO algorithm were reliable. p For 113 human cytokines, there were 58, 50, and 30 cytokines with signiﬁcant diﬀerential expressions for H1N1, H5N1, and H7N9, respectively, (Supplementary Figure S6), in which a few cytokines were strain-speciﬁc and many were shared by two or three IAV strains. Among these shared cytokines, we noticed that H5N1 always exhibited high expression levels while H1N1 and H7N9 showed low expression levels in the early infection stage (Supplementary Figure S6). However, we also observed that several cytokines were highly induced by H1N1 or H7N9 infections in the early stage. For example, CXCL10 and CCL5 were for H1N1 and CXCL5 for H7N9. Moreover, eight type I interferon genes including IFNB1, IFNA4, IFNA6, IFNA7, IFNA8, IFNA10, IFNA14, and IFN16 were more highly induced by H5N1 and repressed by H7N9 in the early stage. There are 13 human IFN-α subtypes, all utilizing a single type I IFN receptor (Gibbert et al., 2013). When treating human plasmacytoid dendritic cells by using various stimuli, Szubin et al. (2008) observed a rigid IFN-α subtype response pattern, in which each subtype was induced at similar relative levels for diﬀerent stimuli. With respect to induction of ISGs, Moll et al. (2011) classiﬁed IFN-α subtypes into low, intermediate and high activity, which was conﬁrmed by the protection of cells against inﬂuenza virus infection. Distinctive Regulation of Cytokines Production Between Human and Avian IAV Thus far, the above results had demonstrated that highly pathogenic avian IAV can disrupt early host cellular antiviral responses in contrast to low pathogenic human IAV. In this work, host cellular antiviral responses were evaluated by the ASG genes that were deﬁned by using transcriptomes under IFN-α treatment. During IAV infection, interferon, a type of cytokine, is usually rapidly induced. Although type I interferon can initiate host cellular antiviral responses, it actually induced few cytokines. In the dataset under IFN-α treatment (Supplementary Table S1), we observed that only 7 out of 113 human cytokines were induced with low expression levels. For severe inﬂuenza, complications or ultimately death are often associated with cytokine storm (Liu et al., 2016). Therefore, we further checked diﬀerences in regulation of multiple cytokines production between human and avian IAV. Avian IAV Caused Distinctive Expression Trends of Early Response ASGs y p Next, the above mentioned 18 gene clusters with rapid responses (early upregulated and early downregulated) or To further identify the consistent co-expression genes induced by interferon treatment and H1N1 infection, we clustered the September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 8 Early Host Responses for Flu Sun et al. FIGURE 6 \| The pathogenesis model of avian IAV (A) A representation of the proposed pathogenesis model. (B) Dynamic expression levels of key genes involved in FIGURE 6 \| The pathogenesis model of avian IAV (A) A representation of the proposed pathogenesis model. (B) Dynamic expression levels of key genes involved in the pathogenesis model. The node size represents signiﬁcance of gene differential expression, and the node color represents up-regulation (red), down-regulation (blue) or not signiﬁcant (gray). September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 9 Early Host Responses for Flu Sun et al. PCC matrix of gene group 8 and 7. As a result, both of them exhibited a big co-expression gene module with 41 genes for interferon and 43 genes for H1N1 (Supplementary Figure S3). The two modules had 40 overlap genes, indicating their important roles in host cellular antiviral responses. Hence, co- expression of the 44 union genes from the two big gene clusters were further checked under avian IAV infections. Clearly, co- expression of the 44 early upregulated ASGs were signiﬁcantly reduced by H5N1 and H7N9 (Figure 3B top). Furthermore, these ﬁndings were well validated by the independent validation datasets (Figure 3B bottom). Hence, avian IAV not only modulated the expression levels of early response ASGs, but also speciﬁcally disrupted the expression synchronization of early upregulated ASGs. including STAT1, IRF7, and TRIM22 showed good prediction performances (Figure 4A and Supplementary Figure S5). On the contrary, STAT2 was not well predicted for avian IAV, indicating that dysregulation of STAT2 was most responsible for impaired co-expression patterns of the 44 genes. In addition, regulation of IRF9 was more disrupted by H5N1 than that of H7N9. In summary, our regulatory network explained impaired co- expression of the 44 genes and provided valuable insights into their potential regulators when binding motifs of many TFs were not available for now. Avian IAV Disrupted Interferon-Induced Normal Regulation of the 44 Early Upregulated ASGs Thus, regulating production of multiple types of type I interferons was crucial to the severity of IAV infection. To determine whether normal gene regulation was disrupted by human or avian IAV, the PCCs between true and predicted expression levels were calculated, where high PCC indicated that the normal regulation induced by interferon was kept. Among the 44 early upregulated ASGs, we found that their regulation patterns were well maintained during H1N1 infection and more disrupted by avian IAV (Figure 4A). To further identify key TFs of the 44 ASGs, a meaningful regulatory network was built by selecting TFs with coeﬃcients at least 0.1 in the LASSO models (Supplementary Table S4). Our regulatory network (Figure 4B) revealed that four TFs including IRF7, IRF9, STAT1, and STAT2 regulated most genes together or alone. In addition, TRIM22, a potential transcription factor, was predicted to contribute to expression levels of speciﬁc genes together with STAT1, STAT2, IRF9, and IRF7. For example, two TFs, including TRIM22 and IRF7, regulated MX1 and MX2, expression levels of which were accurately predicted in H1N1, H5N1, and H7N9 infected cells (Figure 4A and Supplementary Figure S5). Moreover, for H1N1, H5N1, and H7N9, expression levels of three TFs To further explore diﬀerences in regulation of these diﬀerentially expressed cytokines between human and avian September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 10 Sun et al. Early Host Responses for Flu FIGURE 7 \| Validation of the pathogenesis model using avian IAV mutants. (A) Expression trends of all ASGs. The line colors have the same meanings as those of Figure 2D. (B) Expression correlations between the 44 early upregulated ASGs. The annotation colors and gene order are the same as those of Figure 3B. (C) Dynamic expression levels of key genes involving the pathogenesis model. The node size represents signiﬁcance of gene differential expression, and the node color represents up-regulation (red), down-regulation (blue) or not signiﬁcant (gray). FIGURE 7 \| Validation of the pathogenesis model using avian IAV mutants. (A) Expression trends of all ASGs. The line colors have the same meanings as those of Figure 2D. (B) Expression correlations between the 44 early upregulated ASGs. The annotation colors and gene order are the same as those of Figure 3B. (C) Dynamic expression levels of key genes involving the pathogenesis model. Avian IAV Disrupted Interferon-Induced Normal Regulation of the 44 Early Upregulated ASGs The node size represents signiﬁcance of gene differential expression, and the node color represents up-regulation (red), down-regulation (blue) or not signiﬁcant (gray). FIGURE 7 \| Validation of the pathogenesis model using avian IAV mutants. (A) Expression trends of all ASGs. The line colors have the same meanings as those of Figure 2D. (B) Expression correlations between the 44 early upregulated ASGs. The annotation colors and gene order are the same as those of Figure 3B. (C) Dynamic expression levels of key genes involving the pathogenesis model. The node size represents signiﬁcance of gene differential expression, and the node color represents up-regulation (red), down-regulation (blue) or not signiﬁcant (gray). September 2019 \| Volume 10 \| Article 2007 11 Frontiers in Microbiology \| www.frontiersin.org Early Host Responses for Flu Sun et al. IAV infections, we built their strain-speciﬁc transcriptional regulatory networks. Surprisingly, it was observed that human and avian IAV took very diﬀerent approaches to control cytokines production. For H1N1, three regulators including IRF1, IRF7, and STAT1 played critical roles in regulation of cytokines production (Figure 5A). In contrast, H5N1 showed distinctive regulatory pathways, where IRF5 and NFKB1 were responsible for regulation of most cytokines including IFNB1 and IFNA4 (Figure 5B). Notably, for H7N9, IRF5 together with IRF7, contributed to regulation of IFNB1 and IFNA4 despite a small network available due to moderate host cellular immune responses (Figure 5C). These results demonstrated that dysregulation of multiple cytokines production during avian IAV infection arose from activation of completely diﬀerent signaling pathways in contrast to human IAV. Validation of the Pathogenesis Model To further check the proposed pathogenesis model, we used transcriptomes under infection of low pathogenic avian IAV mutants. If the aberrant regulation of host cellular antiviral responses and cytokines production induced by wild-type avian IAV can be rescued, the pathogenesis model will be more reliable. Here, we focused on three mutants including H5N1-PB2-K627E, H5N1-NS1-trunc124, and H7N9-NS1-103F/106M. For H5N1- PB2-K627E, the gene expression trends of the whole ASGs became very similar to those under 500 U/ml IFN-α treatment (Figures 2A, 7D), which likely resulted from decreased viral replication in mammal cells due to the avian speciﬁc mutation PB2-K627E (Subbarao et al., 1993). However, the co-expression pattern of the 44 ASGs was still disrupted (Figures 3B, 7B). A Pathogenesis Model of Avian IAV A Pathogenesis Model of Avian IAV Based on the above ﬁndings, we proposed a uniﬁed model (Figure 6A) for explanation of avian IAV pathogenesis, which was centered on cooperation of the upstream interferon production (denoted by U) and the downstream interferon- induced antiviral responses (denoted by D). For H1N1, the low pathogenicity arose from high cooperation between the U and D with early moderate and late-high responses. However, two highly pathogenic avian IAV strains showed big diﬀerences. For H5N1, the conﬂict between the high U and low D over times caused the high pathogenicity. In contrast, H7N9 exhibited delayed but cooperative features between the U and D, in which the interferon production and antiviral responses were suppressed in the early stage, but remarkably increased in the late stage. These strain-speciﬁc patterns between the U and D were clearly seen from the dynamic patterns of interferon production represented by IFNB1 and IFNA4 (Figure 6B middle row), and antiviral responses represented by MX1 and TRIM22 (Figure 6B bottom row). For interferon production and antiviral responses, their dynamic expression levels were in line with those of their corresponding key TF regulators (Figures 6A,B). Most importantly, it was observed that the key TF IRF1 that regulated H1N1-induced cytokines production was activated by H1N1 but suppressed by H5N1 and H7N9, while the key TF IRF5 for regulation of cytokines production during infection of avian IAV was activated by H5N1 and H7N9 but suppressed by H1N1 (Figure 6B middle row). The mutually exclusive expression of IRF1 and IRF5 prompted us to infer the underlying reasons. Relying on literature search, we found that TLR7, a type of pattern recognition receptor (PRR) for recognizing single strand RNA virus, can activate IRF5, which further induces type I interferon production (Schoenemeyer et al., 2005) and culminates in the activation of the transcription factor NF-KB that controls the expression of an array of inﬂammatory cytokines (Kawai and Akira, 2007). Interestingly, the expression of TLR7 was activated by H5N1 and H7N9 but suppressed by H1N1, while the canonical PRR RIG-I (also known as DDX58) was suppressed by H5N1 and H7N9 but activated by H1N1 (Figure 6B top row). The distinct usages of virus recognition receptors provided a solid support for our pathogenesis model. Avian IAV Disrupted Interferon-Induced Normal Regulation of the 44 Early Upregulated ASGs Furthermore, we observed that dynamic expression levels of two PRRs RIG-I and TLR7, key regulators of cytokines production such as IRF5, and antiviral responses related regulators such as STAT2 and IRF9 were still similar to those of wild-type H5N1 (Figure 7C). These evidences strongly supported that dysregulation of host cellular antiviral responses and cytokines production during H5N1 infection mainly arose from the nature of virus itself because decreased viral replication still caused their dysregulation. For H5N1-NS1-trunc124 that lost the ability of blocking IFN-β production (Qian et al., 2017), it was observed that the gene expression trends of the whole ASGs became very similar to those of H1N1 (Figures 3A, 7A) and the co-expression pattern of the 44 ASGs was rescued (Figures 3B, 7B). Actually, production of IFN-β was highly induced by the H5N1-NS1- trunc124 mutant. Interestingly, we also observed activation of RIG-I, repression of TLR7, and activation of IRF1, STAT2, and IRF9 during H5N1-NS1-trunc124 infection (Figure 7C). These observations were consistent with the fact that viral protein NS1 played a critical role in H5N1 pathogenicity (Zhou et al., 2010). Diﬀerent from the two H5N1 mutants, the H7N9 mutant H7N9- NS1-103F/106M cannot rescue the gene expression trends of the whole ASGs (Figure 7A) and the co-expression pattern of the 44 ASGs (Figure 7B). During H7N9-NS1-103F/106M infection, however, we observed that TLR7 became repressed although RIG-I and key regulators of cytokines production and antiviral responses were not rescued (Figure 7C). These unexpected results suggested that this pair of NS1 mutations contributed less to H7N9 pathogenicity. Taken together, the data from the mutated avian IAV strains demonstrated the rationality of our proposed pathogenesis model. Frontiers in Microbiology \| www.frontiersin.org DISCUSSION In this study, we proposed a pathogenesis model of avian IAV by focusing on regulation of host cellular antiviral responses and cytokines production. In contrast to H1N1, both H5N1, and H7N9 disrupted normal early host cellular antiviral responses and cytokines production. However, time-speciﬁc cooperative patterns of cytokines production and early host cellular antiviral responses were very diﬀerent between H5N1 and H7N9. These ﬁndings were based on the Calu-3 cell from bronchial submucosal gland that, in humans, is preferentially attached September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 12 Early Host Responses for Flu Sun et al. by avian IAV, such as H5N1, than by human IAV (van Riel et al., 2007). Therefore, protection of bronchial submucosal gland will be an eﬀective approach to prevent infection of highly pathogenic avian IAV. well as increased pro-inﬂammatory responses, which not only demonstrated that RIG-I played a critical role in host cellular responses to human IAV, but also supported that cooperation of host cellular antiviral responses and inﬂammatory responses was crucial for the pathogenesis of IAV. Although H5N1 and H7N9 blocked activation of RIG-I, they actually activated expression of TLR7, which is an important membrane-bound receptor triggered by single-stranded RNA and implicated in response to inﬂuenza virus. Wei et al. (2013) revealed that the TLR7 was involved in the early stage of antiviral innate immune responses in geese during infection of highly pathogenic H5N1. Thus, the severity in humans caused by H5N1 and H7N9 was very likely attributed to activation of the TLR7 pathway, which was normal in birds but not in humans. Despite diﬀerent MOIs with 3 for H1N1 and 1 for H5N1 and H7N9, it was believed that diﬀerences of host cellular antiviral responses and cytokines production between human and avian IAV indeed arose from the nature of virus itself. The reasons were as follows: (1) The canonical PRR RIG-I was activated by H1N1 and repressed by H5N1 and H7N9, and the PRR TLR7 potentially recognizing H5N1 and H7N9 was repressed by H1N1. (2) For the same MOI, the H5N1-NS1-trunc124 mutant can activate the expression of RIG-I and inhibit the expression of TLR7. (3) Disruption of early host cellular antiviral responses during avian IAV infections was likely caused by dysregulation of STAT2, which can be rescued by the H5N1-NS1-trunc1241 mutant. DISCUSSION (3) The diﬀerences of controlling host cellular early response antiviral genes between human and avian IAV were from not only gene expression levels (Figure 3A), but also gene co-expression (Figure 3B) that represented gene synchronization and was robust to various MOIs. (4) While the H5N1-PB2-K627E mutant with limited replication eﬃciency in mammalian cells induced normal antiviral gene expression trends but impaired gene co-expression patterns (Figures 7A,B), the H5N1-NS1- trunc124 mutant with decreased eﬃciency of antagonizing IFN-β production led to normal antiviral gene expression trends and gene co-expression patterns (Figures 7A,B). (5) Regulation of cytokines production was likely through completely diﬀerent pathways during human and avian IAV infections (Figure 5). (6) Importantly, when infecting in vitro cells, H5N1 and H7N9 viruses caused infection 3–6 times faster than H1N1 virus (Simon et al., 2016). Hence, high MOI for H1N1 and low MOI for H5N1 and H7N9 were fair on assessing host cellular responses in vitro. All of these evidences demonstrated that speciﬁc host cellular antiviral immune responses to avian IAV likely resulted from inherent properties of virus. Consistent with IAV classiﬁcation by HA and NA groups (Nobusawa et al., 1991), diﬀerential host cellular antiviral responses revealed that H5N1 was more similar to H1N1 than to H7N9 (Figure 3A). Complementary to HA imprinting accounting for age biases of observed human cases between H5N1 and H7N9 (Gostic et al., 2016), our results also showed that expression patterns of early host cellular antiviral genes could explain these biases. For older adults favored by H7N9, their decreased immunity (Lee et al., 2017) cannot resist the early silent but late-high antiviral responses in host cells (Figure 2). However, decreased immune responses in older adults may protect them from H5N1 infection, which causes cytokine storm (Li et al., 2018) in young adults with the help of their strong immunity. g y Unlike well-adapted human IAV that causes massive morbidity every year, avian IAV suddenly infect human with increased pathogenicity. In contrast to H5N1, H7N9 exhibits strange patterns of host cellular antiviral responses and cytokines production. Actually, our data supported that these likely arose from the nature of H7N9 virus because H7N9 not only suppressed the early upregulated ASGs but also activated early downregulated ASGs in the very early stage. DISCUSSION As the normal cellular antiviral state established by IFN-α treatment (Figure 2D) indicated that high expression of early upregulated ASGs together with low expression of early downregulated ASGs were helpful for host cellular defense against virus, H7N9 likely had evolved to inhibit early host cellular antiviral responses for its replication. Since H7N9 was identiﬁed in March 2013, it has caused ﬁve epidemic waves in China. Due to its evolutionary genotypes (Ding et al., 2017), H7N9 may induce wave-speciﬁc regulation of host cellular antiviral responses. So, the identiﬁed 44 early upregulated ASGs may have potentials to evaluate evolution of H7N9 for monitoring host adaptation. For inﬂuenza virus, survival in host cell has been just like a battle of ﬁghting for limited resources, in which virus uses the fewer to defeat the many. Our results implied possible survival strategies of avian IAV that modulated early host cellular antiviral responses. Actually, expression dynamics of most late-response ASGs were caused by early response ASGs in interferon treated cells (Figures 2A,B and Supplementary Table S2). Hence, for viruses, attacking key early response genes seems an eﬀective approach to dominate host cells. Although the used human IAV H1N1 (A/California/04/2009), a major cause of seasonal inﬂuenza nowadays, once caused a pandemic, its pathogenicity is still much lower than avian IAV H5N1 and H7N9 (Morrison et al., 2014). Nevertheless, our results could not clearly explain the pathogenicity of the pandemic H1N1. Frontiers in Microbiology \| www.frontiersin.org REFERENCES Gostic, K. M., Ambrose, M., Worobey, M., and Lloyd-Smith, J. O. (2016). Potent protection against H5N1 and H7N9 inﬂuenza via childhood hemagglutinin imprinting. Science 354, 722–726. doi: 10.1126/science.aag1322 Auewarakul, P., Suptawiwat, O., Kongchanagul, A., Sangma, C., Suzuki, Y., U h k K l (2007) A i i ﬂ H5N1 i h bi d Auewarakul, P., Suptawiwat, O., Kongchanagul, A., Sangma, C., Suzuki, Y., Ungchusak, K., et al. (2007). An avian inﬂuenza H5N1 virus that binds to a human-type receptor. J. Virol. 81, 9950–9955. doi: 10.1128/JVI.00468-467 Hale, B. G., Randall, R. E., Ortin, J., and Jackson, D. (2008). The multifunctional NS1 protein of inﬂuenza A viruses. J. Gen. Virol. 89(Pt 10), 2359–2376. doi: 10.1099/vir.0.2008/004606-4600 Ungchusak, K., et al. (2007). An avian inﬂuenza H5N1 virus that binds to a human-type receptor. J. Virol. 81, 9950–9955. doi: 10.1128/JVI.00468-467 Belser, J. A., and Tumpey, T. M. (2013). H5N1 pathogenesis studies in mammalian models. Virus Res. 178, 168–185. doi: 10.1016/j.virusres.2013.02.003 Jorgensen, S. E., Christiansen, M., Ryo, L. B., Gad, H. H., Gjedsted, J., Staeheli, P., et al. (2018). Defective RNA sensing by RIG-I in severe inﬂuenza virus infection. Jorgensen, S. E., Christiansen, M., Ryo, L. B., Gad, H. H., Gjedsted, J., Staeheli, P., et al. (2018). Defective RNA sensing by RIG-I in severe inﬂuenza virus infection. Clin. Exp. Immunol. 192, 366–376. doi: 10.1111/cei.13120 Carrasco Pro, S., Dafonte Imedio, A., Santoso, C. S., Gan, K. A., Sewell, J. A., Martinez, M., et al. (2018). Global landscape of mouse and human cytokine transcriptional regulation. Nucleic Acids Res. 46, 9321–9337. doi: 10.1093/nar/ gky787 Clin. Exp. Immunol. 192, 366–376. doi: 10.1111/cei.13120 Josset, L., Zeng, H., Kelly, S. M., Tumpey, T. M., and Katze, M. G. (2014). Transcriptomic characterization of the novel avian-origin inﬂuenza A (H7N9) virus: speciﬁc host response and responses intermediate between avian (H5N1 and H7N7) and human (H3N2) viruses and implications for treatment options. mBio 5, e01102–01113 Chasman, D., Walters, K. B., Lopes, T. J., Eisfeld, A. J., Kawaoka, Y., and Roy, S. (2016). Integrating transcriptomic and proteomic data using predictive regulatory network models of host response to pathogens. PLoS Comput. Biol. 12:e1005013. doi: 10.1371/journal.pcbi.1005013 Kawai, T., and Akira, S. (2007). Signaling to NF-kappaB by Toll-like receptors. Trends Mol. Med. 13, 460–469. doi: 10.1016/j.molmed.2007.09.002 Creanga, A., Hang, N. L. K., Cuong, V. D., Nguyen, H. T., Phuong, H. V. M., Thanh, L. T., et al. (2017). REFERENCES Highly pathogenic avian inﬂuenza A(H5N1) viruses at the animal–human interface in vietnam, 2003–2010. J. Infect. Dis. 216(suppl_4), S529–S538. doi: 10.1093/infdis/jix003 Lee, D. H., Kim, M., Kim, M., Lee, Y. J., Yoo, H. J., Lee, S. H., et al. (2017). Age-dependent alterations in serum cytokines, peripheral blood mononuclear cell cytokine production, natural killer cell activity, and prostaglandin F2alpha. Immunol. Res. 65, 1009–1016. doi: 10.1007/s12026-017-8940-8940 Czudai-Matwich, V., Otte, A., Matrosovich, M., Gabriel, G., and Klenk, H. D. (2014). PB2 mutations D701N and S714R promote adaptation of an inﬂuenza H5N1 virus to a mammalian host. J. Virol. 88, 8735–8742. doi: 10.1128/JVI. 00422-414 Levy, D. E., and Garcia-Sastre, A. (2001). The virus battles: IFN induction of the antiviral state and mechanisms of viral evasion. Cytokine Growth. Factor. Rev. 12, 143–156. doi: 10.1016/s1359-6101(00)00027-7 Dankar, S. K., Wang, S., Ping, J., Forbes, N. E., Keleta, L., Li, Y., et al. (2011). Inﬂuenza a virus NS1 gene mutations F103L and M106I increase replication and virulence. Virol. J. 8:13. doi: 10.1186/1743-422X-8-13 Li, C., Bankhead, A. III, Eisfeld, A. J., Hatta, Y., Jeng, S., Chang, J. H., et al. (2011). Host regulatory network response to infection with highly pathogenic H5N1 avian inﬂuenza virus. J. Virol. 85, 10955–10967. doi: 10.1128/JVI.05792- 5711 Ding, X., Luo, J., Quan, L., Wu, A., and Jiang, T. (2017). Evolutionary genotypes of inﬂuenza A (H7N9) viruses over ﬁve epidemic waves in China. Infect. Genet. Evol. 55, 269–276. doi: 10.1016/j.meegid.2017.09.027 Li, X., Fu, Z., Liang, H., Wang, Y., Qi, X., Ding, M., et al. (2018). H5N1 inﬂuenza virus-speciﬁc miRNA-like small RNA increases cytokine production and mouse mortality via targeting poly(rC)-binding protein 2. Cell Res. 28, 157–171. doi: 10.1038/cr.2018.3 Du, X., King, A. A., Woods, R. J., and Pascual, M. (2017). Evolution-informed forecasting of seasonal inﬂuenza A (H3N2). Sci. Transl. Med. 9:eaan5325. doi: 10.1126/scitranslmed.aan5325 Li, Z., Jiang, Y., Jiao, P., Wang, A., Zhao, F., Tian, G., et al. (2006). The NS1 gene contributes to the virulence of H5N1 avian inﬂuenza viruses. J. Virol. 80, 11115–11123. doi: 10.1128/JVI.00993-996 Gao, H. N., Lu, H. Z., Cao, B., Du, B., Shang, H., Gan, J. H., et al. (2013). Clinical ﬁndings in 111 cases of inﬂuenza A (H7N9) virus infection. N. Engl. J. Med. 368, 2277–2285. doi: 10.1056/NEJMoa1305584 Liu, G., Park, H. S., Pyo, H. M., Liu, Q., and Zhou, Y. (2015). Inﬂuenza A virus panhandle structure is directly involved in RIG-I activation and Interferon Induction. J. Virol. 89, 6067–6079. DATA AVAILABILITY Foundation of China (31601082, 81501351, and 31671371), the Jiangsu Natural Science Foundation (BK20160377), the CAMS Initiative for Innovative Medicine (2016-I2M-1-005), and Central Public-Interest Scientiﬁc Institution Basal Research Fund (2017PT31026 and 2018PT31016). Foundation of China (31601082, 81501351, and 31671371), the Jiangsu Natural Science Foundation (BK20160377), the CAMS Initiative for Innovative Medicine (2016-I2M-1-005), and Central Public-Interest Scientiﬁc Institution Basal Research Fund (2017PT31026 and 2018PT31016). Publicly available datasets were analyzed in this study. This data can be found here: https://www.ncbi.nlm.nih.gov/geo/. CONCLUSION Generally, the RIG-I signaling cascade was used to trigger host cellular innate immunity against IAV infection (Liu et al., 2015). However, our results revealed big diﬀerences between human and avian IAV in triggering PRRs, where RIG-I was activated by H1N1 but suppressed by H5N1 and H7N9. A recent study by Jorgensen et al. (2018) reported that a serve inﬂuenza patient with defective RIG-I exhibited decreased antiviral responses as In conclusion, our study provides a new perspective of the pathogenicity of highly pathogenic avian IAV that results from dysregulation of early host cellular antiviral responses and cytokines production, which will be helpful for prevention of avian IAV infection in the future. September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 13 Early Host Responses for Flu Sun et al. FUNDING The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2019.02007/full#supplementary-material This work was supported by the National Program on Key Basic R h P j t (2015CB910501) th N ti l N t l S i This work was supported by the National Program on Key Basic Research Project (2015CB910501), the National Natural Science ACKNOWLEDGMENTS JS, JW, AW, GC, and TJ conceived and designed the experiments. XY contributed to the data collection. JS and AW contributed to the network construction. JS, JW, XW, and TS built the pathogenesis model. JS analyzed the data. JS and TJ wrote the manuscript. TJ, GC, and AW reviewed the manuscript. We thank the labs for publishing their invaluable transcriptome data that were used in this study. REFERENCES doi: 10.1016/j.molcel.2014.01.027 van Riel, D., Munster, V. J., de Wit, E., Rimmelzwaan, G. F., Fouchier, R. A., Osterhaus, A. D., et al. (2007). Human and avian inﬂuenza viruses target diﬀerent cells in the lower respiratory tract of humans and other mammals. Am. Narang, V., Ramli, M. A., Singhal, A., Kumar, P., de Libero, G., Poidinger, M., et al. (2015). Automated identiﬁcation of core regulatory genes in human gene regulatory networks. PLoS Comput. Biol. 11:e1004504. doi: 10.1371/journal. pcbi.1004504 Osterhaus, A. D., et al. (2007). Human and avian inﬂuenza viruses targe diﬀerent cells in the lower respiratory tract of humans and other mammals. Am J. Pathol. 171, 1215–1223. doi: 10.2353/ajpath.2007.070248 J. Pathol. 171, 1215–1223. doi: 10.2353/ajpath.2007.070248 Vogel, A. J., Harris, S., Marsteller, N., Condon, S. A., and Brown, D. M. (2014). Early cytokine dysregulation and viral replication are associated with mortality during lethal inﬂuenza infection. Viral. Immunol. 27, 214–224. doi: 10.1089/ vim.2013.0095 Nobusawa, E., Aoyama, T., Kato, H., Suzuki, Y., Tateno, Y., and Nakajima, K. (1991). Comparison of complete amino acid sequences and receptor-binding properties among 13 serotypes of hemagglutinins of inﬂuenza A viruses. Virology 182, 475–485. doi: 10.1016/0042-6822(91)90588-3 Wei, L., Jiao, P., Yuan, R., Song, Y., Cui, P., Guo, X., et al. (2013). Goose toll- like receptor 7 (TLR7), myeloid diﬀerentiation factor 88 (MyD88) and antiviral molecules involved in anti-H5N1 highly pathogenic avian inﬂuenza virus response. Vet. Immunol. Immunopathol. 153, 99–106. doi: 10.1016/j.vetimm. 2013.02.012 Omranian, N., Eloundou-Mbebi, J. M., Mueller-Roeber, B., and Nikoloski, Z. (2016). Gene regulatory network inference using fused LASSO on multiple data sets. Sci. Rep. 6:20533. Peng, Y., Li, X., Zhou, H., Wu, A., Dong, L., Zhang, Y., et al. (2017). Continual antigenic diversiﬁcation in china leads to global antigenic complexity of avian inﬂuenza H5N1 viruses. Sci. Rep. 7:43566. Wu, A., Su, C., Wang, D., Peng, Y., Liu, M., Hua, S., et al. (2013). Sequential reassortments underlie diverse inﬂuenza H7N9 genotypes in China. Cell Host Microbe 14, 446–452. doi: 10.1016/j.chom.2013.09.001 Qian, W., Wei, X., Guo, K., Li, Y., Lin, X., Zou, Z., et al. (2017). The C-terminal eﬀector domain of non-structural protein 1 of inﬂuenza A virus blocks IFN- beta production by targeting TNF receptor-associated factor 3. Front. Immunol. 8:779. doi: 10.3389/ﬁmmu.2017.00779 Yu, H., Gao, Z., Feng, Z., Shu, Y., Xiang, N., Zhou, L., et al. (2008). Clinical characteristics of 26 human cases of highly pathogenic avian inﬂuenza A (H5N1) virus infection in China. PLoS One 3:e2985. REFERENCES doi: 10.1128/JVI.00232-15 Gerlach, R. L., Camp, J. V., Chu, Y. K., and Jonsson, C. B. (2013). Early host responses of seasonal and pandemic inﬂuenza A viruses in primary well- diﬀerentiated human lung epithelial cells. PLoS One 8:e78912. doi: 10.1371/ journal.pone.0078912 Liu, Q., Zhou, Y. H., and Yang, Z. Q. (2016). The cytokine storm of severe inﬂuenza and development of immunomodulatory therapy. Cell Mol. Immunol. 13, 3–10. doi: 10.1038/cmi.2015.74 Gibbert, K., Schlaak, J. F., Yang, D., and Dittmer, U. (2013). IFN-alpha subtypes: distinct biological activities in anti-viral therapy. Br. J. Pharmacol. 168, 1048–1058. doi: 10.1111/bph.12010 Loo, Y. M., and Gale, M. (2011). Immune signaling by RIG-I-like receptors. Immunity 34, 680–692. doi: 10.1016/j.immuni.2011.05.003 September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 14 Sun et al. Early Host Responses for Flu McNab, F., Mayer-Barber, K., Sher, A., Wack, A., and O’Garra, A. (2015). Type I interferons in infectious disease. Nat. Rev. Immunol. 15, 87–103. doi: 10.1038/ nri3787 Su, S., Gu, M., Liu, D., Cui, J., Gao, G. F., Zhou, J., et al. (2017). Epidemiology, evolution, and pathogenesis of H7N9 inﬂuenza viruses in ﬁve epidemic waves since 2013 in China. Trends Microbiol. 25, 713–728. doi: 10.1016/j.tim.2017.06. 008 Min, J. Y., Santos, C., Fitch, A., Twaddle, A., Toyoda, Y., DePasse, J. V., et al. (2013). Mammalian adaptation in the PB2 gene of avian H5N1 inﬂuenza virus. J. Virol. 87, 10884–10888. doi: 10.1128/JVI.01016-1013 Subbarao, E. K., London, W., and Murphy, B. R. (1993). A single amino acid in the PB2 gene of inﬂuenza A virus is a determinant of host range. J. Virol. 67, 1761–1764. Moll, H. P., Maier, T., Zommer, A., Lavoie, T., and Brostjan, C. (2011). The diﬀerential activity of interferon-alpha subtypes is consistent among distinct target genes and cell types. Cytokine 53, 52–59. doi: 10.1016/j.cyto.2010.09.006 Szubin, R., Chang, W. L., Greasby, T., Beckett, L., and Baumgarth, N. (2008). Rigid interferon-alpha subtype responses of human plasmacytoid dendritic cells. J. Interferon Cytokine Res. 28, 749–763. doi: 10.1089/jir.2008. 0037 Morrison, J., Josset, L., Tchitchek, N., Chang, J., Belser, J. A., Swayne, D. E., et al. (2014). H7N9 and other pathogenic avian inﬂuenza viruses elicit a three- pronged transcriptomic signature that is reminiscent of 1918 inﬂuenza virus and is associated with lethal outcome in mice. J. Virol. 88, 10556–10568. doi: 10.1128/JVI.00570-514 Upton, J. W., and Chan, F. K. (2014). Staying alive: cell death in antiviral immunity. Mol. Cell. 54, 273–280. REFERENCES doi: 10.1371/journal.pone. 0002985 Ritchie, M. E., Phipson, B., Wu, D., Hu, Y., Law, C. W., Shi, W., et al. (2015). limma powers diﬀerential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43:e47. doi: 10.1093/nar/gkv007 Zhou, H., Zhu, J., Tu, J., Zou, W., Hu, Y., Yu, Z., et al. (2010). Eﬀect on virulence and pathogenicity of H5N1 inﬂuenza A virus through truncations of NS1 eIF4GI binding domain. J. Infect. Dis. 202, 1338–1346. doi: 10.1086/ 656536 Schneider, W. M., Chevillotte, M. D., and Rice, C. M. (2014). Interferon-stimulated genes: a complex web of host defenses. Annu. Rev. Immunol. 32, 513–545. doi: 10.1146/annurev-immunol-032713-120231 Zhu, Y., Chidekel, A., and Shaﬀer, T. H. (2010). Cultured human airway epithelial cells (calu-3): a model of human respiratory function, structure, and inﬂammatory responses. Crit. Care Res. Pract. 2010:394578. doi: 10.1155/2010/ 394578 Schoenemeyer, A., Barnes, B. J., Mancl, M. E., Latz, E., Goutagny, N., Pitha, P. M., et al. (2005). The interferon regulatory factor, IRF5, is a central mediator of toll-like receptor 7 signaling. J. Biol. Chem. 280, 17005–17012. doi: 10.1074/jbc. M412584200 Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Simon, P. F., de La Vega, M. A., Paradis, E., Mendoza, E., Coombs, K. M., Kobasa, D., et al. (2016). Avian inﬂuenza viruses that cause highly virulent infections in humans exhibit distinct replicative properties in contrast to human H1N1 viruses. Sci. Rep. 6:24154. doi: 10.1038/srep24154 Copyright © 2019 Sun, Wang, Yuan, Wu, Sui, Wu, Cheng and Jiang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Copyright © 2019 Sun, Wang, Yuan, Wu, Sui, Wu, Cheng and Jiang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). REFERENCES The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Copyright © 2019 Sun, Wang, Yuan, Wu, Sui, Wu, Cheng and Jiang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Simon, P. F., McCorrister, S., Hu, P., Chong, P., Silaghi, A., Westmacott, G., et al. (2015). Highly pathogenic H5N1 and novel H7N9 Inﬂuenza A viruses induce more profound proteomic host responses than seasonal and pandemic H1N1 strains. J. Proteome Res. 14, 4511–4523. doi: 10.1021/acs.jproteome.5b00196 Proteome Res. 14, 4511–4523. doi: 10.1021/acs.jproteome.5b00196 Su, S., Bi, Y., Wong, G., Gray, G. C., Gao, G. F., and Li, S. (2015). Epidemiology, evolution, and recent outbreaks of avian inﬂuenza virus in china. J. Virol. 89, 8671–8676. doi: 10.1128/JVI.01034-1015 September 2019 \| Volume 10 \| Article 2007 Frontiers in Microbiology \| www.frontiersin.org 15
https://openalex.org/W4312053268	https://www.repository.cam.ac.uk/bitstreams/440ce3a2-60ed-4cea-af8c-ffae8ed5d1ba/download	English	null	Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior	Nature medicine	2,022	cc-by	26,423	Article https://doi.org/10.1038/s41591-022-02106-5 1USDA/ARS Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA. 2University of Cambridge Metabolic Research Laboratories and NIHR Cambridge Biomedical Research Centre, Wellcome-MRC Institute of Metabolic Science, Addenbrooke’s Hospital, Cambridge, UK. 3Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX, USA. 4Division of Diabetes, Endocrinology and Metabolism, Department of Medicine, Baylor College of Medicine, Houston, TX, USA. 5Exeter Centre of Excellence for Diabetes Research (EXCEED), University of Exeter Medical School, Exeter, UK. 6These authors contributed equally: Yang He, Bas Brouwers, Hesong Liu. e-mail: yongx@bcm.edu; isf20@cam.ac.uk nature medicine nature medicine Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior 0 1 2 3 4 5 6 0 5 10 15 20 25 30 35 40 Age of females (years) Weight (kg) HTR2C A171V mutation carrier b NH2- 1 9 21 31 -COOH Cell membrane 41 51 59 73 84 93 108 118 132 146 158 166 174 187 199 210 220 235 250 260 270 280 290 299 314 328 338 351 364 374 385 395 405 415 425 435 439 458 449 K K L H ICL1 Y V W P L ECL1 P I E H S R F N ICL2 E E K V F V N N T T C V L N ECL2 G H T E E P P G L S L D F L K C C K R N T A E E E N S A N P N Q D Q N A R R R K K K E R R P R G T ICL3 E K S C ECL3 M V N L R N A V H S F L V H L I G L L V W Q C D I S V S P V A A I V T D I F N T S D G G R F K F P D G V Q N-term N Y K V E K K P P V R Q I P R V A A T A L S G R E L N V N I Y R H T N E P V I E K A S D N E P G I E M Q V E N L E L P V N P S S V V S E R I S S V C-term N W P A L S I V I I I I M T I G G N I L V I M A V S M E N A T N Y F L M S L A I A D M L V G L L V M P L S L L A I L Y D P R Y L C P V W I S L D V L F S T A S I M H L C A I S L D R Y V A I R N S R T K A I M K I A I V W A I S I G V S V P I P V I G L R D D P N F V L I G S F V A F F I P L T I M V I T Y C L T I Y V L R R Q A L M L L H M Q A I N N E R K A S K V L G I V F F V F L I M W C P F F I T N I L S V L C N Q K L M E K L L N V F V W I G Y V C S G I N P L V Y T L F N K I F S N Y R R A Y L R C a b a Age of females (years) 0 1 2 3 4 5 6 0 5 10 15 20 25 30 35 40 Age of females (years) Weight (kg) HTR2C C266R mutation carrier Age of females (years) Age of females (years) Fig. Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior Yang He1,6, Bas Brouwers 2,6, Hesong Liu1,6, Hailan Liu1, Katherine Lawler2, Edson Mendes de Oliveira 2, Dong-Kee Lee3, Yongjie Yang1, Aaron R. Cox 4, Julia M. Keogh2, Elana Henning2, Rebecca Bounds2, Aliki Perdikari2, Vikram Ayinampudi2, Chunmei Wang1, Meng Yu1, Longlong Tu1, Nan Zhang1, Na Yin1, Junying Han1, Nikolas A. Scarcelli1, Zili Yan1, Kristine M. Conde 1, Camille Potts1, Jonathan C. Bean 1, Mengjie Wang1, Sean M. Hartig 3,4, Lan Liao3, Jianming Xu 3, Inês Barroso 5, Jacek Mokrosinski2, Yong Xu 1,3 & I. Sadaf Farooqi 2 Serotonin reuptake inhibitors and receptor agonists are used to treat obesity, anxiety and depression. Here we studied the role of the serotonin 2C receptor (5-HT2CR) in weight regulation and behavior. Using exome sequencing of 2,548 people with severe obesity and 1,117 control individuals without obesity, we identified 13 rare variants in the gene encoding 5-HT2CR (HTR2C) in 19 unrelated people (3 males and 16 females). Eleven variants caused a loss of function in HEK293 cells. All people who carried variants had hyperphagia and some degree of maladaptive behavior. Knock-in male mice harboring a human loss-of-function HTR2C variant developed obesity and reduced social exploratory behavior; female mice heterozygous for the same variant showed similar deficits with reduced severity. Using the 5-HT2CR agonist lorcaserin, we found that depolarization of appetite-suppressing proopiomelanocortin neurons was impaired in knock-in mice. In conclusion, we demonstrate that 5-HT2CR is involved in the regulation of human appetite, weight and behavior. Our findings suggest that melanocortin receptor agonists might be effective in treating severe obesity in individuals carrying HTR2C variants. We suggest that HTR2C should be included in diagnostic gene panels for severe childhood-onset obesity. Drugs that alter levels of the neurotransmitter serotonin (5-hydroxytryptamine, 5-HT) are widely prescribed for the treatment of obesity and neuropsychiatric disorders; however, they often exert adverse effects due to a lack of receptor specificity1 as 5-HT signals through at least 14 different receptors to regulate body weight, mood and behavior2,3. For example, second-generation antipsychotic drugs (clozapine and olanzapine) are highly effective at reducing psychotic symptoms, but cause increased hunger and weight gain in up to 60% through at least 14 different receptors to regulate body weight, mood and behavior2,3. Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior For example, second-generation antipsychotic drugs (clozapine and olanzapine) are highly effective at reducing psychotic symptoms, but cause increased hunger and weight gain in up to 60% Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2537 https://doi.org/10.1038/s41591-022-02106-5 Article NH2- 1 9 21 31 -COOH Cell membrane 41 51 59 73 84 93 108 118 132 146 158 166 174 187 199 210 220 235 250 260 270 280 290 299 314 328 338 351 364 374 385 395 405 415 425 435 439 458 449 K K L H ICL1 Y V W P L ECL1 P I E H S R F N ICL2 E E K V F V N N T T C V L N ECL2 G H T E E P P G L S L D F L K C C K R N T A E E E N S A N P N Q D Q N A R R R K K K E R R P R G T ICL3 E K S C ECL3 M V N L R N A V H S F L V H L I G L L V W Q C D I S V S P V A A I V T D I F N T S D G G R F K F P D G V Q N-term N Y K V E K K P P V R Q I P R V A A T A L S G R E L N V N I Y R H T N E P V I E K A S D N E P G I E M Q V E N L E L P V N P S S V V S E R I S S V C-term N W P A L S I V I I I I M T I G G N I L V I M A V S M E N A T N Y F L M S L A I A D M L V G L L V M P L S L L A I L Y D P R Y L C P V W I S L D V L F S T A S I M H L C A I S L D R Y V A I R N S R T K A I M K I A I V W A I S I G V S V P I P V I G L R D D P N F V L I G S F V A F F I P L T I M V I T Y C L T I Y V L R R Q A L M L L H M Q A I N N E R K A S K V L G I V F F V F L I M W C P F F I T N I L S V L C N Q K L M E K L L N V F V W I G Y V C S G I N P L V Y T L F N K I F S N Y R R A Y L R C 0 1 2 3 4 5 6 0 5 10 15 20 25 30 35 40 Age of females (years) Weight (kg) 0 1 2 3 4 5 6 0 5 10 15 20 25 30 35 40 Age of females (years) Weight (kg) HTR2C A171V mutation carrier HTR2C C266R mutation carrier a b Fig. Clinical phenotype of carriers of rare variants in HTR2C All probands had a history of hyperphagia, impaired satiety and weight gain from early childhood (Fig. 1b) leading to severe obesity (mean BMI = 36.5 kg m−2 in adults; mean BMI SDS = 3.5 in children) (Table 1). Eight of 19 probands had learning difficulties or developmental delay; longitudinal follow-up of patients revealed a wide spectrum of maladap- tive behaviors that started in childhood, including emotional lability (frequent outbursts of crying and/or aggressive behavior in the absence of obvious triggers) and maladaptive behavior, particularly in social settings (Table 1). The clinical features seen in people carrying HTR2C variants show some overlap with other conditions associated with onset of severe obesity in childhood (Fig. 2). For example, reactive aggression is seen in people with deletions and loss-of-function (LOF) mutations in the adaptor molecule, SH2B1, which modulates signaling by leptin, insulin and brain-derived neurotrophic factor (BDNF)14,15. In our study, people carrying HTR2C variants generally reported features of anxiety and/ or social anxiety from childhood. To ascertain the prevalence of these conditions in a large clinically ascertained cohort, we reviewed the records of 7,775 children recruited to the Genetics of Obesity Study (GOOS; www.goos.org.uk). Criteria for inclusion in this study were severe early-onset obesity (defined as BMI SDS > 3, onset of obesity before 10 years of age). Learning difficulties (n = 650, 8.3%), speech and Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior 1 \| Rare variants affecting 5-HT2CR identified in people with severe obesity. a, Rare variants identified in individuals with severe early-onset obesity shown on a schematic of the 5-HT2CR protein; ECL and ICL refer to extra- and intracellular loops of the G-protein-coupled receptor (GPCR), respectively; C-term, C-terminal domain of the protein. b, Weight charts of two female probands (5th and 95th percentiles based on reference data for the UK population shown as dashed lines). of patients, which represents a major barrier to their long-term use4. Understanding the mechanisms by which serotonin’s effects on food intake, body weight, mood and behavior are mediated in humans could inform the development of more targeted therapies for a range of clini- cal disorders. lies on the X chromosome in humans; 16 girls carried a heterozygous variant and 3 boys carried a hemizygous variant (variant on their only X chromosome). Three families of probands carrying variants that were not found in controls, consented to co-segregation studies. The four people carrying rare HTR2C variants in these families had overweight or obesity (Table 1). Studies in mice have shown that the appetite-suppressing actions of 5-HT are largely mediated by 5-HT2CRs expressed on hypothalamic proopiomelanocortin (POMC) neurons5–7, which play a major role in weight regulation. A complete lack of POMC due to bi-allelic loss-of-function mutations causes hyperphagia and severe childhood-onset obesity8. This genetic obesity syndrome and other closely related disorders of the leptin–melanocortin pathway are treatable by setmelanotide, a melanocortin 4 receptor (MC4R) agonist, which has been licensed for clinical use in the UK, Europe and the US9–11. Here, we set out to investigate the potential contribution of 5-HT2CR signaling to human weight regulation and the interaction between 5-HT2CRs and the melanocortin pathway. Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior 1 \| Rare variants affecting 5-HT2CR identified in people with severe obesity. a, Rare variants identified in individuals with severe early-onset obesity shown on a schematic of the 5-HT2CR protein; ECL and ICL refer to extra- and intracellular loops of the G-protein-coupled receptor (GPCR), respectively; C-term, C-terminal domain of the protein. b, Weight charts of two female probands (5th and 95th percentiles based on reference data for the UK population shown as dashed lines). https://doi.org/10.1038/s41591-022-02106-5 https://doi.org/10.1038/s41591-022-02106-5 Article Table 1 \| Phenotypes seen in people carrying rare variants in HTR2C Variant Age Sex Weight Height BMI Measured BMR Predicted BMR RQ Glucose Insulin SBP/DBP Leptin TSH FT4 Medical history (years) kg cm (SDS) kg m−2 (SDS) MJ d−1 MJ d−1 mmol l−1 pmol l−1 mm Hg nmol l−1 IU l−1 IU l−1 V2L* 27.3 F 161.2 154.6 67.4 9.3 10.5 0.9 4.7 123/59 2.9 13.2 Exomphthalos, mild learning difficulties, sleep apnea N6K* 5.6 F 40.1 119.2 (1.4) 28.2 (4.3) 4.4 41 31.1 A7V* 19.3 F 107.6 162.5 40.8 7.5 7.9 0.9 4.8 130 116/70 3.8 Dyslexia, learning difficulties, emotional lability, volatile behavior, delayed puberty A7V (m) 44 F 75.5 165.1 27.7 A7V* 13.4 F 101 154.0 (−0.5) 42.6 (3.9) V61I* 13.5 F 102 158.3 (0.1) 40.7 (3.8) Normal - ADHD, social anxiety, depression, ODD, borderline personality disorder V61I* 19 F 114.4 155.3 47.4 8.5 7.4 0.7 4.6 212 105/59 1.4 12.1 Dyslexia, learning difficulties, Table 1 \| Phenotypes seen in people carrying rare variants in HTR2C Variant Age Sex Weight Height BMI Measured BMR Predicted BMR RQ Glucose Insulin SBP/DBP Leptin TSH FT4 Medical history (years) kg cm (SDS) kg m−2 (SDS) MJ d−1 MJ d−1 mmol l−1 pmol l−1 mm Hg nmol l−1 IU l−1 IU l−1 V2L* 27.3 F 161.2 154.6 67.4 9.3 10.5 0.9 4.7 123/59 2.9 13.2 Exomphthalos, mild learning difficulties, sleep apnea N6K* 5.6 F 40.1 119.2 (1.4) 28.2 (4.3) 4.4 41 31.1 A7V* 19.3 F 107.6 162.5 40.8 7.5 7.9 0.9 4.8 130 116/70 3.8 Dyslexia, learning difficulties, emotional lability, volatile behavior, delayed puberty A7V (m) 44 F 75.5 165.1 27.7 A7V* 13.4 F 101 154.0 (−0.5) 42.6 (3.9) V61I* 13.5 F 102 158.3 (0.1) 40.7 (3.8) Normal - ADHD, social anxiety, depression, ODD, borderline personality disorder V61I* 19 F 114.4 155.3 47.4 8.5 7.4 0.7 4.6 212 105/59 1.4 12.1 Dyslexia, learning difficulties, Table 1 \| Phenotypes seen in people carrying rare variants in HTR2C Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2539 (−0.5) V61I* 13.5 F 102 158.3 (0.1) 40.7 (3.8) Normal - ADHD, social anxiety, depression, ODD, borderline personality disorder V61I* 19 F 114.4 155.3 47.4 8.5 7.4 0.7 4.6 212 105/59 1.4 12.1 Dyslexia, learning difficulties, volatile behavior, anxiety, aggression, depression, emotional lability, impulsivity, mild OCD and social anxiety V61I (m) 43.9 F 89.1 160.0 34.8 Depression, anxiety V61I (s) 25.4 F 102.2 159.0 40.4 7.5 7.7 0.8 4.3 95 134/85 0.7 11.7 Anxiety, asthma I97V* 11 F 82 157.0 (1.9) 33.3 (3.4) IFG 431 Normal Normal A171V* 14.6 M 108.5 172.2 (0.7) 36.6 (3.4) 9.8 9.2 5.4 265 110/63 89 2.2 16.4 A171V (m) 39.6 F 64.1 173.8 21.2 6.4 5.9 0.9 5.8 68 124/83 35.3 0.7 17 A171V (b) 9.9 M 31.2 142.2 (0.7) 15.4 (-0.6) 4.5 4.7 0.9 4.7 94 99/64 19.8 2.2 14.6 A171V(mgf) 71.3 M 86.7 176.1 28 7.6 7.9 0.9 5.1 44 127/73 2.0 15.7 A171V(ma) 43.1 F 76.4 167.6 27.2 A171V* 5.1 F 29.7 110.2 (0.1) 24.5 (3.7) 4.7 202 39.1 Learning difficulties, emotional lability. Rare variants in HTR2C in people with severe obesity l d d d Rare variants in HTR2C in people with severe obesity We analyzed exome sequencing and targeted resequencing on 2,548 (46% male and 54% female) European ancestry individuals with severe, early-onset obesity (mean body mass index (BMI) s.d. score (SDS) > 3; age of onset <10 years) and 1,117 (40% male, 60% female) ancestry-matched controls analyzed using the same methods12,13. We identified 13 rare variants (minor allele frequency <1%) in HTR2C in 19 unrelated people with severe obesity (Fig. 1a); these variants were either not found or very rarely found, in publicly available exomes (Table 1). One rare variant (A171V) was identified in obese individuals (n = 5) and controls (n = 2); no rare variants were found in controls only. https://doi.org/10.1038/s41591-022-02106-5 Facial dysmorphia A171V* 1.2 M 15.2 83.5 (2.0) 21.8 (2.7) Normal - A171V* 2.6 F 20.5 84.0 (−2.1) 29.1 (5.4) Normal - 5.0 16.6 Learning difficulties, severe delay. Slight facial dysmorphia, increased infections A171V* 7 F 34.7 116.5 (−0.9) 25.6 (3.3) 4.5 43 Normal Normal Speech delay, attention deficit disorder, hyperactivity V208M* 3.1 F 16.5 85.7 (−2.6) 22.5 (3.5) - - Normal Normal Learning disability Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2539 anxiety V61I (m) 43.9 F 89.1 160.0 34.8 Depression, anxiety V61I (s) 25.4 F 102.2 159.0 40.4 7.5 7.7 0.8 4.3 95 134/85 0.7 11.7 Anxiety, asthma I97V* 11 F 82 157.0 (1.9) 33.3 (3.4) IFG 431 Normal Normal A171V* 14.6 M 108.5 172.2 (0.7) 36.6 (3.4) 9.8 9.2 5.4 265 110/63 89 2.2 16.4 A171V (m) 39.6 F 64.1 173.8 21.2 6.4 5.9 0.9 5.8 68 124/83 35.3 0.7 17 A171V (b) 9.9 M 31.2 142.2 (0.7) 15.4 (-0.6) 4.5 4.7 0.9 4.7 94 99/64 19.8 2.2 14.6 A171V(mgf) 71.3 M 86.7 176.1 28 7.6 7.9 0.9 5.1 44 127/73 2.0 15.7 A171V(ma) 43.1 F 76.4 167.6 27.2 A171V* 5.1 F 29.7 110.2 (0.1) 24.5 (3.7) 4.7 202 39.1 Learning difficulties, emotional lability. Facial dysmorphia A171V* 1.2 M 15.2 83.5 (2.0) 21.8 (2.7) Normal - A171V* 2.6 F 20.5 84.0 (−2.1) 29.1 (5.4) Normal - 5.0 16.6 Learning difficulties, severe delay. Slight facial dysmorphia, increased infections A171V* 7 F 34.7 116.5 (−0.9) 25.6 (3.3) 4.5 43 Normal Normal Speech delay, attention deficit disorder, hyperactivity V208M* 3.1 F 16.5 85.7 (−2.6) 22.5 (3.5) - - Normal Normal Learning disability Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2539 anxiety V61I (m) 43.9 F 89.1 160.0 34.8 Depression, anxiety V61I (s) 25.4 F 102.2 159.0 40.4 7.5 7.7 0.8 4.3 95 134/85 0.7 11.7 Anxiety, asthma I97V* 11 F 82 157.0 (1.9) 33.3 (3.4) IFG 431 Normal Normal A171V* 14.6 M 108.5 172.2 (0.7) 36.6 (3.4) 9.8 9.2 5.4 265 110/63 89 2.2 16.4 A171V (m) 39.6 F 64.1 173.8 21.2 6.4 5.9 0.9 5.8 68 124/83 35.3 0.7 17 A171V (b) 9.9 M 31.2 142.2 (0.7) 15.4 (-0.6) 4.5 4.7 0.9 4.7 94 99/64 19.8 2.2 14.6 A171V(mgf) 71.3 M 86.7 176.1 28 7.6 7.9 0.9 5.1 44 127/73 2.0 15.7 A171V(ma) 43.1 F 76.4 167.6 27.2 A171V* 5.1 F 29.7 110.2 (0.1) 24.5 (3.7) 4.7 202 39.1 Learning difficulties, emotional lability. Rare variants in HTR2C in people with severe obesity l d d d HTR2C Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2538 Article https://doi.org/10.1038/s41591-022-02106-5 Table 1 \| Phenotypes seen in people carrying rare variants in HTR2C Variant Age Sex Weight Height BMI Measured BMR Predicted BMR RQ Glucose Insulin SBP/DBP Leptin TSH FT4 Medical history (years) kg cm (SDS) kg m−2 (SDS) MJ d−1 MJ d−1 mmol l−1 pmol l−1 mm Hg nmol l−1 IU l−1 IU l−1 V2L* 27.3 F 161.2 154.6 67.4 9.3 10.5 0.9 4.7 123/59 2.9 13.2 Exomphthalos, mild learning difficulties, sleep apnea N6K* 5.6 F 40.1 119.2 (1.4) 28.2 (4.3) 4.4 41 31.1 A7V* 19.3 F 107.6 162.5 40.8 7.5 7.9 0.9 4.8 130 116/70 3.8 Dyslexia, learning difficulties, emotional lability, volatile behavior, delayed puberty A7V (m) 44 F 75.5 165.1 27.7 A7V* 13.4 F 101 154.0 (−0.5) 42.6 (3.9) V61I* 13.5 F 102 158.3 (0.1) 40.7 (3.8) Normal - ADHD, social anxiety, depression, ODD, borderline personality disorder V61I* 19 F 114.4 155.3 47.4 8.5 7.4 0.7 4.6 212 105/59 1.4 12.1 Dyslexia, learning difficulties, volatile behavior, anxiety, aggression, depression, emotional lability, impulsivity, mild OCD and social anxiety V61I (m) 43.9 F 89.1 160.0 34.8 Depression, anxiety V61I (s) 25.4 F 102.2 159.0 40.4 7.5 7.7 0.8 4.3 95 134/85 0.7 11.7 Anxiety, asthma I97V* 11 F 82 157.0 (1.9) 33.3 (3.4) IFG 431 Normal Normal A171V* 14.6 M 108.5 172.2 (0.7) 36.6 (3.4) 9.8 9.2 5.4 265 110/63 89 2.2 16.4 A171V (m) 39.6 F 64.1 173.8 21.2 6.4 5.9 0.9 5.8 68 124/83 35.3 0.7 17 A171V (b) 9.9 M 31.2 142.2 (0.7) 15.4 (-0.6) 4.5 4.7 0.9 4.7 94 99/64 19.8 2.2 14.6 A171V(mgf) 71.3 M 86.7 176.1 28 7.6 7.9 0.9 5.1 44 127/73 2.0 15.7 A171V(ma) 43.1 F 76.4 167.6 27.2 A171V* 5.1 F 29.7 110.2 (0.1) 24.5 (3.7) 4.7 202 39.1 Learning difficulties, emotional lability. Facial dysmorphia A171V* 1 2 M 15 2 83 5 21 8 (2 7) Normal - https://doi.org/10.1038/s41591-022-02106-5 Article Variant Age Sex Weight Height BMI Measured BMR Predicted BMR RQ Glucose Insulin SBP/DBP Leptin TSH FT4 Medical history (years) kg cm (SDS) kg m−2 (SDS) MJ d−1 MJ d−1 mmol l−1 pmol l−1 mm Hg nmol l−1 IU l−1 IU l−1 N213T* 6.4 M 35.9 122.0 (0.8) 24.1 (3.6) 5.2 72 16.7 Normal Normal Epilepsy S260G* 16.1 F 127 154.0 (−1.5) 53.6 (4.4) Type 2 diabetes 559 114 Depression, social anxiety, PCOS C266R* 15.8 F 112 156.4 (−1.1) 45.8 (4.0) 6.8 7.9 0.9 3.8 251 113/68 3.3 14.6 Learning difficulties, delayed speech, autistic traits, social anxiety, depression, aggression, emotional lability, behavioral & concentration difficulties C266R (s) 33.9 F 171.8 170.2 59.3 Aggression C266R (m) 56.1 F 104.6 165.1 38.4 Hypertension Q282H* 15.9 F 82.7 156.0 (−1.2) 34.0 (3.0) 3.9 62 23.1 Depression, anxiety, sudden mood swings and PCOS F327L* 7 F 49.8 126.4 (1.0) 31.2 (4.2) 6.3 103 44 Normal Normal Learning difficulties, volatile behavior and PCOS F327L (f) 30.9 M 88.9 1.78 28.1 Tf419A* 7.1 F 40.8 128.3 (1.3) 24.8 (3.1) Normal Normal Hypothyroidism, social anxiety Data for probands with severe obesity (denoted as ) from the GOOS cohort in whom HTR2C variants were identified. Variants were found in hemizygous form in males and heterozygous form in females (transcript ID, ENST00000276198). Additional information was obtained on their family members who carried variants, which is included directly beneath the proband where available; family relationships are indicated in brackets (m, mother; f, father; ma, maternal aunt; s, sister; b, brother; mgf, maternal grandfather.) Age and sex-adjusted SDS for height and BMI were included for individuals up to 18 years; sex (M, male; F, female); TSH, thyroid stimulating hormone (0.35–5.50 mIU l−1); FT, free thyroxine (10.0–19.8 pmol l−1); fasting plasma insulin (0–60 pmol l−1); fasting glucose (3.5–5.5 mmol l−1). In some cases, biochemical test results were not available but values were recorded as within the normal range in clinical letters (recorded as normal here). Basal metabolic rate (BMR) and respiratory quotient (RQ) were measured by indirect calorimetry in the fasted state and BMR was predicted on the basis of age, sex and body composition shown for comparison. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured in the rested, fasted state in mm Hg. Blank indicates data not available. https://doi.org/10.1038/s41591-022-02106-5 Additional information was obtained on their family members who carried variants, which is included directly beneath the proband where available; family relationships are indicated in brackets (m, mother; f, father; ma, maternal aunt; s, sister; b, brother; mgf, maternal grandfather.) Age and sex-adjusted SDS for height and BMI were included for individuals up to 18 years; sex (M, male; F, female); TSH, thyroid stimulating hormone (0.35–5.50 mIU l−1); FT, free thyroxine (10.0–19.8 pmol l−1); fasting plasma insulin (0–60 pmol l−1); fasting glucose (3.5–5.5 mmol l−1). In some cases, biochemical test results were not available but values were recorded as within the normal range in clinical letters (recorded as normal here). Basal metabolic rate (BMR) and respiratory quotient (RQ) were measured by indirect calorimetry in the fasted state and BMR was predicted on the basis of age, sex and body composition shown for comparison. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured in the rested, fasted state in mm Hg. Blank indicates data not available. ADHD, attention deficit hyperactivity disorder; ODD, oppositional defiant disorder; OCD, obsessive-compulsive disorder; IFG, impaired fasting glycemia; PCOS, polycystic ovarian syndrome. Genetic obesity syndromes Developmental delay, learning and/or behavioral dificulties Normal development Retinal dystrophies Skeletal abnormalities Dysmorphic features Autistic traits Anxiety Aggression Hyperactivity Endocrine features Prader–Willi Ciliopathies (Bardet–Biedl, Alstrom’s, Tub) Albright’s hereditary osteodystrophy (PHP) Cohen syndrome Single minded 1 orthopedia Brain-derived neurotrophic factor Tyrosine receptor kinase B Src homology 2B adaptor protein 1 Pleckstrin domain homology interacting protein Serotonin 2C receptor Leptin Leptin receptor Pro-opiomelanocortin Prohormone convertase 1 Carboxypeptidase E Steroid receptor co-activator-1 Semaphorin 3A-G Plexin A1-4 Neuropilin 1 and 2 Melanocortin 4 receptor Melanocortin receptor accessory protein 2 Adenylate cyclase 3 Kinase suppressor of Ras-2 Fig. 2 \| Genetic obesity syndromes. A schematic depicting genetic obesity syndromes, conditions where severe childhood-onset obesity is a major presenting clinical feature. Conditions in bold are those where the gene/genes have been shown to affect signaling through the leptin–melanocortin pathway and could be treated by an MC4R agonist. Clinical features seen in people with LOF variants in the gene encoding the serotonin 2C receptor (orange) are reported in this study. https://doi.org/10.1038/s41591-022-02106-5 Facial dysmorphia A171V 1.2 M 15.2 83.5 (2.0) 21.8 (2.7) Normal - A171V* 2.6 F 20.5 84.0 (−2.1) 29.1 (5.4) Normal - 5.0 16.6 Learning difficulties, severe delay. Slight facial dysmorphia, increased infections A171V* 7 F 34.7 116.5 (−0.9) 25.6 (3.3) 4.5 43 Normal Normal Speech delay, attention deficit disorder, hyperactivity V208M* 3.1 F 16.5 85.7 (−2.6) 22.5 (3.5) - - Normal Normal Learning disability Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 https://doi.org/10.1038/s41591-022-02106-5 ADHD, attention deficit hyperactivity disorder; ODD, oppositional defiant disorder; OCD, obsessive-compulsive disorder; IFG, impaired fasting glycemia; PCOS, polycystic ovarian syndrome. Table 1 (continued) \| Phenotypes seen in people carrying rare variants in HTR2C Variant Age Sex Weight Height BMI Measured BMR Predicted BMR RQ Glucose Insulin SBP/DBP Leptin TSH FT4 Medical history (years) kg cm (SDS) kg m−2 (SDS) MJ d−1 MJ d−1 mmol l−1 pmol l−1 mm Hg nmol l−1 IU l−1 IU l−1 N213T* 6.4 M 35.9 122.0 (0.8) 24.1 (3.6) 5.2 72 16.7 Normal Normal Epilepsy S260G* 16.1 F 127 154.0 (−1.5) 53.6 (4.4) Type 2 diabetes 559 114 Depression, social anxiety, PCOS C266R* 15.8 F 112 156.4 (−1.1) 45.8 (4.0) 6.8 7.9 0.9 3.8 251 113/68 3.3 14.6 Learning difficulties, delayed speech Table 1 (continued) \| Phenotypes seen in people carrying rare variants in HTR2C Table 1 (continued) \| Phenotypes seen in people carrying rare variants in HTR2C N M di i \| V l 28 \| D b 2022 \| 2537 2546 S260G* 16.1 F 127 154.0 (−1.5) 53.6 (4.4) Type 2 diabetes 559 114 Depression, social anxiety, PCOS C266R* 15.8 F 112 156.4 (−1.1) 45.8 (4.0) 6.8 7.9 0.9 3.8 251 113/68 3.3 14.6 Learning difficulties, delayed speech, autistic traits, social anxiety, depression, aggression, emotional lability, behavioral & concentration difficulties C266R (s) 33.9 F 171.8 170.2 59.3 Aggression C266R (m) 56.1 F 104.6 165.1 38.4 Hypertension Q282H* 15.9 F 82.7 156.0 (−1.2) 34.0 (3.0) 3.9 62 23.1 Depression, anxiety, sudden mood swings and PCOS F327L* 7 F 49.8 126.4 (1.0) 31.2 (4.2) 6.3 103 44 Normal Normal Learning difficulties, volatile behavior and PCOS F327L (f) 30.9 M 88.9 1.78 28.1 Tf419A* 7.1 F 40.8 128.3 (1.3) 24.8 (3.1) Normal Normal Hypothyroidism, social anxiety Data for probands with severe obesity (denoted as ) from the GOOS cohort in whom HTR2C variants were identified. Variants were found in hemizygous form in males and heterozygous form in females (transcript ID, ENST00000276198). https://doi.org/10.1038/s41591-022-02106-5 difficulties C266R (s) 33.9 F 171.8 170.2 59.3 Aggression C266R (m) 56.1 F 104.6 165.1 38.4 Hypertension Q282H 15.9 F 82.7 156.0 (−1.2) 34.0 (3.0) 3.9 62 23.1 Depression, anxiety, sudden mood swings and PCOS F327L* 7 F 49.8 126.4 (1.0) 31.2 (4.2) 6.3 103 44 Normal Normal Learning difficulties, volatile behavior and PCOS F327L (f) 30.9 M 88.9 1.78 28.1 Tf419A* 7.1 F 40.8 128.3 (1.3) 24.8 (3.1) Normal Normal Hypothyroidism, social anxiety Data for probands with severe obesity (denoted as ) from the GOOS cohort in whom HTR2C variants were identified. Variants were found in hemizygous form in males and heterozygous form in females (transcript ID, ENST00000276198). Additional information was obtained on their family members who carried variants, which is included directly beneath the proband where available; family relationships are indicated in brackets (m, mother; f, father; ma, maternal aunt; s, sister; b, brother; mgf, maternal grandfather.) Age and sex-adjusted SDS for height and BMI were included for individuals up to 18 years; sex (M, male; F, female); TSH, thyroid stimulating hormone (0.35–5.50 mIU l−1); FT, free thyroxine (10.0–19.8 pmol l−1); fasting plasma insulin (0–60 pmol l−1); fasting glucose (3.5–5.5 mmol l−1). In some cases, biochemical test results were not available but values were recorded as within the normal range in clinical letters (recorded as normal here). Basal metabolic rate (BMR) and respiratory quotient (RQ) were measured by indirect calorimetry in the fasted state and BMR was predicted on the basis of age, sex and body composition shown for comparison. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured in the rested, fasted state in mm Hg. Blank indicates data not available. ADHD, attention deficit hyperactivity disorder; ODD, oppositional defiant disorder; OCD, obsessive-compulsive disorder; IFG, impaired fasting glycemia; PCOS, polycystic ovarian syndrome. Data for probands with severe obesity (denoted as ) from the GOOS cohort in whom HTR2C variants were identified. Variants were found in hemizygous form in males and heterozygous form in females (transcript ID, ENST00000276198). Normal development Developmental delay, learning and/or behavioral dificulties Normal development Retinal dystrophies Skeletal abnormalities Dysmorphic features Autistic traits Anxiety Aggression Hyperactivity Endocrine features Prader–Willi Ciliopathies (Bardet–Biedl, Alstrom’s, Tub) Albright’s hereditary osteodystrophy (PHP) Cohen syndrome Single minded 1 orthopedia Brain-derived neurotrophic factor Tyrosine receptor kinase B Src homology 2B adaptor protein 1 Pleckstrin domain homology interacting protein Serotonin 2C receptor Leptin Leptin receptor Pro-opiomelanocortin Prohormone convertase 1 Carboxypeptidase E Steroid receptor co-activator-1 Semaphorin 3A G Melanocortin 4 re Melanocortin rec Adenylate cyclas Kinase suppresso Normal development https://doi.org/10.1038/s41591-022-02106-5 Additional information was obtained on their family members who carried variants, which is included directly beneath the proband where available; family relationships are indicated in brackets (m, mother; f, father; ma, maternal aunt; s, sister; b, brother; mgf, maternal grandfather.) Age and sex-adjusted SDS for height and BMI were included for individuals up to 18 years; sex (M, male; F, female); TSH, thyroid stimulating hormone (0.35–5.50 mIU l−1); FT, free thyroxine (10.0–19.8 pmol l−1); fasting plasma insulin (0–60 pmol l−1); fasting glucose (3.5–5.5 mmol l−1). In some cases, biochemical test results were not available but values were recorded as within the normal range in clinical letters (recorded as normal here). Basal metabolic rate (BMR) and respiratory quotient (RQ) were measured by indirect calorimetry in the fasted state and BMR was predicted on the basis of age, sex and body composition shown for comparison. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured in the rested, fasted state in mm Hg. Blank indicates data not available. ADHD, attention deficit hyperactivity disorder; ODD, oppositional defiant disorder; OCD, obsessive-compulsive disorder; IFG, impaired fasting glycemia; PCOS, polycystic ovarian syndrome. Genetic obesity syndromes Developmental delay, learning and/or behavioral dificulties Normal development Retinal dystrophies Skeletal abnormalities Dysmorphic features Autistic traits Anxiety Aggression Hyperactivity Endocrine features Prader–Willi Ciliopathies (Bardet–Biedl, Alstrom’s, Tub) Albright’s hereditary osteodystrophy (PHP) Cohen syndrome Single minded 1 orthopedia Brain-derived neurotrophic factor Tyrosine receptor kinase B Src homology 2B adaptor protein 1 Pleckstrin domain homology interacting protein Serotonin 2C receptor Leptin Leptin receptor Pro-opiomelanocortin Prohormone convertase 1 Carboxypeptidase E Steroid receptor co-activator-1 Semaphorin 3A-G Plexin A1-4 Neuropilin 1 and 2 Melanocortin 4 receptor Melanocortin receptor accessory protein 2 Adenylate cyclase 3 Kinase suppressor of Ras-2 Fig. 2 \| Genetic obesity syndromes. A schematic depicting genetic obesity syndromes, conditions where severe childhood-onset obesity is a major presenting clinical feature. Conditions in bold are those where the gene/genes have been shown to affect signaling through the leptin–melanocortin pathway and could be treated by an MC4R agonist. Clinical features seen in people with LOF variants in the gene encoding the serotonin 2C receptor (orange) are reported in this study. Genetic obesity syndromes Developmental delay, learning and/or behavioral dificulties Developmental delay, learning and/or behavioral dificulties Retinal dystrophies Skeletal abnormalities Dysmorphic features Autistic traits Anxiety Aggression Hyperactivity Prader–Willi Ciliopathies (Bardet–Biedl, Alstrom’s, Tub) Albright’s hereditary osteodystrophy (PHP) Cohen syndrome Single minded 1 orthopedia Brain-derived neurotrophic factor Tyrosine receptor kinase B Src homology 2B adaptor protei Pleckstrin domain homology inte Serotonin 2C receptor Melanocortin 4 receptor Melanocortin receptor accessory protein 2 Adenylate cyclase 3 Kinase suppressor of Ras-2 Melanocortin 4 receptor Melanocortin receptor accessory protein 2 Adenylate cyclase 3 Kinase suppressor of Ras-2 Brain-derived neurotrophic factor Tyrosine receptor kinase B Src homology 2B adaptor protein 1 Pleckstrin domain homology interacting protein Serotonin 2C receptor Leptin Leptin receptor Pro-opiomelanocortin Prohormone convertase 1 Carboxypeptidase E Steroid receptor co-activator-1 Semaphorin 3A-G Plexin A1-4 Neuropilin 1 and 2 Fig. 2 \| Genetic obesity syndromes. A schematic depicting genetic obesity syndromes, conditions where severe childhood-onset obesity is a major presenting clinical feature. Conditions in bold are those where the gene/genes have been shown to affect signaling through the leptin–melanocortin pathway Fig. 2 \| Genetic obesity syndromes. A schematic depicting genetic obesity syndromes, conditions where severe childhood-onset obesity is a major presenting clinical feature. Conditions in bold are those where the gene/genes have been shown to affect signaling through the leptin–melanocortin pathway and could be treated by an MC4R agonist. Clinical features seen in people with LOF variants in the gene encoding the serotonin 2C receptor (orange) are reported in this study. and could be treated by an MC4R agonist. Clinical features seen in people with LOF variants in the gene encoding the serotonin 2C receptor (orange) are reported in this study. Developmental delay, learning and/or behavioral dificulties Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2540 Article https://doi.org/10.1038/s41591-022-02106-5 WT V2L N6K A7V V61I I97V A171V V208M N213T S260G C266R Q282H F327L T419A WT V2L N6K A7V V61I I97V A171V V208M N213T S260G C266R Q282H F327L T419A 0 50 100 150 200 Percentage WT 0 50 100 150 200 Percentage WT 0 50 100 150 200 Percentage WT 0 50 100 150 200 Percentage WT Cell surface expression (ELISA) Cell surface expression (fluorescence) WT F327L ER-to-cytoplasm ratio –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) 0 50 100 150 200 Percentage WT 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) log [5-HT] (M) 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) IP1 accumulation IP1 accumulation Mock WT V2L N6K A7V V61I I97V A171V V208M Mock WT N213T S260G C266R Q282H F327L T419A β-arrestin-1 recruitment WT V2L N6K A7V V61I I97V A171V V208M β-arrestin-1 recruitment WT S260G C266R Q282H F327L T419A N213T β-arrestin-2 recruitment β-arrestin-2 recruitment V2L N6K A7V V61I I97V A171V V208M WT WT N213T S260G C266R Q282H F327L T419A –10 –8 –6 -4 0 3H mesulergine binding Mock WT F327L a b c d e f g i h j P = 0.0235 P = 0.0343 P = 0.0013 P = 0.0423 P = 0.0137 TM-VI TM-VII TM-V TM-VI TM-V TM-IV TM-III TM-II TM-VII Phe327 Ritanserin Phe327 Ergotamine TM-III TM-II TM-I TM-I Not shown Not shown Not shown Not shown TM-IV Mock WT F327L 5HT2CR 5HT2CR ER Merge DAPI Non-permeabilized cells Permeabilized cells Fig. 3 \| Human variants in the gene encoding 5-HT2CR cause a loss of function in cells. Wild-type (WT) and mutant forms of 5-HT2CR were studied in cells. a,b, Cell surface localization of WT and mutant receptors measured by ELISA (a) and high-content confocal microscopy (b). Values are expressed as percentage of WT and represent mean ± s.e.m., n = 3–4 independent experiments; mock transfected cells served as negative controls. Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 Developmental delay, learning and/or behavioral dificulties Data are shown as sum curves normalized to WT response ± s.e.m. EC50 and maximum response (Emax) for each variant is shown in Supplementary Table 1. h, Competitive binding assay for WT versus F327L 5-HT2CR using the receptor antagonist 3H-mesulergine with increasing concentrations of the agonist, 5-HT. Sum curves are shown from n = 3 independent experiments normalized to WT max binding ± s.e.m. Affinity (IC50) and number of binding sites (Bmax) are shown in Supplementary Table 1; significant differences from receptor were determined using a two-tailed Student’s t-test with Welch correction. i,j, Structural analysis of the 5-HT2CR in an active (i) and inactive conformation (j), bound to the agonist ergotamine and the inverse agonist ritanserin, respectively. TM, transmembrane domains (I–VII). The F327 (Phe327) residue is highlighted and predicted to interact with both ligands21. Effects on receptor-mediated activation of Gαq/11-regulated inositol triphosphate signaling (IP1 accumulation; n = 3–4 independent experiments) (e) and coupling to β-arrestin-1 and -2 (n = 4–5 independent experiments) (f,g). Data are shown as sum curves normalized to WT response ± s.e.m. EC50 and maximum response (Emax) for each variant is shown in Supplementary Table 1. h, Competitive binding assay for WT versus F327L 5-HT2CR using the receptor antagonist 3H-mesulergine with increasing concentrations of the agonist, 5-HT. Sum curves are shown from n = 3 independent experiments normalized to WT max binding ± s.e.m. Affinity (IC50) and number of binding sites (Bmax) are shown in Supplementary Table 1; significant differences from receptor were determined using a two-tailed Student’s t-test with Welch correction. i,j, Structural analysis of the 5-HT2CR in an active (i) and inactive conformation (j), bound to the agonist ergotamine and the inverse agonist ritanserin, respectively. TM, transmembrane domains (I–VII). The F327 (Phe327) residue is highlighted and predicted to interact with both ligands21. Fig. 3 \| Human variants in the gene encoding 5-HT2CR cause a loss of function in cells. Wild-type (WT) and mutant forms of 5-HT2CR were studied in cells. a,b, Cell surface localization of WT and mutant receptors measured by ELISA (a) and high-content confocal microscopy (b). Values are expressed as percentage of WT and represent mean ± s.e.m., n = 3–4 independent experiments; mock transfected cells served as negative controls. Differences between WT and mutant receptors were compared using two-tailed Student’s t-tests with Welch correction. Developmental delay, learning and/or behavioral dificulties https://doi.org/10.1038/s41591-022-02106-5 Article WT V2L N6K A7V V61I I97V A171V V208M N213T S260G C266R Q282H F327L T419A WT V2L N6K A7V V61I I97V A171V V208M N213T S260G C266R Q282H F327L T419A 0 50 100 150 200 Percentage WT 0 50 100 150 200 Percentage WT 0 50 100 150 200 Percentage WT 0 50 100 150 200 Percentage WT Cell surface expression (ELISA) Cell surface expression (fluorescence) WT F327L ER-to-cytoplasm ratio –10 –9 –8 –7 –6 –5 0 [ ] ) 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 [ ] ) IP1 accumulation IP1 accumulation Mock WT V2L N6K A7V V61I I97V A171V V208M Mock WT N213T S260G C266R Q282H F327L T419A a b c d e P = 0.0235 P = 0.0343 P = 0.0013 P = 0.0423 P = 0.0137 Mock WT F327L 5HT2CR 5HT2CR ER Merge DAPI Non-permeabilized cells Permeabilized cells WT V2L N6K A7V V61I I97V A171V V208M N213T S260G C266R Q282H F327L T419A 0 50 100 150 200 Percentage WT Cell surface expression (fluorescence) b P = 0.0423 WT V2L N6K A7V V61I I97V A171V V208M N213T S260G C266R Q282H F327L T419A 0 50 100 150 200 Percentage WT Cell surface expression (ELISA) a P = 0.0235 P = 0.0343 P = 0.0013 0 50 100 150 200 Percentage WT WT F327L ER-to-cytoplasm ratio c P = 0.0137 d Mock WT F327L 5HT2CR 5HT2CR ER Merge DAPI Non-permeabilized cells Permeabilized cells b d N a Mock 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) IP1 accumulation Mock WT V2L N6K A7V V61I I97V A171V V208M e 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) IP1 accumulation Mock WT N213T S260G C266R Q282H F327L T419A e 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) β-arrestin-2 recruitment WT N213T S260G C266R Q282H F327L T419A 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) β-arrestin-1 recruitment WT V2L N6K A7V V61I I97V A171V V208M f 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) β-arrestin-1 recruitment WT V2L N6K A7V V61I I97V A171V V208M β-arrestin-1 recruitment WT S260G C266R Q282H F327L T419A N213T f 0 50 100 150 200 Percentage WT log [5-HT] (M) –10 –8 –6 -4 0 3H mesulergine binding Mock WT F327L h 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) β-arrestin-1 recruitment WT S260G C266R Q282H F327L T419A N213T 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) 0 50 100 150 200 Percentage WT –10 –9 –8 –7 –6 –5 0 log [5-HT] (M) β-arrestin-2 recruitment β-arrestin-2 recruitment V2L N6K A7V V61I I97V A171V V208M WT WT N213T S260G C266R Q282H F327L T419A g f h g j TM-VI TM-V TM-IV TM-III TM-II TM-VII Phe327 Ritanserin TM-I Not shown Not shown i TM-VI TM-VII TM-V Phe327 Ergotamine TM-III TM-II TM-I Not shown Not shown TM-IV j i i Effects on receptor-mediated activation of Gαq/11-regulated inositol triphosphate signaling (IP1 accumulation; n = 3–4 independent experiments) (e) and coupling to β-arrestin-1 and -2 (n = 4–5 independent experiments) (f,g). Developmental delay, learning and/or behavioral dificulties Differences between WT and mutant receptors were compared using two-tailed Student’s t-tests with Welch correction. c, Expression of WT versus F327L 5-HT2CR in the ER compared to the cytoplasm, quantified by high-content confocal microscopy. Data represent mean ± s.e.m., n = 3 independent experiments; differences between WT and mutant receptors were determined by two-tailed Student’s t-tests with Welch correction. d, A representative micrograph of subcellular localization of WT and F327L 5-HT2CR in non-permeabilized and permeabilized cells. Scale bars, 50 μm. 4,6-Diamidino-2-phenylindole (DAPI) was used to stain the nucleus. e–g, Effects on receptor-mediated activation of Gαq/11-regulated inositol triphosphate signaling (IP1 accumulation; n = 3–4 independent experiments) (e) and coupling to β-arrestin-1 and -2 (n = 4–5 independent experiments) (f,g). Data are shown as sum curves normalized to WT response ± s.e.m. EC50 and maximum response (Emax) for each variant is shown in Supplementary Table 1. h, Competitive binding assay for WT versus F327L 5-HT2CR using the receptor antagonist 3H-mesulergine with increasing concentrations of the agonist, 5-HT. Sum curves are shown from n = 3 independent experiments normalized to WT max binding ± s.e.m. Affinity (IC50) and number of binding sites (Bmax) are shown in Supplementary Table 1; significant differences from receptor were determined using a two-tailed Student’s t-test with Welch correction. i,j, Structural analysis of the 5-HT2CR in an active (i) and inactive conformation (j), bound to the agonist ergotamine and the inverse agonist ritanserin, respectively. TM, transmembrane domains (I–VII). The F327 (Phe327) residue is highlighted and predicted to interact with both ligands21. Developmental delay, learning and/or behavioral dificulties c, Expression of WT versus F327L 5-HT2CR in the ER compared to the cytoplasm, quantified by high-content confocal microscopy. Data represent mean ± s.e.m., n = 3 independent experiments; differences between WT and mutant receptors were determined by two-tailed Student’s t-tests with Welch correction. d, A representative micrograph of subcellular localization of WT and F327L 5-HT2CR in non-permeabilized and permeabilized cells. Scale bars, 50 μm. 4,6-Diamidino-2-phenylindole (DAPI) was used to stain the nucleus. e–g, Effects on receptor-mediated activation of Gαq/11-regulated inositol triphosphate signaling (IP1 accumulation; n = 3–4 independent experiments) (e) and coupling to β-arrestin-1 and -2 (n = 4–5 independent experiments) (f,g). Data are shown as sum curves normalized to WT response ± s.e.m. EC50 and maximum response (Emax) for each variant is shown in Supplementary Table 1. h, Competitive binding assay for WT versus F327L 5-HT2CR using the receptor antagonist 3H-mesulergine with increasing concentrations of the agonist, 5-HT. Sum curves are shown from n = 3 independent experiments normalized to WT max binding ± s.e.m. Affinity (IC50) and number of binding sites (Bmax) are shown in Supplementary Table 1; significant differences from receptor were determined using a two-tailed Student’s t-test with Welch correction. i,j, Structural analysis of the 5-HT2CR in an active (i) and inactive conformation (j), bound to the agonist ergotamine and the inverse agonist ritanserin, respectively. TM, transmembrane domains (I–VII). The F327 (Phe327) residue is highlighted and predicted to interact with both ligands21. Fig. 3 \| Human variants in the gene encoding 5-HT2CR cause a loss of function in cells. Wild-type (WT) and mutant forms of 5-HT2CR were studied in cells. a,b, Cell surface localization of WT and mutant receptors measured by ELISA (a) and high-content confocal microscopy (b). Values are expressed as percentage of WT and represent mean ± s.e.m., n = 3–4 independent experiments; mock transfected cells served as negative controls. Differences between WT and mutant receptors were compared using two-tailed Student’s t-tests with Welch correction. c, Expression of WT versus F327L 5-HT2CR in the ER compared to the cytoplasm, quantified by high-content confocal microscopy. Data represent mean ± s.e.m., n = 3 independent experiments; differences between WT and mutant receptors were determined by two-tailed Student’s t-tests with Welch correction. d, A representative micrograph of subcellular localization of WT and F327L 5-HT2CR in non-permeabilized and permeabilized cells. Scale bars, 50 μm. 4,6-Diamidino-2-phenylindole (DAPI) was used to stain the nucleus. e–g, Fig. n Htr2cF327L mice develop hyperphagia and obesity Knock in Htr2c mice develop hyperphagia and obesity To determine whether a human LOF variant affecting 5-HT2CR (F327L) can cause obesity, we used a CRISPR-Cas9 approach to generate a knock-in Htr2cF327L mouse line (Methods) (Extended Data Fig. 3a). Experiments were performed in heterozygous female mice (carry- ing a variant on one X chromosome; wild-type gene sequence on the other X chromosome) and hemizygous male mice (carrying a vari- ant on their only X chromosome). When fed with a regular chow diet, male Htr2cF327L/Y hemizygous mice showed modest body weight gain compared to wild-type littermates, associated with increased fat mass (Fig. 4a,b). In addition, male Htr2cF327L/Y mice exhibited increased food intake compared to wild-type littermates from 7 weeks of age before body weight divergence was apparent (Fig. 4c). We confirmed hyper- phagia in male Htr2cF327L/Y mice using TSE PhenoMaster metabolic cages (Fig. 4d), which also revealed an increase in physical activity (Fig. 4e,f). No changes were detected in overall O2 consumption, CO2 and heat production when using body weight as a covariate, whereas subtle increases were detected at intervals during the 2-d record- ing period in male Htr2cF327L/Y mice (Extended Data Fig. 3b–d). When challenged with chronic high fat diet (HFD) feeding, male Htr2cF327L/Y mice showed greater body weight gain, associated with increased fat mass and lean mass (Fig. 4g,h and Extended Data Fig. 4a). Consistent with increased adiposity, male Htr2cF327L/Y mice displayed impaired glucose tolerance and impaired insulin sensitivity (Extended Data Fig. 4b,c). Male Htr2cF327L/Y mice rapidly developed hyperphagia on an HFD before divergence in body weight (Fig. 4i). The TSE PhenoM- aster metabolic cage study confirmed increased HFD consumption (Fig. 4j) and revealed increases in physical activity and no changes in O2 consumption, CO2 and heat production (Extended Data Fig. 4d–h). Collectively, these results indicate that the F327L 5-HT2CR variant leads to weight gain in male mice due primarily to hyperphagia, despite increased physical activity. Peng and colleagues reported the crystal structure of 5-HT2CR bound to the agonist ergotamine and the inverse agonist ritanserin21. Their model predicted the F327 residue to be critical to binding of these compounds to the receptor. Moreover, substituting F327 to a leucine impaired predicted intermolecular (π–π) interactions and reduced the receptor’s affinity for ergotamine and 5-HT21. Here, in a competitive binding assay with the 5-HT2CR antagonist mesulergine, we observed decreased 5-HT binding affinity and higher maximal mesulergine bind- ing (Fig. Functional characterization of HTR2C variants To investigate the functional consequences of rare variants in HTR2C, we performed a series of experiments in cells transiently transfected with constructs encoding wild-type or mutant human 5-HT2CR. The functional activity of Gαq/11-coupled 5-HT2CRs is regulated by desensi- tization and re-sensitization which involves the recruitment of intra- cellular β-arrestins that promote the phosphorylation of extracellular signal-regulated kinase 1/2 (ERK1/2)16,17 and thus, gene transcription. The site where 5-HT2CR interacts with G proteins is prone to RNA edit- ing, which affects the receptor’s affinity for G proteins and hence its constitutive activity18,19. We studied variants in the ‘VSV’ isoform, the most prevalent isoform in the human hypothalamus20. We performed exploratory analysis of mental health data collected as part of the UK Biobank (n = 53,174; Methods and Supplementary Table 2). The V61I HTR2C variant was identified in 24 unrelated people in the UK Biobank; Mental Health Questionnaire data were available for ten of these individuals, five of whom (2 males and 3 females) had a diagnosis of depression; Fisher’s exact test, OR (95% CI) = 3.7 (0.8–16), P = 0.04 and three had additional diagnoses of ‘anxiety, nerves or gen- eralized anxiety disorder’ or ‘phobias’; while individual associations do not survive multiple testing across traits and variants, these findings are compatible with our hypothesis that rare HTR2C variants contribute to obesity and maladaptive behavior in humans. Using a cell surface ELISA and high-content confocal microscopy to quantify receptors in different cellular compartments, we observed markedly reduced plasma membrane expression of one mutant, F327L (53% of wild-type, by ELISA; and 52% of wild-type, by microscopy) and slightly reduced expression of another (V2L) (83% of wild-type, by ELISA) (Fig. 3a,b, Extended Data Fig. 2a, and Supplementary Table 1). F327L 5-HT2CR colocalized with the endoplasmic reticulum (ER) marker, calreticulin (Fig. 3c,d and Extended Data Fig. 2b,c) consistent with its partial retention in this cellular compartment; other mutants did not affect cellular localization. We next quantified ligand-induced IP3 turnover and found that, compared to wild-type 5-HT2CR, F327L resulted in a LOF, decreasing both the maximal efficacy (Emax) of sign- aling and increasing the half-maximal effective concentration (EC50) (Fig. 3e, Supplementary Table 1); none of the other mutants affected signaling in this assay. Using a NanoBiT protein–protein interaction assay, we found that nine mutants reduced agonist-induced β-arrestin-1 recruitment (Fig. 3f and Supplementary Table 1) and eight reduced β-arrestin-2 recruitment, (Fig. Developmental delay, learning and/or behavioral dificulties 3 \| Human variants in the gene encoding 5-HT2CR cause a loss of function in cells. Wild-type (WT) and mutant forms of 5-HT2CR were studied in cells. a,b, Cell surface localization of WT and mutant receptors measured by ELISA (a) and high-content confocal microscopy (b). Values are expressed as percentage of WT and represent mean ± s.e.m., n = 3–4 independent experiments; mock transfected cells served as negative controls. Differences between WT and mutant receptors were compared using two-tailed Student’s t-tests with Welch correction. c, Expression of WT versus F327L 5-HT2CR in the ER compared to the cytoplasm, quantified by high-content confocal microscopy. Data represent mean ± s.e.m., n = 3 independent experiments; differences between WT and mutant receptors were determined by two-tailed Student’s t-tests with Welch correction. d, A representative micrograph of subcellular localization of WT and F327L 5-HT2CR in non-permeabilized and permeabilized cells. Scale bars, 50 μm. 4,6-Diamidino-2-phenylindole (DAPI) was used to stain the nucleus. e–g, language delay (n = 609, 7.8%) and autism spectrum disorder (n = 382, 4.9%) were frequently reported at the time of recruitment (mean age 10 years). Behavioral issues (not solely around food) were reported in 271 children (3.4%), hyperactivity (n = 112, 1.4%) and anxiety or social anxiety (n = 77, 1%) were sometimes reported. take part. Body composition measured by dual energy X-ray absorpti- ometry showed excess body fat mass (mean (± s.d.) percentage body fat 46.6 ± 14.5) and a normal bone mineral density (BMD) SDS (mean 0.38). Basal metabolic rate measured by indirect calorimetry after an overnight fast was comparable to that predicted on the basis of age, sex and body composition (Table 1). Carriers of variants had normal blood pressures measured in the rested, fasted state (Table 1). Fasting glucose and insulin levels and HOMA-IR (homeostasis model assessment of To investigate the clinical phenotypes seen in people carrying HTR2C variants, we invited all probands and their family members to take part in physiological studies (Methods); six individuals agreed to Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2541 https://doi.org/10.1038/s41591-022-02106-5 Article as a continuous trait (Methods and Supplementary Table 2). Developmental delay, learning and/or behavioral dificulties In males, gene-region rare variant tests showed a nominal association of rare (allele frequency <0.1%) non-synonymous exonic or splice-site vari- ants and obesity (BMI > 30 kg m−2, P = 0.01, SKAT binary robust burden; P = 0.01, SKATO; odds ratio (OR) (95% CI) = 1.9 (1.1–3.1); P = 0.01, Fisher’s exact test). Among females, we found eight non-synonymous coding variants with an allele count >20 including three variants identified in the GOOS cohort (V61I, A171V and T419A); none was associated with BMI in single-variant analyses (Methods). Among variants identified in the GOOS cohort and shown to cause LOF in cells (allele count <10 in the UK Biobank), four of the five people carrying the Q282H HTR2C variant had obesity (BMI > 30 kg m−2; OR (95% CI) = 14 (1.3–670), nomi- nal P = 0.01, Fisher’s exact test), 5 of 9 males carrying A7V had obesity (BMI > 30 kg m−2; P = 0.03, SKAT robust; OR (95% CI) = 3.8 (0.8–19), nominal P = 0.05, Fisher’s exact test). insulin resistance) were appropriate for the degree of obesity when people carrying variants were compared to 2,138 controls of a com- parable age, sex and BMI (Extended Data Fig. 1). Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 n Htr2cF327L mice develop hyperphagia and obesity 3h), the latter suggesting that the F327L substitution stabilizes the inactive conformation of the receptor (Fig. 3i,j). Cumulatively, these studies demonstrate that the F327L variant causes severe LOF by disrupting a critical residue involved in the binding of 5-HT. Ten other missense mutations found in individuals with severe obesity caused a partial LOF, predominantly by affecting the recruitment of β-arrestins. Functional characterization of HTR2C variants 3g and Supplementary Table 1); F327L increased the EC50 in assays for both β-arrestin-1 and -2 recruitment (Supplementary Table 1). In total, 11 of 13 mutants found in individuals with severe obesity caused a LOF in at least one functional assay. The variant found in people with obesity and controls (A171V) caused a LOF in HEK293 cells. Article 6 9 12 15 18 15 20 25 30 35 Age (weeks) g WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y * P = 0.2957 Fat Lean 0 5 10 20 25 30 35 g P = 0.029 6 9 12 15 18 16 18 20 22 24 26 28 Age (weeks) g * * * P = 0.7815 0 2 4 6 8 Time g 18:00 6:00 18:00 6:00 18:00 P > 0.9999 Dark Light 24 h 0 1 2 3 4 5 g P = 0.012 0 1 2 3 4 5 6 7 8 0 5 10 15 20 Weeks on HFD g * * * P = 0.003 Fat Lean 0 10 20 30 40 50 g P = 0.026 P = 0.003 1 2 3 4 5 6 7 8 10 15 20 25 30 Weeks on HFD g * * * * 0 3 6 9 12 Time g * * * 18:00 6:00 18:00 6:00 18:00 Dark Light 24 h 0 2 4 6 8 g P = 0.023 18:00 6:00 18:00 6:00 18:00 0 20,000 40,000 60,000 Time 18:00 6:00 18:00 6:00 18:00 Time Counts h–1 Counts h–1 Counts h–1 WT Htr2cF327L/Y WT Htr2cF327L/Y * * * * * * * P = 0.5152 P = 0.4596 P = 0.9925 Dark Light 0 20,000 40,000 60,000 P = 0.012 0 2,000 4,000 6,000 8,000 Counts h–1 0 2,000 4,000 6,000 8,000 * * * * P = 0.0178 Dark Light P = 0.022 a b c d g h i e f k l WT Htr2cF327L/Y Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y 0 0.5 1.0 1.5 g Saline Lorcaserin P = 0.004 P = 0.516 0 0.5 1.0 1.5 g Saline Leptin P < 0.0001 P = 0.0003 42.86% depolarized 57.14% no response WT 18 24 Total = 42 20.00% depolarized 80.00% no response 7 28 Total = 35 0 2 4 6 0 2 4 6 Depolarization (mV) P = 0.017 ∆Firing frequency (s–1) P = 0.008 j m n o p WT Htr2cF327L/Y Lorcaserin Lorcaserin 20 s P = 0.0329 in χ2 test Fig. 4 \| Knock-in Htr2cF327L mice develop hyperphagic obesity and are less responsive to 5-HT2CR agonism. Article WT Htr2cF327L/Y * 0 2 4 6 8 Time g 18:00 6:00 18:00 6:00 18:00 P > 0.9999 Dark Light 24 h 0 1 2 3 4 5 g P = 0.012 d WT Htr2cF327L/Y * 0 2 4 6 8 Time g 18:00 6:00 18:00 6:00 18:00 P > 0.9999 Da 0 1 2 3 4 5 g d WT Htr2cF327L/Y Fat Lean 0 5 10 20 25 30 35 g P = 0.029 b WT Htr2cF327L/Y 6 9 12 15 18 16 18 20 22 24 26 28 Age (weeks) g * * P = 0.7815 c 6 9 12 15 18 15 20 25 30 35 Age (weeks) g WT Htr2cF327L/Y * P = 0.2957 a d b c a WT Htr2cF327L/Y 18:00 6:00 18:00 6:00 18:00 Time Counts h–1 WT Htr2cF327L/Y 0 2,000 4,000 6,000 8,000 Counts h–1 0 2,000 4,000 6,000 8,000 * * * * P = 0.0178 Dark Light P = 0.022 f WT Htr2cF327L/Y WT Htr2cF327L/Y 18:00 6:00 18:00 6:00 18:00 0 20,000 40,000 60,000 Time Counts h–1 Counts h–1 * * * * * * * P = 0.5152 Dark Light 0 20,000 40,000 60,000 P = 0.012 e WT Htr2cF327L/Y Fat Lean 0 10 20 30 40 50 g P = 0.026 P = 0.003 h f WT Htr2cF327L/Y 0 1 2 3 4 5 6 7 8 0 5 10 15 20 Weeks on HFD g * * * P = 0.003 g h g e k WT Htr2cF327L/Y 0 0.5 1.0 1.5 g Saline Lorcaserin P = 0.004 P = 0.516 l 1 j l WT Htr2cF327L/Y 0 0.5 1.0 1.5 g Saline Leptin P < 0.0001 P = 0.0003 l WT Htr2cF327L/Y 0 0.5 1.0 1.5 g Saline Leptin P < 0.0001 P = 0.0003 WT Htr2cF327L/Y 1 2 3 4 5 6 7 8 10 15 20 25 30 Weeks on HFD g * * * * P = 0.4596 i 0 3 6 9 12 Time g * * * 18:00 6:00 18:00 6:00 18:00 WT Htr2cF327L/Y P = 0.9925 j i WT Htr2cF327L/Y Dark Light 24 h 0 2 4 6 8 g P = 0.023 k 1 k Htr2cF327L/Y olarized esponse 20.00% depolarized 80.00% no response 7 28 Total = 35 P = 0.0329 in χ2 test WT Htr2cF327L/Y 0 2 4 6 Depolarization (mV) P = 0.017 o m WT Htr2cF327L/Y Lorcaserin Lorcaserin 20 s WT Htr2cF327L/Y 0 2 4 6 ∆Firing frequency (s–1) P = 0.008 p 42.86% depolarized 57.14% no response WT 18 24 Total = 42 n P = 0.032 p n m o Total = 35 20.00% depol 80.00% no res Total = 35 20.00% depolarized 80.00% no response 42.86% depolarized 57.14% no response injections of saline or lorcaserin (3 mg kg–1). Article a–c, Body weight curves (a), body composition (b) and weekly chow intake (c) of male WT (n = 10) and Htr2cF327L/Y mice (n = 10) fed on regular chow. d–f, Cumulative chow intake (d), temporal levels of xy axis activity (e) and z axis activity (f) during a 2-d period measured by the TSE PhenoMaster apparatus in male WT (n = 7) and Htr2cF327L/Y (n = 9) mice (left). Averaged values during the dark or light cycle on left (right). g–i, Body weight curves (g), body composition (h) and weekly HFD intake (i) of male WT (n = 7) and Htr2cF327L/Y (n = 9) mice fed on HFD. j, Cumulative HFD intake during a 2-d period in male WT (n = 7) and Htr2cF327L/Y (n = 9) mice (left). Averaged HFD intake during the dark cycle, light cycle or 24 h (right). k, One-hour food intake in chow-fed male WT (n = 8) and Htr2cF327L/Y (n = 8) mice (6 months of age) after i.p. injections of saline or lorcaserin (3 mg kg–1). l, One-hour food intake in chow-fed male WT (n = 9) and Htr2cF327L/Y mice (n = 14) (2 months of age) after i.p. injections of saline or leptin (5 mg kg−1). m, Representative traces in POMC neurons from WT and Htr2cF327L/Y mice treated with lorcaserin. n, Percentage/number of POMC neurons from WT and Htr2cF327L/Y mice that were depolarized by or irresponsive to lorcaserin. o,p, Resting membrane potential, depolarization (o) and increases in firing frequency (p) induced by lorcaserin in WT (n = 18) and Htr2cF327L/Y (n = 7) mice. Two-way analysis of variance followed by Sidak’s multiple comparisons were performed in a,c–f (left), g,i–I. P values in some individual data points in a,c,d–f (left), i,j were determined by two-tailed unpaired Student’s t-test. Two- tailed unpaired Student’s t-tests were performed in b,d,e,h,j (right), f (left), o,p. P < 0.05, P < 0.01 and *P < 0.001. Data are presented as mean ± s.e.m. Article d–f, Cumulative chow intake (d), temporal levels of xy axis activity (e) and z axis activity (f) during a 2-d period measured by the TSE PhenoMaster apparatus in male WT (n = 7) and Htr2cF327L/Y (n = 9) mice (left). Averaged values during the dark or light cycle on left (right). g–i, Body weight curves (g), body composition (h) and weekly HFD intake (i) of male WT (n = 7) and Htr2cF327L/Y (n = 9) mice fed on HFD. j, Cumulative HFD intake during a 2-d period in male WT (n = 7) and Htr2cF327L/Y (n = 9) mice (left). Averaged HFD intake during the dark cycle, light cycle or 24 h (right). k, One-hour food intake in chow-fed male WT (n = 8) and Htr2cF327L/Y (n = 8) mice (6 months of age) after i.p. wild-type mice, P = 0.0329, chi-squared test; Fig. 4m,n). We found that lorcaserin-induced depolarization and increased firing frequency were attenuated in POMC neurons from mutant mice compared to wild-type mice (Fig. 4o,p). Consistently, lorcaserin-induced c-fos expression in POMC neurons was reduced in both male Htr2cF327L/Y mice and female Htr2cF327L/+ mice (Extended Data Fig. 6f–i). obesity associated with glucose intolerance and insulin resistance (Extended Data Fig. 5c–i). Modest increases in physical activity were detected in HFD-fed female Htr2cF327L/+ mice; there was no change in O2 consumption, CO2 and heat production (Extended Data Fig. 5j–n). These results indicate that the heterozygous F327L 5-HT2CR mutation caused hyperphagic obesity in female mice on a HFD. Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 Rare HTR2C variants in a population-based cohort To investigate whether rare variants in HTR2C are associated with increased BMI or obesity in the population, we examined 200,000 exomes from a large population-derived cohort, the UK Biobank22. As HTR2C is located on the X chromosome, we performed analyses separately for unrelated males (n = 69,488) and females (n = 83,864) of European ancestry. Five people in the UK Biobank dataset had frameshift or stop-gain mutations in HTR2C; three of the five males had a BMI > 30 kg m−2; there were no female carriers. We studied rare HTR2C variants using gene-region and single-variant association analysis for obesity (BMI > 30 kg m−2), severe obesity (BMI > 40 kg m−2) and for BMI As the F327L 5-HT2CR variant was identified in the heterozygous form in a female proband with obesity, we characterized metabolic phenotypes in female Htr2cF327L/+ heterozygous mice. When fed on a chow diet, female Htr2cF327L/+ mice showed comparable body weight and food intake to female wild-type littermates (Extended Data Fig. 5a,b). When fed on HFD, female Htr2cF327L/+ mice developed hyperphagic Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2542 https://doi.org/10.1038/s41591-022-02106-5 Article Article l, One-hour food intake in chow-fed male WT (n = 9) and Htr2cF327L/Y mice (n = 14) (2 months of age) after i.p. injections of saline or leptin (5 mg kg−1). m, Representative traces in POMC neurons from WT and Htr2cF327L/Y mice treated with lorcaserin. n, Percentage/number of POMC neurons from WT and Htr2cF327L/Y mice that were depolarized by or irresponsive to lorcaserin. o,p, Resting membrane potential, depolarization (o) and increases in firing frequency (p) induced by lorcaserin in WT (n = 18) and Htr2cF327L/Y (n = 7) mice. Two-way analysis of variance followed by Sidak’s multiple comparisons were performed in a,c–f (left), g,i–I. P values in some individual data points in a,c,d–f (left), i,j were determined by two-tailed unpaired Student’s t-test. Two- tailed unpaired Student’s t-tests were performed in b,d,e,h,j (right), f (left), o,p. P < 0.05, P < 0.01 and P < 0.001. Data are presented as mean ± s.e.m. Fig. 4 \| Knock-in Htr2cF327L mice develop hyperphagic obesity and are less responsiveto5-HT Ragonism a–c Bodyweightcurves(a) bodycompositio Fig. 4 \| Knock-in Htr2cF327L mice develop hyperphagic obesity and are less responsive to 5-HT2CR agonism. a–c, Body weight curves (a), body composition (b) and weekly chow intake (c) of male WT (n = 10) and Htr2cF327L/Y mice (n = 10) fed on regular chow. d–f, Cumulative chow intake (d), temporal levels of xy axis activity (e) and z axis activity (f) during a 2-d period measured by the TSE PhenoMaster apparatus in male WT (n = 7) and Htr2cF327L/Y (n = 9) mice (left). Averaged values during the dark or light cycle on left (right). g–i, Body weight curves (g), body composition (h) and weekly HFD intake (i) of male WT (n = 7) and Htr2cF327L/Y (n = 9) mice fed on HFD. j, Cumulative HFD intake during a 2-d period in male WT (n = 7) and Htr2cF327L/Y (n = 9) mice (left). Averaged HFD intake during the dark cycle, light cycle or 24 h (right). k, One-hour food intake in chow-fed male WT (n = 8) and Htr2cF327L/Y (n = 8) mice (6 months of age) after i.p. Fig. 4 \| Knock-in Htr2cF327L mice develop hyperphagic obesity and are less responsive to 5-HT2CR agonism. a–c, Body weight curves (a), body composition (b) and weekly chow intake (c) of male WT (n = 10) and Htr2cF327L/Y mice (n = 10) fed on regular chow. Htr2cF327L mice are less responsive to 5-HT2CR agonism Htr2cF327L mice are less responsive to 5-HT2CR agonism We next examined the effect of a 5-HT2CR agonist, lorcaserin, on food intake. Lorcaserin-induced anorexia was blunted in both male Htr2cF327L/Y mice and female Htr2cF327L/+ mice compared to wild-type mice (Fig. 4k and Extended Data Fig. 6a). The F327L mutation did not affect leptin-induced anorexia in either male or female mice (Fig. 4l and Extended Data Fig. 6b). We and others have previously demonstrated that the anorectic effects of 5-HT2CR-mediated signaling are largely mediated through POMC neurons7,23. Here we found that the F327L 5-HT2CR mutation reduced the baseline firing frequency of POMC neu- rons without altering baseline resting membrane potential (Extended Data Fig. 6c–e). Notably, fewer POMC neurons from Htr2cF327L/Y mice were depolarized by lorcaserin (20% in Htr2cF327L/Y mice versus 43% in Serotonin signaling is involved in coordinating defensive responses to threats in experimental animals24. We therefore examined the behavior of male Htr2cF327L/Y mice in a resident-intruder test. In this test, a male ‘intruder’ mouse is placed in the home cage of a ‘resident’ mouse (male wild-type or Htr2cF327L/Y) that has been individually housed. Htr2cF327L/Y mice showed reduced social exploration, whereas non-social explora- tion was not altered (Fig. 5a,b and Methods). Htr2cF327L/Y mice showed significantly increased defensive behaviors compared to wild-type lit- termates (Fig. 5c) and increased offensive attempts toward the intruder (89.5% in Htr2cF327L/Y mice versus 53.3% in wild-type mice, P = 0.0177, chi-squared test) with reduced latency (Fig. Htr2cF327L mice are less responsive to 5-HT2CR agonism 5d,e) although the average number of offensive episodes and time spent in these behaviors was Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2543 https://doi.org/10.1038/s41591-022-02106-5 Article Total = 15 53.33% ofense 46.67% no ofense WT 7 8 Total = 19 89.47% ofense 10.53% no ofense Htr2cF327L/Y 17 2 WT Htr2cF327L/Y 0 100 200 300 400 500 Latency (s) P = 0.006 P = 0.0177 in χ2 test d e WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y WT Htr2cF327L/Y 0 50 100 150 Duration (s) P = 0.001 0 100 200 300 400 Duration (s) 0 100 200 300 Duration (s) P = 0.034 Total = 15 53.33% ofense 46.67% no ofense WT 7 8 Total = 19 89.47% ofense 10.53% no ofense Htr2cF327L/Y 17 2 WT Htr2cF327L/Y 0 100 200 300 400 500 Latency (s) P = 0.006 0 5 10 15 20 Number of ofenses 0 50 100 150 Duration (s) P = 0.0177 in χ2 test a b c d e g h 0 5 10 15 20 Number P = 0.013 i 0 20 40 60 80 Time (s) P < 0.0001 j 0 5 10 15 20 25 Frequency P = 0.028 k 0 5 10 15 20 25 Number P = 0.013 l 0 100 200 300 400 Time (s) P = 0.014 m 0 10 20 30 40 50 Time (s) P = 0.031 n 0 0.5 1.0 1.5 Ratio WT Htr2cF327L/Y WT Htr2cF327L/Y 0 0.5 1.0 1.5 Ratio P = 0.035 f o Fig. 5 \| Knock-in male Htr2cF327L/Y mice develop social anxiety. a–c, Social exploration time (a), non-social exploration time (b) and defensive time (c) spent by male WT (n = 15) and Htr2cF327L/Y mice (n = 19) (4–5 months of age) in the resident-intruder test. d, Percentage/number of male WT and Htr2cF327L/Y mice that exhibited offensive behaviors toward the intruder. e,g, Offensive latency, number and time quantified in male WT (n = 15) and Htr2cF327L/Y (n = 19) mice in the intruder test. h, Ratio of time spent sniffing a mouse when interacting with a mouse compared to an object in the three-chamber social interaction test quantified in male WT (n = 7) and Htr2cF327L/Y (n = 8) mice. Htr2cF327L mice are less responsive to 5-HT2CR agonism Accordingly, we consider that the spectrum of behavioral changes seen in the mouse model of a human HTR2C mutation (reactive aggression, locomotor hyperactivity and anxiety) are modulated by the coexistence of cognitive deficits that impair the animals’ ability to learn how to respond to environmental challenges. Mendelian inheritance (LEP, LEPR, POMC and PCSK1), rare heterozygous variants in HTR2C are generally not fully penetrant. Our results align with findings reported for an increasing number of genes (KSR2 (ref. 28), MRAP2 (ref. 29) and SRC-1 (ref. 30)) where complete disruption in mice causes obesity but where the penetrance of heterozygous rare LOF variants in humans is more variable and likely to be modified by other genetic and environmental factors. There are a number of limitations associated with our human studies. Given the rarity of the variants identified here and the fact they do not display Mendelian inheritance, association studies in tens of thousands of individuals with severe obesity versus a comparable number of control individuals will be needed to find statistical evidence for enrichment31,32; however, to our knowledge, no clinical cohort of this size currently exists. Association studies in very large population-based cohorts such as the UK Biobank can be informative, but people carry- ing pathogenic variants in the gene encoding 5-HT2CR are likely to be under-represented in such cohorts33, as exemplified by studies of other genes involved in severe obesity, for example MC4R34,35. Further analysis is warranted in larger cohorts, but additional experimental evidence (for example using animal models of human disease) is often needed to establish the causal role of rare variants with variable penetrance. Animal studies also have limitations. For example, the laboratory set- ting where metabolic and behavioral phenotypes were characterized does not fully model the complex nutritional and social environment people live in. We observed that anxiety and maladaptive behavior are frequently reported in a large cohort of children with severe obesity, findings that suggest shared mechanistic origins for these disorders; however, as a medical history of these conditions is sometimes not obtained, or if obtained it is attributed to weight stigma or family circumstances, their true prevalence may be underestimated in this study. Further studies of large clinical cohorts in whom the prevalence of neurobehavioral Htr2cF327L mice are less responsive to 5-HT2CR agonism e,g, Offensive latency, number and time quantified in male WT (n = 15) and Htr2cF327L/Y (n = 19) mice in the intruder test. h, Ratio of time spent sniffing a mouse when interacting with a mouse compared to an object in the three-chamber social interaction test quantified in male WT (n = 7) and Htr2cF327L/Y (n = 8) mice. i, Ratio of time spent sniffing a novel mouse when interacting with a new mouse compared to a familiar mouse in the three-chamber social interaction test in male WT (n = 7) and Htr2cF327L/Y (n = 8) mice. l,m, Number of entries to (l) and time spent on (m) the open arms of the EPM apparatus by male WT (n = 8) and Htr2cF327L/Y (n = 8) mice. n,o, Number of (n) and time (o) spent in head dipping on the open arms of the EPM apparatus by male WT (n = 8) and Htr2cF327L/Y (n = 8) mice. P values were determined by two-tailed unpaired Student’s t-tests. Data are presented as mean ± s.e.m. with individual data points. exploration time (a), non-social exploration time (b) and defensive time (c) spent by male WT (n = 15) and Htr2cF327L/Y mice (n = 19) (4–5 months of age) in the resident-intruder test. d, Percentage/number of male WT and Htr2cF327L/Y mice that exhibited offensive behaviors toward the intruder. e,g, Offensive latency, number and time quantified in male WT (n = 15) and Htr2cF327L/Y (n = 19) mice in the intruder test. h, Ratio of time spent sniffing a mouse when interacting with a mouse compared to an object in the three-chamber social interaction test quantified in male WT (n = 7) and Htr2cF327L/Y (n = 8) mice. i, Ratio of time comparable between male Htr2cF327L/Y and wild-type mice (Fig. 5f,g). In a three-chamber social interaction test, compared to male wild-type mice, male Htr2cF327L/Y mice displayed reduced interaction with another mouse (Fig. 5h), demonstrating reduced sociability. Notably, both groups showed comparable preference to interact with a novel mouse versus a familiar mouse (Fig. 5i), excluding neophobia as a potential confounder in this test. The elevated plus-maze (EPM) test is classically used to assess anxiety-related behavior and screen anxiolytic com- pounds in rodents25. In this test, male Htr2cF327L/Y mice paradoxically showed reduced risk assessment behavior before they entered the open arms (Fig. 5j,k) and spent more time in the open arms (Fig. Htr2cF327L mice are less responsive to 5-HT2CR agonism i, Ratio of time spent sniffing a novel mouse when interacting with a new mouse compared to a familiar mouse in the three-chamber social interaction test in male WT (n = 7) and Htr2cF327L/Y (n = 8) mice. l,m, Number of entries to (l) and time spent on (m) the open arms of the EPM apparatus by male WT (n = 8) and Htr2cF327L/Y (n = 8) mice. n,o, Number of (n) and time (o) spent in head dipping on the open arms of the EPM apparatus by male WT (n = 8) and Htr2cF327L/Y (n = 8) mice. P values were determined by two-tailed unpaired Student’s t-tests. Data are presented as mean ± s.e.m. with individual data points. WT Htr2cF327L/Y 0 50 100 150 Duration (s) P = 0.001 a WT Htr2cF327L/Y 0 100 200 300 Duration (s) P = 0.034 c WT Htr2cF327L/Y 0 100 200 300 400 Duration (s) b e WT Htr2cF327L/Y 0 100 200 300 400 500 Latency (s) P = 0.006 e b d c a WT Htr2cF327L/Y 0 5 10 15 20 Number of ofenses f WT Htr2cF327L/Y 0 50 100 150 Duration (s) g WT Htr2cF327L/Y h 0 0.5 1.0 1.5 Ratio P = 0.035 i 1 i 0 5 10 15 20 Number P = 0.013 j WT Htr2cF327L/Y h i WT Htr2cF327L/Y 0 0.5 1.0 1.5 Ratio 2 j f g WT Htr2cF327L/Y 0 5 10 15 20 25 Frequency P = 0.028 l WT Htr2cF327L/Y 0 20 40 60 80 Time (s) P < 0.0001 k WT Htr2cF327L/Y 0 5 10 15 20 25 Number P = 0.013 n WT Htr2cF327L/Y 0 100 200 300 400 Time (s) P = 0.014 m l WT Htr2cF327L/Y 0 10 20 30 40 50 Time (s) P = 0.031 o k n o n o m Htr2cF327L/Y WT WT Fig. 5 \| Knock-in male Htr2cF327L/Y mice develop social anxiety. a–c, Social Fig. 5 \| Knock-in male Htr2cF327L/Y mice develop social anxiety. a–c, Social exploration time (a), non-social exploration time (b) and defensive time (c) spent by male WT (n = 15) and Htr2cF327L/Y mice (n = 19) (4–5 months of age) in the resident-intruder test. d, Percentage/number of male WT and Htr2cF327L/Y mice that exhibited offensive behaviors toward the intruder. Htr2cF327L mice are less responsive to 5-HT2CR agonism 5l,m), whereas on the open arms, Htr2cF327L/Y mice showed significantly increased head dipping (Fig. 5n,o), an exploratory behavior that can be induced by antagonists of 5-HT2CR26. Female Htr2cF327L/+ mice partially recapitulated the behaviors seen in male Htr2cF327L/Y mice, including increased offensive behavior in the resident-intruder test, increased entry into the open arms and increased number of head dipping epi- sodes on the EPM apparatus, whereas social behavior and risk assess- ment behavior were not altered (Extended Data Fig. 7). Previous studies have shown that Htr2c knockout mice exhibit impaired performance in the Morris water maze test of spatial learning, with reduced aversion to a novel environment27. Accordingly, we consider that the spectrum of behavioral changes seen in the mouse model of a human HTR2C mutation (reactive aggression, locomotor hyperactivity and anxiety) are modulated by the coexistence of cognitive deficits that impair the animals’ ability to learn how to respond to environmental challenges. comparable between male Htr2cF327L/Y and wild-type mice (Fig. 5f,g). In a three-chamber social interaction test, compared to male wild-type mice, male Htr2cF327L/Y mice displayed reduced interaction with another mouse (Fig. 5h), demonstrating reduced sociability. Notably, both groups showed comparable preference to interact with a novel mouse versus a familiar mouse (Fig. 5i), excluding neophobia as a potential confounder in this test. The elevated plus-maze (EPM) test is classically used to assess anxiety-related behavior and screen anxiolytic com- pounds in rodents25. In this test, male Htr2cF327L/Y mice paradoxically showed reduced risk assessment behavior before they entered the open arms (Fig. 5j,k) and spent more time in the open arms (Fig. 5l,m), whereas on the open arms, Htr2cF327L/Y mice showed significantly increased head dipping (Fig. 5n,o), an exploratory behavior that can be induced by antagonists of 5-HT2CR26. Female Htr2cF327L/+ mice partially recapitulated the behaviors seen in male Htr2cF327L/Y mice, including increased offensive behavior in the resident-intruder test, increased entry into the open arms and increased number of head dipping epi- sodes on the EPM apparatus, whereas social behavior and risk assess- ment behavior were not altered (Extended Data Fig. 7). Previous studies have shown that Htr2c knockout mice exhibit impaired performance in the Morris water maze test of spatial learning, with reduced aversion to a novel environment27. Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 Article Article in women with obesity47. To date, the assumption has been that this association may reflect the widespread stigma and discrimination people with obesity face and/or early life experiences which influence susceptibility to obesity and anxiety; however, our findings suggest another possible explanation. In some people (particularly women), impaired serotonin signaling may directly contribute to both their obesity and their social anxiety/maladaptive behavior. Other serotonin receptors may also contribute to human neuropsychiatric behaviors. Goldman and colleagues48 identified an LOF mutation in 5-HT2BR in an isolated Finnish extended family with a strong history of aggressive behavior and impulsivity. In line with these clinical observations, they showed that mice lacking 5-HT2BR exhibited impulsivity and aggressive behavior in a resident-intruder paradigm. While LOF variants in the gene encoding 5-HT2CR are rare, other genes and/or mechanisms which perturb signaling through 5-HT2CR may contribute to obesity and anxi- ety in some people. Indeed, selective serotonin reuptake inhibitors are effective in reducing symptoms of anxiety and specifically are first-line treatment for panic disorder, obsessive-compulsive disorder and social anxiety disorder. Our findings suggest that enhanced signaling through 5-HT2CR could explain the clinical effectiveness of these drugs in the treatment of a range of disorders characterized by anxiety. Further studies are needed to identify additional regulators of 5-HT2CR expres- sion and signaling, which may further add to our understanding of the association between obesity and anxiety. Targeting the mechanisms that regulate both canonical and β-arrestin-mediated signaling by 5-HT2CR could inform the development of medications that are both effective and safe for chronic weight management. conditions is measured using standardized clinical questionnaires are needed. Our findings indicate that a careful history of anxiety, mood disturbances and maladaptive behavior should be obtained in people presenting with severe early-onset obesity as positive clinical findings may guide genetic testing and have relevance for treatment. In keeping with studies in mice which showed that the anorectic effects of 5-HT are mediated by 5-HT2CRs expressed on hypothalamic POMC neurons6,7, we show that knock-in mice carrying F327L 5-HT2CR have impaired depolarization of POMC neurons when stimulated by the serotoninergic agonist, lorcaserin. As such, severe obesity in some people with functional variants in HTR2C is likely to be driven by impaired melanocortin signaling. References O 1. Marston, O. J., Garfield, A. S. & Heisler, L. K. Role of central serotonin and melanocortin systems in the control of energy balance. Eur. J. Pharmacol. 660, 70–79 (2011). 1. Marston, O. J., Garfield, A. S. & Heisler, L. K. Role of central serotonin and melanocortin systems in the control of energy balance. Eur. J. Pharmacol. 660, 70–79 (2011). 2. Donovan, M. H. & Tecott, L. H. Serotonin and the regulation of mammalian energy balance. Front. Neurosci. 7, 36 (2013). 3 L D D G fi ld A S M O J Sh J & H i l L K 2. Donovan, M. H. & Tecott, L. H. Serotonin and the regulation of mammalian energy balance. Front. Neurosci. 7, 36 (2013). 3. Lam, D. D., Garfield, A. S., Marston, O. J., Shaw, J. & Heisler, L. K. Brain serotonin system in the coordination of food intake and body weight. Pharmacol. Biochem. Behav. 97, 84–91 (2010). 3. Lam, D. D., Garfield, A. S., Marston, O. J., Shaw, J. & Heisler, L. K. Brain serotonin system in the coordination of food intake and body weight. Pharmacol. Biochem. Behav. 97, 84–91 (2010). 4. Alvarez-Jimenez, M. et al. Antipsychotic-induced weight gain in chronic and first-episode psychotic disorders: a systematic critical reappraisal. CNS Drugs 22, 547–562 (2008). 4. Alvarez-Jimenez, M. et al. Antipsychotic-induced weight gain in chronic and first-episode psychotic disorders: a systematic critical reappraisal. CNS Drugs 22, 547–562 (2008). 5. Xu, Y. et al. 5-HT2CRs expressed by pro-opiomelanocortin neurons regulate insulin sensitivity in liver. Nat. Neurosci. 13, 1457–1459 (2010). 5. Xu, Y. et al. 5-HT2CRs expressed by pro-opiomelanocortin neurons regulate insulin sensitivity in liver. Nat. Neurosci. 13, 1457–1459 (2010). 5-HT2CRs are expressed along ascending dopaminergic pathways and their activation in regions such as the nucleus accumbens regulates inhibitory control over dopamine release39,40 suggesting a plausible mechanism underlying hyperactivity41. In a series of behavioral studies, we found that male hemizygous Htr2cF327L/Y and female heterozygous Htr2cF327L/+ mice exhibited decreased social exploration and increased aggression toward an intruder. 5-HT plays an evolutionarily conserved role in the regulation of social and emotional behaviors in many spe- cies;24,39,42–44 mice given selective serotonin reuptake inhibitors that enhance central 5-HT levels, show reduced aggression45. Neural circuits in the amygdala (a prominent site of 5-HT2CR expression) play a critical role in linking fear processing and defensive responses (freeze/flight)43,46, which enable a rapid response to a natural predator, promoting survival. 6. Article These findings have clinical relevance as a melanocortin 4 receptor agonist, setmelanotide, causes weight loss in people with impaired melanocortin signaling due to POMC, PCSK1 and LEPR mutations9–11 and is being trialed (ClinicalTrials.gov identifier NCT04963231) in people with rare variants in other genes that regulate the depolarization of POMC neurons and the transcription of POMC (Fig. 2). Clinically meaningful weight loss in people carrying HTR2C variants receiving an MC4R agonist may provide biological validation of the pathogenicity of specific variants and inform the inclusion of HTR2C in diagnostic gene panels for severe childhood-onset obesity36. g g p y In our studies, knock-in mutant mice displayed increased food intake and locomotor hyperactivity as seen in knockout mice;37,38 there was minimal difference in energy expenditure. The metabolic and behavioral phenotype of knock-in mice was more marked in hemizy- gous males (minimal functioning 5-HT2CRs due to the presence of the mutation on their single X chromosome) compared to heterozygous female knock-in mice (one wild-type X chromosome and one mutant X chromosome). Modest differences in phenotype are likely to be explained by a gene-dosage effect, although other explanations for sex-specific differences cannot be excluded. It is noteworthy that 16 of 19 probands in whom HTR2C variants were identified in our study were female, whereas 50% of the GOOS cohort who were screened were female. Mutations in genes on the X chromosome have different consequences in males and females. To date, more than 100 X-linked inherited diseases have been identified; the majority are recessively inherited, a much smaller number are dominantly inherited and a few are dominant and lethal in hemizygous males. The excess of affected female heterozygotes compared to male heterozygotes in this study is suggestive of X-linked dominant inheritance. X-linked dominant disorders that are lethal in males in utero are, by definition, seen only in female heterozygotes, with affected (hemizygous) males manifesting as an excess of spontaneous miscarriages in affected families as seen in Rett syndrome. In view of the more severe metabolic and behavioral phenotype seen in male versus female mice, the low prevalence of males carrying HTR2C variants seen in this study might be explained by excess mortality in affected males. Further studies to establish the frequency of miscarriage in these families will be needed to explore this further. Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 Online content Any methods, additional references, Nature Portfolio reporting sum- maries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author con- tributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41591-022-02106-5. Discussion We provide evidence for a functional role of rare variants in HTR2C (encoding 5-HT2CR) in the development of obesity in humans. These findings have both diagnostic and therapeutic implications. In con- trast to some monogenic obesity syndromes, which display classical Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2544 https://doi.org/10.1038/s41591-022-02106-5 References O 5-HT2C receptor structures reveal the structural basis of GPCR polypharmacology. Cell 172, 719–730 (2018). 44. Kulikov, A. V., Osipova, D. V., Naumenko, V. S. & Popova, N. K. The C1473G polymorphism in the tryptophan hydroxylase-2 gene and intermale aggression in mice. Dokl. Biol. Sci. 402, 208–210 (2005). 22. Szustakowski, J. D. et al. Advancing human genetics research and drug discovery through exome sequencing of the UK Biobank. Nat. Genet. 53, 942–948 (2021). 45. Sanchez, C. & Meier, E. Behavioral profiles of SSRIs in animal models of depression, anxiety and aggression. Are they all alike? Psychopharmacology 129, 197–205 (1997). 23. Xu, Y. et al. 5-HT2CRs expressed by pro-opiomelanocortin neurons regulate energy homeostasis. Neuron 60, 582–589 (2008). 46. Marcinkiewcz, C. A. et al. Serotonin engages an anxiety and fear-promoting circuit in the extended amygdala. Nature 537, 97–101 (2016). 24. Haller, J., Mikics, E., Halasz, J. & Toth, M. Mechanisms differentiating normal from abnormal aggression: glucocorticoids and serotonin. Eur. J. Pharmacol. 526, 89–100 (2005). 47. Petry, N. M., Barry, D., Pietrzak, R. H. & Wagner, J. A. Overweight and obesity are associated with psychiatric disorders: results from the national epidemiologic survey on alcohol and related conditions. Psychosom. Med. 70, 288–297 (2008). 25. Walf, A. A. & Frye, C. A. The use of the elevated plus maze as an assay of anxiety-related behavior in rodents. Nat. Protoc. 2, 322–328 (2007). 48. Bevilacqua, L. et al. A population-specific HTR2B stop codon predisposes to severe impulsivity. Nature 468, 1061–1066 (2010). 26. Griebel, G., Rodgers, R. J., Perrault, G. & Sanger, D. J. Risk assessment behaviour: evaluation of utility in the study of 5-HT-related drugs in the rat elevated plus-maze test. Pharmacol. Biochem. Behav. 57, 817–827 (1997). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 27. Tecott, L. H., Logue, S. F., Wehner, J. M. & Kauer, J. A. Perturbed dentate gyrus function in serotonin 5-HT2C receptor mutant mice. Proc. Natl Acad. Sci. USA 95, 15026–15031 (1998). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. References O 27, 1088–1096 (2021). 36. Styne, D. M. et al. Pediatric obesity-assessment, treatment, and prevention: an endocrine society clinical practice guideline. J. Clin. Endocrinol. Metab. 102, 709–757 (2017). J. Clin. Endocrinol. Metab. 102, 709–757 (2017). 14. Doche, M. E. et al. Human SH2B1 mutations are associated with maladaptive behaviors and obesity. J. Clin. Invest. 122, 4732–4736 (2012). 37. Tecott, L. H. et al. Eating disorder and epilepsy in mice lacking 5-HT2c serotonin receptors. Nature 374, 542–546 (1995). 15. Jiang, L. et al. Neural deletion of Sh2b1 results in brain growth retardation and reactive aggression. FASEB J. 32, 1830–1840 (2018). 38. Nonogaki, K., Abdallah, L., Goulding, E. H., Bonasera, S. J. & Tecott, L. H. Hyperactivity and reduced energy cost of physical activity in serotonin 5-HT(2C) receptor mutant mice. Diabetes 52, 315–320 (2003). 16. Labasque, M., Reiter, E., Becamel, C., Bockaert, J. & Marin, P. Physical interaction of calmodulin with the 5-hydroxytryptamine2C receptor C-terminus is essential for G protein-independent, arrestin-dependent receptor signaling. Mol. Biol. Cell 19, 4640–4650 (2008).f 39. Ferrari, P. F., van Erp, A. M., Tornatzky, W. & Miczek, K. A. Accumbal dopamine and serotonin in anticipation of the next aggressive episode in rats. Eur. J. Neurosci. 17, 371–378 (2003). 17. He, Y. et al. Barbadin potentiates long-term effects of lorcaserin on POMC neurons and weight loss. J. Neurosci. 41, 5734–5746 (2021). 40. Abdallah, L. et al. Impact of serotonin 2C receptor null mutation on physiology and behavior associated with nigrostriatal dopamine pathway function. J. Neurosci. 29, 8156–8165 (2009). 41. Bonasera, S. J. et al. Mice lacking serotonin 2C receptors have increased affective responses to aversive stimuli. PLoS ONE 10, e0142906 (2015). 18. Burns, C. M. et al. Regulation of serotonin-2C receptor G-protein coupling by RNA editing. Nature 387, 303–308 (1997). 19. Gumpper, R. H., Fay, J. F. & Roth, B. L. Molecular insights into the regulation of constitutive activity by RNA editing of 5HT2C serotonin receptors. Cell Rep. 40, 111211 (2022). 42. Popova, N. K. From genes to aggressive behavior: the role of serotonergic system. Bioessays 28, 495–503 (2006). 20. Wang, Q. et al. Altered G protein-coupling functions of RNA editing isoform and splicing variant serotonin2C receptors. J. Neurochem. 74, 1290–1300 (2000). 43. Heisler, L. K., Zhou, L., Bajwa, P., Hsu, J. & Tecott, L. H. Serotonin 5-HT(2C) receptors regulate anxiety-like behavior. Genes Brain Behav. 6, 491–496 (2007). 21. Peng, Y. et al. References O Berglund, E. D. et al. Serotonin 2C receptors in pro-opiomelanocortin neurons regulate energy and glucose homeostasis. J. Clin. Invest. 123, 5061–5070 (2013). 6. Berglund, E. D. et al. Serotonin 2C receptors in pro-opiomelanocortin neurons regulate energy and glucose homeostasis. J. Clin. Invest. 123, 5061–5070 (2013). 7. Heisler, L. K. et al. Central serotonin and melanocortin pathways regulating energy homeostasis. Ann. NY Acad. Sci. 994, 169–174 (2003).f 7. Heisler, L. K. et al. Central serotonin and melanocortin pathways regulating energy homeostasis. Ann. NY Acad. Sci. 994, 169–174 (2003).f 8. Krude, H. et al. Severe early-onset obesity, adrenal insufficiency and red hair pigmentation caused by POMC mutations in humans. Nat. Genet. 19, 155–157 (1998). 8. Krude, H. et al. Severe early-onset obesity, adrenal insufficiency and red hair pigmentation caused by POMC mutations in humans. Nat. Genet. 19, 155–157 (1998). 9. Kuhnen, P. et al. Proopiomelanocortin deficiency treated with a melanocortin-4 receptor agonist. N. Engl. J. Med. 375, 240–246 (2016). Clinically, anxiety disorders (including social anxiety) are gener- ally considered an extreme manifestation of the anxiety response to perceived threats. Several epidemiological studies have found that anx- iety is a common feature in people living with obesity and particularly 10. Clement, K. et al. MC4R agonism promotes durable weight loss in patients with leptin receptor deficiency. Nat. Med. 24, 551–555 (2018). 2545 Article https://doi.org/10.1038/s41591-022-02106-5 32. Agarwala, V., Flannick, J., Sunyaev, S., Go, T. D. C. & Altshuler, D. Evaluating empirical bounds on complex disease genetic architecture. Nat. Genet. 45, 1418–1427 (2013). 11. Clement, K. et al. Efficacy and safety of setmelanotide, an MC4R agonist, in individuals with severe obesity due to LEPR or POMC deficiency: single-arm, open-label, multicentre, phase 3 trials. Lancet Diabetes Endocrinol. 8, 960–970 (2020). 33. Tyrrell, J. et al. Genetic predictors of participation in optional components of UK Biobank. Nat. Commun. 12, 886 (2021). 12. Hendricks, A. E. et al. Rare variant analysis of human and rodent obesity genes in individuals with severe childhood obesity. Sci. Rep. 7, 4394 (2017). 34. Khera, A. V. et al. Polygenic prediction of weight and obesity trajectories from birth to adulthood. Cell 177, 587–596 (2019). 13. Marenne, G. et al. Exome sequencing identifies genes and gene sets contributing to severe childhood obesity, linking PHIP Variants to repressed POMC transcription. Cell Metab. 31, 1107–1119 (2020). 35. Wade, K. H. et al. Loss-of-function mutations in the melanocortin 4 receptor in a UK birth cohort. Nat. Med. Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 References O The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 28. Pearce, L. R. et al. KSR2 mutations are associated with obesity, insulin resistance, and impaired cellular fuel oxidation. Cell 155, 765–777 (2013). 29. Asai, M. et al. Loss of function of the melanocortin 2 receptor accessory protein 2 is associated with mammalian obesity. Science 341, 275–278 (2013). 30. Yang, Y. et al. Steroid receptor coactivator-1 modulates the function of POMC neurons and energy homeostasis. Nat. Commun. 10, 1718 (2019). 31. Zuk, O. et al. Searching for missing heritability: designing rare variant association studies. Proc. Natl Acad. Sci. USA 111, E455–E464 (2014). © The Author(s) 2022 Nature Medicine \| Volume 28 \| December 2022 \| 2537–2546 2546 https://doi.org/10.1038/s41591-022-02106-5 Article Methods Human studies diagnoses in 16 categories (https://biobank.ctsu.ox.ac.uk/showcase/ field.cgi?id=20544); limitations include lack of complete data (returned questionnaires), ascertainment bias and self-reporting of professional diagnoses. Of 16 categories of mental distress, we selected three categories of interest: ‘psychological over-eating or binge-eating’, ‘anxiety, nerves or generalized anxiety disorder’ (anxiety) and ‘depres- sion’, and for selected variants we further inspected the reporting of any category of mental distress in people who carried variants and those who did not. These studies were approved by the Multi-regional Ethics Committee and Cambridge Local Research Ethics Committee (03/103, 03/104 and 18/EE/0032) and conducted in accordance with the principles of the Declaration of Helsinki. Each participant or their legal guardian (for children under 16 years of age) provided written, informed consent and minors provided verbal or written consent. We studied people with severe obesity (BMI (kg m−2) SDS > 3) of early-onset (<10 years of age) recruited to the GOOS cohort (www.goos. org.uk) between 1997 and 2022. Referring physicians completed stand- ardized questionnaires that captured medical history, family history and findings on clinical examination, including anthropometric data and the results of investigations. The sex of participants was assigned by clinical examination by physicians. To describe the prevalence of neurobehavioral conditions in the GOOS cohort, we excluded people in whom genetic obesity syndromes had been identified. As a result data on 7,775 people were included. Exome sequencing and targeted resequencing were performed in a subset of the cohort as previously described;12,13 all variants referred to in this study were verified by Sanger sequencing. Single-variant association analysis with BMI, obesity and severe obesity Association analysis of single variants with allele count >20 was per- formed separately for females and males among unrelated European exomes (plink v.2.00a1LM; www.cog-genomics.org/plink/2.0/). Case– control association with severe obesity (BMI > 40 kg m−2) or obesity (BMI > 30 kg m−2) was performed using plink2 –glm with Firth regres- sion; n = 2,725 individuals (1,868 females and 857 males) were severely obese; n = 36,304 individuals (19,018 females and 17,286 males) were obese. Association with BMI as a continuous trait was performed using plink2 –glm. Covariates were age (Field 21003) and 40 genetic principal components (Fields 22009.0.1–40). At very rare variants (allele count <20), ORs were calculated for the number of variant carriers using Fisher’s exact test. Probands and their families were invited to participate in physi- ological studies at the Wellcome-MRC Institute of Metabolic Science Translational Research Facility, Addenbrooke’s Hospital. Weight and height were measured barefoot in light clothing. Dual X-ray absorptiom- etry (DPX software; Lunar Corp) was used to determine body composi- tion. BMR was determined by indirect calorimetry after a 10-h overnight fast using an open circuit, ventilated, canopy measurement system (Europa Gas Exchange Monitor; NutrEn Technology). BMR adjusted for body composition was compared to predicted metabolic rate based on standard age and sex-specific equations. Blood pressure was measured in the rested fasted state using wrist monitors (OMRON Healthcare). For the analysis of insulin and glucose, patients with type 2 diabetes were excluded. Blood samples were taken in the fasted state for all assays. cDNA constructs and site-directed mutagenesis g The human WT HTR2C complementary DNA (Ensembl transcript ENST00000371951; unedited, INI isoform) was cloned into the pcDNA3.1(+) vector (Invitrogen) and a sequence encoding a FLAG tag was inserted downstream of the sequence encoding the receptor’s signal peptide using a Q5 site-directed mutagenesis kit (New England Biolabs). The edited VSV isoform was generated by changing aden- osines at positions A, B, C and D to guanosines20, using the QuickChange II XL site-directed mutagenesis kit (Agilent). For β-arrestin protein–pro- tein interaction assays, the receptor VSV isoform lacking a FLAG tag was cloned into the pBiT1.1-C (TK/LgBiT) vector and cDNAs encoding human β-arrestin-1 and -2 were ligated in pBiT2.1-N (TK/SmBiT) vectors (Promega) as described before49. Variant constructs of both backbones were generated using QuickChange II XL site-directed mutagenesis kit (Agilent). All constructs were verified using Sanger sequencing. In females, we used genotype calls provided in the pVCF file. In males, we treated any non-homozygous-REF genotype as hemizygous, consistent with reported allelic depths. Relatedness was obtained from the UK Biobank (ukbgene rel) and one person was excluded from each related pair among all the OQFE exomes (kinship ≥ 0.0442, KING, third-degree kinship or closer; retained pairs contained within OQFE exomes and excluded individuals in column ‘ID2’; n = 15,547 people). Exomes were restricted European genetic ethnic grouping (Field 22006, self-reported ‘White British’ and tight cluster in genotype principal-component analysis). Unrelated European OQFE exomes were taken forward for analysis (n = 153,352). We obtained BMI (kg m−2) from the UK Biobank initial assessment visit (Field 21001, Instance 0), available to us for n = 152,837 of 153,352 unrelated European exomes (83,864 females and 69,488 males). To investigate phenotypes related to anxiety or mood, we used the mental distress (‘Mental health problems ever diagnosed by a professional’, Field 20544) section of online follow-up Mental Health questionnaire (category 136) (rationale and design is provided at https://biobank. ctsu.ox.ac.uk/showcase/refer.cgi?id=22), which had an available ‘Date of completing’ (Field 20400) for n = 53,174 of 153,352 unrelated Euro- pean exomes. Mental distress data provide self-reported professional Rare variant association tests We performed gene-based association tests separately for males and females using non-synonymous exonic or splicing variants and severe obesity (BMI > 40 kg m−2), BMI > 30 kg m−2 and for continuous BMI. Gene-based burden and SKATO tests were performed using the SKAT- Binary_Robust function for dichotomized BMI or the SKAT function for continuous BMI, in R package SKAT v.2.0.1 (method = ‘Burden’ or ‘SKATO’ with default settings). Null models were calculated using SKAT_Null_Model(y~X) where covariates matrix (X) contained age (Field 21003.0.0), sex (Field 31.0.0), ten genetic principal components (Fields 22009.0.1–10) and sequencing batch (UKB 50K or 150K exomes). Single-variant P values reported from gene-based analyses are SKAT- Binary_Robust output values for p.value_singlevariant. ORs were also calculated for the number of variant carriers using Fisher’s exact test. UK Biobank 200K OQFE exomes and clinical phenotypes This research was conducted using the UK Biobank Resource22 under application no. 53821. We used pVCF variant file (chrX, block 17) from OQFE exome pipeline (UK Biobank Field 23156; n = 200,629 exomes available to us). Reported sex was obtained from Field 31 (Supple- mentary Table 2). We split and left-normalized multiallelic entries (bcftools v.1.9) and defined variant consequences with respect to Ensembl canonical transcript ENST00000276198 using Ensembl Vari- ant Effect Predictor (Ensembl release v.102). Nature Medicine Cell surface expression ELISA HEK293 cells seeded onto poly-d-lysine coated 96-well plates (40,000 cells per well) were transfected with 5 ng per well of FLAG-tagged HTR2C constructs as described above. The day after transfection, the cells were fixed with 3.7% paraformaldehyde in PBS for 15 min at room tem- perature and washed three times with PBS. Subsequently, nonspecific binding sites were blocked with 3% non-fat dry milk in 50 mM Tris-PBS pH 7.4 (blocking buffer) for 1 h at room temperature. Cells were incu- bated with a mouse monoclonal anti-FLAG antibody (Sigma-Aldrich, F1804), diluted 1,000× in blocking buffer overnight at 4 °C. Next, cells were washed three times with PBS and incubated with goat anti-mouse IgG (H + L)-HRP conjugate (Bio-Rad Laboratories, 172-1011) (1:1,250 dilution in 1.5% non-fat dry milk in 50 mM Tris-PBS) for 2 h at room temperature. Finally, cells were washed three times with PBS and the β-arrestin protein–protein interaction assay Coupling between 5-HT2CR and β-arrestin-1 or β-arrestin-2 was quan- tified using the NanoBiT protein–protein interaction assay (Pro- mega, M2014). Assays were performed in HEK293SL cells seeded in poly-d-lysine coated, white 96-well plates (15,000 cells per well) and transiently transfected with 50 ng per well of each of the two constructs as described above. As a negative control, the β-arrestin-SmBiT con- structs were substituted with the HaloTag-SmBiT negative control vector. The day after transfection and half an hour before assay, cul- ture medium was substituted for 70 μl per well serum-free Opti-MEM (Gibco, 31985). Nanoluciferase activity was measured at 37 °C and in the presence of 5% CO2 using a Spark 10 M microplate reader (Tecan). After measurement of the background signal, 20 μl per well Nano-Glo Live Cell Assay System (Promega, N2013) was added and basal luciferase activity was measured for 10 min in 30-s intervals. Subsequently, cells were stimulated with 10 μl of 10 × 5-HT stock solution and the lumines- cent signal was quantified for 20 min in 30-s intervals. Dose–response curves (sigmoidal with variable slope) were plotted from total peak area under the curve values calculated from each 5-HT concentration, ranging from 0 to 10−5 M. Normalized data from independent experi- ments were merged and presented as sum curves. Results are from 4–5 independent experiments. Cell culture and transfection HEK293 or HEK293SL cells were maintained in high glucose Dulbecco’s modified eagle medium (Gibco, 31966), supplemented with 10% fetal bovine serum (Gibco, 10270, South America origin), 1% GlutaMAX (100×) (Gibco, 35050) and 100 U ml−1 penicillin and 100 μg ml−1 strep- tomycin (Sigma-Aldrich, P0781) and cultured at 37 °C in humidified air containing 5% CO2. Cells were transfected using Lipofectamine 2000 (Gibco, 11668) in serum-free Opti-MEM (Gibco, 31985) according to the manufacturer’s protocol. Nature Medicine https://doi.org/10.1038/s41591-022-02106-5 Article Inositol triphosphate turnover assay chromogenic substrate 3,3′,5,5′-tetramethylbenzidine (TMB CORE+, Bio-Rad Laboratories, BUF062) was used to detect HRP activity. The reaction was stopped with 0.5 M H2SO4 and absorbance at 450 nm was quantified using a Tecan Infinite M1000 PRO microplate reader. p p y HEK293 cells were seeded in 96-well plates at a density of 40,000 cells per well and transiently transfected on the next day with 5 ng per well WT HTR2C or variant cDNA construct. After transfection, cells were cultured overnight in growth medium supplemented with 5 μl per well [3H]-myo-inositol (PerkinElmer, NET115600). Cells were washed once with Hank’s balanced salt solution (HBSS; Gibco, 14025) and subsequently stimulated with 20× 5-hydroxytryptamine stock solution (5-HT; Sigma-Aldrich, 85036) in HBSS containing 10 mM LiCl (Sigma-Aldrich, L9650) for 90 min at 37 °C. Cells were stimulated with different concentrations of 5-HT, ranging from 0 to 10−5 M. After aspi- rating the stimulation buffer, cells were lysed on ice for 30 min using 50 μl per well 10 mM formic acid (Sigma-Aldrich, F0507). Then 20 μl of lysate was transferred to a white 96-well plate containing 80 μl per well 12.5 mg ml−1 yttrium silicate poly-lysine-coated scintillation proximity beads (PerkinElmer, RPNQ0010) in ddH2O. Plates were sealed, shaken vigorously for 5 min and the relative amount of radiolabeled inositol monophosphate (IP1) was quantified after 8 h of settle time using a Top- Count 9012 Microplate Counter (Packard). Results were analyzed using GraphPad Prism 8 (GraphPad Software). Sigmoidal dose–response curves with variable slope (four-parameter logistic regression) were plotted. Results from each assay were normalized to mean counts from the bottom of the 5-HT2CR WT curve set as basal and the top value of the 5-HT2CR WT curve (maximal efficacy; Emax) as 100%. Cells transfected with pcDNA3.1(+) (mock) were used as a negative control. Normalized data for each experiment were merged and presented as sum curves, results are from 3–4 independent experiments. Immunofluorescence HEK293SL cells (10,000 cells per well) were seeded onto poly-d-lysine coated CellCarrier-96 Ultra plates (PerkinElmer, 6055302) and trans- fected on the following day with 40 ng per well FLAG-tagged constructs as described above. The next day, the cells were fixed with 3.7 % para- formaldehyde in PBS for 15 min and washed twice with PBS. Nonspecific antibody binding sites were blocked with blocking buffer (3 % BSA in PBS) for 1 h. The cells were then incubated with primary antibody (mouse monoclonal anti-FLAG; Sigma-Aldrich, F1804; 1:100 dilution in blocking buffer) for 1 h. After three washing steps with PBS, cells were incubated with secondary antibody (Alexa Fluor 488 goat anti-mouse IgG (H + L); Life Technologies A11029; 1:400 in blocking buffer) for 1 h in the dark. Plates were washed thrice with PBS and cells were per- meabilized with 0.1 % Triton X-100 (VWR, 306324 N) in PBS for 5 min. In parallel experiments to capture intracellular receptors, cells were permeabilized before adding the primary anti-FLAG antibody together with an anti-calreticulin antibody, to visualize the ER (Invitrogen, PA3- 900, 1:100 dilution), which was recognized using a Alexa Fluor 647 donkey anti-rabbit IgG (H + L) (Thermo Fisher Scientific, A-31573) at a 1:400 dilution in blocking buffer. Next, cell nuclei were stained with DAPI (Merck, D9542; 0.1 μg ml−1 final concentration in PBS) for 10 min, followed by three washing steps with PBS. Finally, cells were incubated with DyLight 554 Phalloidin (Cell Signaling Technology, 13054; 1:200 dilution in PBS) for 15 min to stain cytoskeletal F-actin. Cells were washed twice with PBS and stored in the dark at 4 °C before imaging on an Opera Phenix High-Content Screening System (PerkinElmer). For high-content analysis of WT and mutant protein levels at the cell surface and in the cytoplasm, the PerkinElmer Harmony software was used (HH17000012, version 5.0); the DyLight Phalloidin and DAPI channels were used to identify cells and the signal strength of the 488 nm channel was calculated per cell. For ER localization, the overlap between the calreticulin and the 488 nm channel was quantified. Each condition was performed in four technical replicates (on average 400 cells per well) and the mean intensity of all cells per well was calculated. Results shown are from three independent experiments. Radioligand binding assay g g y HEK293 cells were seeded in solid white 96-well plates coated with poly-d-lysine and transiently transfected with 50 ng per well FLAG-tagged HTR2C constructs. The day after transfection, cells were washed once and incubated on ice with binding buffer (50 μl per well; 20 mM HEPES pH 7.4, 118 mM NaCl, 4.7 mM KCl, 5 mM MgCl2, 5.5 mM d-glucose and 0.1% BSA) for 30 min. Next, varying doses of unlabeled 5-HT (0 M to 10−4 M) were added to the cells, immediately followed by 50 μl per well of the 3H-labeled 5-HT2CR antagonist mesulergine (PerkinElmer, NET1148; 1:5,000 dilution in binding buffer) and cells were incubated on ice for 3 h. After washing twice with ice-cold bind- ing buffer, 20 μl per well 0.1 M NaOH was dispensed and plates were shaken, followed by adding 80 μl per well MicroScint-20 scintillation fluid (PerkinElmer, 6013621). Plates were shaken and activity of bound 3H-mesulergine was quantified after a 3 h settle time, using a TopCount 9012 Microplate Counter (Packard). Approval for studies in mice Care of animals and procedures were approved by the Baylor College of Medicine Institutional Animal Care and Use Committee. Generation of Htr2cF327L mice The F327 amino acid residue in the human 5-HT2CR protein is equivalent to the F328 of the mouse 5-HT2CR protein. We generated a knock-in Nature Medicine Article https://doi.org/10.1038/s41591-022-02106-5 https://doi.org/10.1038/s41591-022-02106-5 mouse with the F328L mutation but referred it as Htr2cF327L for simplicity. A single-guide RNA (sgRNA) sequence (5′-TTTCATCACCAATATCCTGT) was selected to target the mouse Htr2c gene encoding the F328 and the sgRNA was purchased from Synthego. Cas9 Nuclease was purchased from IDT (Alt-R S.p. Cas9 Nuclease V3). The donor ssDNA template to introduce the F328L point mutation and a silent mutation C326C to remove the restriction site for NlaIV were pur- chased from IDT. The sequence of ssDNA is as follows: 5′-AATGAG AAGAAAGCTTCCAAAGTCCTTGGCATTGTATTCTTTGTGTTTCTGA TCATGTGGTGTCCGCTTTTCATCACCAATATCCTGTCGGTGCTT TGTGGGAAGGCCTGTAACCAAAAGCTAATGGAGAAACTTCTCAATG TGTTTGTTTGGATT. The BCM Genetically Engineered Murine Model Core microinjected Cas9 (20 ng μl−1), ssDNA (20 ng µl−1) and sgRNA (20 ng μl−1) into the pronuclei of 139 one-cell stage C57BL/6J embryos. Founder animals (F0) were identified by PCR-based restriction diges- tion to detect the CRISPR generated point mutations in the Htr2c gene. PCR product was amplified with the primer pairs: 5′-ACGTCGAAAGA AGAAAGAAAAGC and 5′-GGTAAATTTTGTTGAAGAGAGTGTAC. The 266-bp PCR products were then digested with NlaIV. After the diges- tion, 153, 79 and 34-bp fragments could be detected for PCR products from a WT allele; 232 and 34 bp fragments could be detected from a mutant Htr2cF327L allele. Three independent lines were sequenced for the further confirmation of the point mutation. One of these lines was crossed to C57BL/6J to produce study cohorts. In some breeding, the POMC-CreER/Rosa26-LSL-td TOMATO alleles were introduced to allow specific labeling of POMC neurons. Male WT and Htr2cF327L/Y mice (at 7 months old) and female WT and Htr2cF327L/+ mice (at 4 months old) were briefly fasted for 2 h (14:00 to 16:00) to empty the stomach and then received i.p. injections of saline or lorcaserin (3 mg kg−1). One hour later, mice were anesthetized with inhaled isoflurane and quickly perfused with saline, followed by 10% formalin. The brain sections were cut at 25 µm and collected into five consecutive series. One series of brain sections were blocked with 3% normal donkey serum for 1 h incubated with rabbit anti-β-endorphin antibody (1:10,000 dilution; no. TSE PhenoMaster metabolic cages Male mice were acclimated into the TSE PhenoMaster system at 18 weeks (the end of chow feeding period) and again at 27 weeks (the end of HFD feeding period). Female mice were acclimated to the TSE PhenoMaster system at 27 weeks (the end of HFD feeding period). After acclimation for 2 d, food intake, O2 consumption, CO2 production, heat production, xy axis and z axis movements were continuously monitored for 4 d and data collected from the last 2 d were used for analysis. O2 consumption, CO2 production and heat production were analyzed using the online CalR tool50. Slices were transferred to the recording chamber at 32 °C and perfused continuously with oxygenated artificial cerebrospinal fluid at a flow rate of 1.8–2.0 ml min−1. Slices were allowed to equilibrate for at least 5 min before recording. tdTOMATO-labeled neurons in the ARH were visualized using epifluorescence and infrared–differential interference contrast (IR–DIC) imaging on an upright microscope (Eclipse FN-1, Nikon) equipped with a moveable stage (MP-285, Sutter Instrument). Patch pipettes with resistances of 3–5 MΩω were filled with intracellular solution (pH 7.3) containing 128 mM potassium glu- conate, 10 mM KCl, 10 mM HEPES, 0.1 mM EGTA, 2 mM MgCl2, 0.05 mM GTP (sodium salt) and 0.05 mM ATP (magnesium salt). Recordings were made using a MultiClamp 700B amplifier (Axon Instruments), filtered at 1 kHz and sampled at 10 kHz using Digidata 1440A and analyzed offline with pClamp v.10.3 software (Axon Instruments). Series resist- ance was monitored during the recording and the values were generally <10 MΩ and were not compensated. The liquid junction potential was monitored and corrected. Data were excluded if the series resistance Glucose and insulin tolerance tests For glucose tolerance tests, after an overnight fast, mice received intra- peritoneal (i.p.) injections of 1 g kg−1 d-glucose (G8270, Sigma) at 10:00. Blood glucose was measured from tail blood using a glucometer (One- TouchUltra) at 0, 15, 30, 60 and 120 min. For insulin tolerance tests, after a 4-h fast to empty the stomach, mice received i.p. injections of insulin (0.75 U kg−1). Blood glucose was measured at 0, 15, 30, 60 and 90 min. Electrophysiology Male POMC-CreER/Rosa26-LSL-tdTOMATO and POMC-CreER/ Rosa26-LSL-td TOMATO/Htr2cF327L/Y littermates at 5–6 months old were used for recordings from POMC neurons in the ARH. Mice were anesthetized with isoflurane and were transcardially perfused with a modified ice-cold sucrose-based cutting solution (pH 7.4; containing 10 mM NaCl, 25 mM NaHCO3, 195 mM sucrose, 5 mM glucose, 2.5 mM KCl, 1.25 mM NaH2PO4, 2 mM sodium pyruvate, 0.5 mM CaCl2 and 7 mM MgCl2, bubbled continuously with 95% O2 and 5% CO2). The mice were then decapitated and the entire brain was removed and immediately submerged in the cutting solution. Coronal slices (220 μm) were cut with a Microm HM 650V vibratome (Thermo Scientific). Brain slices containing the ARH were collected and recordings were made at levels throughout this brain region. The slices were recovered for ~ 30 min at 32 °C and then maintained at room temperature for another 1 h in oxygenated (95% O2 and 5% CO2) artificial cerebrospinal fluid (pH 7.4; containing 126 mM NaCl, 2.5 mM KCl, 2.4 mM CaCl2, 1.2 mM NaH2PO4, 1.2 mM MgCl2, 11.1 mM glucose and 21.4 mM NaHCO3) before recording. Food intake, body weight and body composition Food intake, body weight and body composition Male and female WT and Htr2cF327L littermates were singly housed from 5 weeks of age. Mice were fed ad libitum with a regular chow diet (5V5R-Advanced Protocol PicoLab Select Rodent 50 IF/6F, PicoLab) from weaning to 18 weeks of age and then fed with a HFD (60% fat, no. D12492i, Research Diets) from 18 weeks to 27 weeks of age. Body weight and food intake were measured weekly. Fat mass and lean mass were determined by quantitative magnetic resonance at 18 weeks of age in male mice and at 27 weeks of age in both male and female mice. Approval for studies in mice H-02233, Phoenix Peptide) and mouse anti-c-Fos antibody (1:1,000 dilution, Ab208942, Abcam) on shaker at room temperature for overnight, followed by the donkey anti-rabbit Alexa Fluor 488 (1:200 dilution, A21206, Invitrogen) and donkey anti-mouse Alexa Fluor 594 (1:200 dilution, A21203, Invitro- gen) for 2 h. Slides were cover-slipped and analyzed using a fluores- cence microscope. The numbers of β-endorphin-positive cells and c-Fos/β-endorphin double positive neurons in the arcuate nucleus of the hypothalamus were counted in all brain sections and the ratio of c-Fos/β-endorphin double positive neurons to β-endorphin-positive neurons was used to reflect the data value for that mouse. Three mice were included in each group for statistical analyses. Leptin-induced anorexia Male WT and Htr2cF327L/Y mice (at 6 months old) and female WT and Htr2cF327L/+ mice (at 2 months old) were briefly fasted for 2 h (16:00 to 18:00). Then, these mice received i.p. injections of saline or leptin (5 mg kg−1) at 18:00. Food was provided to the cages immediately after the injections and food intake was measured for 1 h. Three days later, these mice were subjected to the same protocol with leptin or saline injections in a crossover fashion. References 49. Lotta, L. A. et al. Human gain-of-function MC4R variants show signaling bias and protect against obesity. Cell 177, 597–607 (2019). 50. Amir I., Mina Raymond A., LeClair Katherine B., LeClair David E., Cohen Louise, Lantier Alexander S., Banks (2018) CalR: A Web-Based Analysis Tool for Indirect Calorimetry Experiments. Cell Metabolism 28(4) 656-666.e1 S1550413118304017 https://doi. org/10.1016/j.cmet.2018.06.019 Reporting summary Further information on research design is available in the Nature Port- folio Reporting Summary linked to this article. Three-chamber social interaction test This test was used to evaluate social behavior in male WT and Htr2cF327L/Y littermates (4–5 months old) or in female WT and Htr2cF327L/+ littermates (4 months old). The social interaction test used a three-chambered box with openings between chambers for the mouse to pass through. The test had three sessions: habituation, sociability and social novelty. During the habituation, we put an empty pencil cup upside down into each of the side chambers and a test mouse was released into the center chamber and allowed to explore all the chambers for 15 min. After the habituation session, one novel object was put into the pencil cup on one side and a never-before-met intruder mouse (mouse 1) was placed under the cup on the other side. The sociability session took 15 min. A second never-before-met intruder mouse (mouse 2) was used during the social novelty session to swap out the novel object under the pencil cup. The test mouse again had 15 min to investigate each chamber. The time spent sniffing each pencil cup was recorded. During the sociabil- ity session, we calculated the ratio between time spent sniffing mouse 1 and total time spent sniffing mouse 1 or the object to reflect test mouse’s sociability; in the social novelty session, we calculated the ratio between time spent sniffing mouse 2 and total time spent sniffing mouse 1 or mouse 2 to reflect test mouse’s social novelty. Exome sequencing data are accessible from the European Genome Archive under a managed access agreement (EGAS00001000124 and EGAS00001000825). Source data from cell based and animal stud- ies are provided. Anonymized clinical data are available to bonafide researchers from the corresponding author with no restrictions (I.S.F.). Source data are provided with this paper. Statistical analysis ll l l All results were analyzed using GraphPad Prism 8 (GraphPad Soft- ware) to evaluate normal distribution and variations within and among groups. For inositol triphosphate turnover assays β-arrestin coupling assays, statistical significance of differences in Emax and EC50 between WT and variant receptors was determined using unpaired Student’s t-tests with Welch correction. For all animal studies, the minimal sam- ple size was predetermined by the nature of experiments. For most of physiological readouts (such as body weight, food intake, energy expenditure and body composition), at least six different mice per group were included. For histology studies, the same experiment was repeated in at least three different mice. For electrophysiological studies, at least 30 different neurons from three different mice were included. The data are presented as mean ± s.e.m. or as individual data points. Methods of statistical analyses were chosen based on the design of each experiment and are indicated in figure legends. P < 0.05 was considered to be statistically significant. https://doi.org/10.1038/s41591-022-02106-5 To ensure that each recorded neuron receive same amount of lorcaserin, the neurons located on the surface of the slice were selected for recording and the puff pipette was always put at a 100 μm horizontal and 100 μm vertical distance from the recorded neurons. The puff strength was maintained at a same level using a repeatable pressure pulse system (Picospritzer III, Parker). Each neuron was recorded at least 1 min baseline and only the neurons with stable baseline were used to test the lorcaserin treatment. The values of resting membrane potential and firing frequency were averaged in baseline and in a 1-min range containing the point with the maximal change in resting membrane potential after lorcaserin puff. A neuron was considered activated if a change in membrane potential was at least 2 mV, whereas values between a 2 mV were defined as ‘non-responsive’. Clampfit v.10.6 was used to analyze electrophysiology data. Acknowledgements We are indebted to the participants and their families for their participation and to the physicians involved in GOOS (www.goos. org.uk). This study was supported by funding from Wellcome (207462/Z/17/Z to I.S.F.), the National Institute for Health and Care Research (NIHR) Cambridge Biomedical Research Centre, a NIHR Senior Investigator Award, the Botnar Foundation and the Bernard Wolfe Health Neuroscience Endowment (to I.S.F.). B.B. was supported by an EMBO Long-term Fellowship. Y.X. was supported by grants from the National Institutes of Health (NIH) (P01DK113954, https://doi.org/10.1038/s41591-022-02106-5 https://doi.org/10.1038/s41591-022-02106-5 Article 6 min. After the test, videos were analyzed using Noldus EthoVision XT (v.14.0) or by one tester for different behaviors. The number of visits to open arms and the time spent in open arms were calculated. One tester was blind to the group information of each mouse and counted other behaviors in the videos. When a mouse stands at the center area decid- ing which arm to enter, the mouse will show a stretch-attend posture with head and shoulders stretching toward the open arms. This posture is defined at ‘risk assessment behavior’. The tester counted the number and the duration of risk assessment behavior in the boundary between the center area and the open arms. When the mouse explores the open arm, the mouse sometimes will show a head dipping posture standing at the side edge of the open arms with head and shoulders dipping down toward the floor. The number and duration of head dipping behavior was counted by the tester as an indication of fearless behavior. exceeding 20% change during the experiment or without overshoot for action potential. Current clamp was engaged to test neural firing frequency and resting membrane potential at the baseline and after puff delivery of lorcaserin (5 s at 30 μM). To ensure that each recorded neuron receive same amount of lorcaserin, the neurons located on the surface of the slice were selected for recording and the puff pipette was always put at a 100 μm horizontal and 100 μm vertical distance from the recorded neurons. The puff strength was maintained at a same level using a repeatable pressure pulse system (Picospritzer III, Parker). Each neuron was recorded at least 1 min baseline and only the neurons with stable baseline were used to test the lorcaserin treatment. The values of resting membrane potential and firing frequency were averaged in baseline and in a 1-min range containing the point with the maximal change in resting membrane potential after lorcaserin puff. A neuron was considered activated if a change in membrane potential was at least 2 mV, whereas values between a 2 mV were defined as ‘non-responsive’. Clampfit v.10.6 was used to analyze electrophysiology data. exceeding 20% change during the experiment or without overshoot for action potential. Current clamp was engaged to test neural firing frequency and resting membrane potential at the baseline and after puff delivery of lorcaserin (5 s at 30 μM). Lorcaserin-induced anorexia and c-fos expression in POMC neurons Male WT and Htr2cF327L/Y mice (at 6 months old) and female WT and Htr2cF327L/+ mice (at 2 months old) were briefly fasted for 2 h (16:00 to 18:00). Then, these mice received i.p. injections of saline or lorcaserin (3 mg kg−1) at 18:00. Food was provided to the cages immediately after the injections and food intake was measured for 1 h. Three days later, these mice were subjected to the same protocol with lorcaserin or saline injections in a crossover fashion. Nature Medicine Resident-intruder test We used the resident-intruder paradigm to measure social behaviors of male WT and Htr2cF327L/Y littermates (4–5 months old) or in female WT and Htr2cF327L/+ littermates (4 months old) in a semi-natural set- ting7. Mice were singly housed in the resident cage for at least 1 week before testing. The cage remained uncleaned and unchanged for 1 week before testing so that there were olfactory cues to enhance the resident mouse’s territoriality. We started the test by introducing an unfamiliar retired male breeder into the home cage in the afternoon and 7 min later, the intruder was removed from the home cage. All behaviors of mice were continuously recorded with a video camera during the 7-min period. The recorded videos were analyzed in a blinded fashion to measure the time spent by the resident mouse in various behaviors listed below: social exploration (nose–nose sniffing and anogenital sniffing), non-social exploration (self-grooming and cage exploration), defensive behaviors (move away from the intruder, flight and freeze) and offensive behaviors (lateral threat, upright position, clinch attack, keep down and chasing). Additional information Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41591-022-02106-5. Risk assessment in the elevated plus-maze test The EPM apparatus is a ‘+’ shaped maze made of plexi-glass. It consists of two open arms, two closed arms and a center area that are elevated above the floor. The test starts once the animal is placed into the center area. As the animal freely explores the whole maze areas, its behavior is recorded by a camera mounted above the maze. The test lasts for Nature Medicine Article https://doi.org/10.1038/s41591-022-02106-5 R01DK115761, R01DK117281 and R01DK120858) and from the US Department of Agriculture Current Research Information System (51000-064-01S). C.W. was supported by NIH K01DK119471. J.X. was supported by NIH R01CA193455. S.M.H. was supported by Nancy Chang Award for Research Excellence at Baylor College of Medicine, American Diabetes Association (1-18-IBS-105) and NIH R01DK126042 and NIH R01DK114356 awards. A.R.C. was supported by Baylor College of Medicine Bridge to Independence Award. Y.H. and L.T. were supported by fellowship awards from the American Heart Association. K.C. was supported by a US Department of Agriculture Agricultural Research Service fellowship. I.B. acknowledges support from an ‘Expanding excellence in England’ award from Research England. Clinical studies were performed on the Wellcome-MRC Institute of Metabolic Science Translational Research Facility funded by Wellcome (208363/Z/17/Z). The authors thank the Genetically Engineered Rodent Model Core at Baylor College of Medicine, partially supported by NIH grant P30CA125123, for generating the mouse model. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. M.Y., L.T., N.Z., N.Y., J.H., N.A.S., Z.Y., K.M.C., C.P., J.C.B., M.W. and S.M.H. assisted in production of study mice, phenotypic studies and data analyses. Y.X. and I.S.F. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors approved the final manuscript. Competing interests I.S.F. has consulted for a number of companies involved in the development of weight loss drugs (Rhythm Pharmaceuticals, Eli Lilly and Novo Nordisk). All other authors have no competing interests. Author contributions J.M.K., R.B., E.H. and I.S.F. were involved in recruiting the cohort, clinical studies and analyzing clinical data. K.L., I.B. and I.S.F. were involved in genetic analyses. B.B., E.M.O., A.P., V.A. and J.M. performed the functional characterization of human mutations in cells. Y.H., Hesong Liu and Y.X. were involved in animal experimental design and most procedures, data acquisition and analyses and writing the manuscript. Hailan Liu assisted with histology studies. D.K.L., L.L. and J.X. assisted in the generation of the mouse line. Y.Y., A.R.C., C.W., Peer review information Nature Medicine thanks Tamas Horvath, Anke Hinney and Michael Cowley for their contribution to the peer review of this work. Primary Handling Editor: Jennifer Sargent, in collaboration with the Nature Medicine team. Reprints and permissions information is available at www.nature.com/reprints. Reprints and permissions information is available at www.nature.com/reprints. Nature Medicine Article https://doi.org/10.1038/s41591-022-02106-5 Extended Data Fig. 1 \| Glucose homeostasis in HTR2C variant carriers compared to age and BMI matched controls. a, Fasting plasma insulin levels (log10-transformed) in 14 HTR2C variant carriers compared to 2138 severely obese children as controls [GOOS]. b and c show the mean fasting plasma insulin levels in HTR2C variant carriers of all ages (N = 14) or under 12 years of age (N = 6). d, Fasting plasma glucose levels in 12 HTR2C variant carriers compared to 2248 severely obese children as controls [GOOS]. e and f show the mean fasting plasma glucose levels in HTR2C variant carriers of all ages (N = 12); and in HTR2C variant carriers up to 12 years of age (N = 5), respectively. In a and d, the 5th and 95th percentiles are shown as dashed lines and the regression slopes are shown as solid lines in gray [GOOS] or orange [HTR2C variant carriers]. g, The relationship between HOMA-IR (Homeostasis Model Assessment of insulin resistance) scores and age in 12 HTR2C variant carriers compared to 2024 severely obese children as controls [GOOS]. HOMA-IR was calculated as [(fasting insulin in μU/mL) × (fasting glucose in mg/dL)] / 405. Dashed lines represent 95% confidence intervals; regression slope shown as solid lines in gray [GOOS] or orange [HTR2C variant carriers]. A dotted horizontal line indicates the threshold above which values are consistent with insulin resistance. Author contributions h and i show the mean HOMA-IR scores in HTR2C variant carriers of all ages (N = 12); and on HTR2C variant carriers up to 12 years of age (N = 5), respectively. Statistical significance was determined using two-tailed Mann-Whitney U test. Data are shown as mean ± SEM. Article https://doi.org/10.1038/s41591-022-02106-5 https://doi.org/10.1038/s41591-022-02106-5 Extended Data Fig. 1 \| Glucose homeostasis in HTR2C variant carriers between HOMA-IR (Homeostasis Model Assessment of insulin resistance) scores and age in 12 HTR2C variant carriers compared to 2024 severely obese children as controls [GOOS]. HOMA-IR was calculated as [(fasting insulin in μU/mL) × (fasting glucose in mg/dL)] / 405. Dashed lines represent 95% confidence intervals; regression slope shown as solid lines in gray [GOOS] or orange [HTR2C variant carriers]. A dotted horizontal line indicates the threshold above which values are consistent with insulin resistance. h and i show the mean HOMA-IR scores in HTR2C variant carriers of all ages (N = 12); and on HTR2C variant carriers up to 12 years of age (N = 5), respectively. Statistical significance was determined using two-tailed Mann-Whitney U test. Data are shown as mean ± SEM. g \| compared to age and BMI matched controls. a, Fasting plasma insulin levels (log10-transformed) in 14 HTR2C variant carriers compared to 2138 severely obese children as controls [GOOS]. b and c show the mean fasting plasma insulin levels in HTR2C variant carriers of all ages (N = 14) or under 12 years of age (N = 6). d, Fasting plasma glucose levels in 12 HTR2C variant carriers compared to 2248 severely obese children as controls [GOOS]. e and f show the mean fasting plasma glucose levels in HTR2C variant carriers of all ages (N = 12); and in HTR2C variant carriers up to 12 years of age (N = 5), respectively. In a and d, the 5th and 95th percentiles are shown as dashed lines and the regression slopes are shown as solid lines in gray [GOOS] or orange [HTR2C variant carriers]. g, The relationship Nature Medicine https://doi.org/10.1038/s41591-022-02106-5 Article tended Data Fig. 2 \| Functional characterization of human variants 5-HT2CR. a, Representative micrographs of cell surface and intracellular pression of WT vs mutant receptors (green), quantified by high-content nfocal microscopy. Phalloidin and DAPI (blue) were used to stain the cytoplasm d nuclei, respectively and were used for cell identification. Scale bar = 50 μm. b-c, Number of receptor-containing vesicles (b), and the size of these receptor positive vesicles (c), expressed as % of WT. Differences between WT and mutant receptors were compared with two-tailed student’s t-test with Welch correction. Results represent mean ± SEM of 3 independent experiments. https://doi.org/10.1038/s41591-022-02106-5 b-c, Number of receptor-containing vesicles (b), and the size of these receptor positive vesicles (c), expressed as % of WT. Differences between WT and mutant receptors were compared with two-tailed student’s t-test with Welch correction. Results represent mean ± SEM of 3 independent experiments. b-c, Number of receptor-containing vesicles (b), and the size of these receptor positive vesicles (c), expressed as % of WT. Differences between WT and mutant receptors were compared with two-tailed student’s t-test with Welch correction. Results represent mean ± SEM of 3 independent experiments. Extended Data Fig. 2 \| Functional characterization of human variants in 5-HT2CR. a, Representative micrographs of cell surface and intracellular expression of WT vs mutant receptors (green), quantified by high-content confocal microscopy. Phalloidin and DAPI (blue) were used to stain the cytoplasm and nuclei, respectively and were used for cell identification. Scale bar = 50 μm. Nature Medicine https://doi.org/10.1038/s41591-022-02106-5 Article Article https://doi.org/10.1038/s41591 022 02106 5 Extended Data Fig. 3 \| Metabolic phenotype in chow-fed male Htr2cF327L/Y mice. a, The PCR products around the mouse F328 (equivalent to human F327) were amplified from genomic DNA extracts and incubated with NlaIV. A WT allele was cut into 153, 79 and 34 bp bands. A F327L mutant allele resulted in 232 and 34 bp bands. Heterozygous alleles (a WT allele + a F327L mutant allele) resulted in 232, 153, 79 and 34 bp bands (repeated 3 times independently for all groups). b-d, Left panels: temporal levels of O2 consumption (b), CO2 production (c) and heat production (d) during a 2-day period measured by the TSE PhenoMaster in 18- week old chow-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice. Right panels: the regression of O2 consumption, CO2 production and heat production with body mass. Two way ANOVA followed by Sidak’s multiple comparisons were performed in b (left panel), c (left panel), d (left panel). P values in b (left panel), c (left panel), d (left panel) were determined by two-tailed unpaired t-test. Regression-based ANCOVA analysis was performed in b (right panel), c (right panel) and d (right panel). , p < 0.05. Data are presented as mean ± SEM., p < 0.05. Extended Data Fig. 3 \| Metabolic phenotype in chow-fed male Htr2cF327L/Y mice a ThePCRproductsaroundthemouseF328(equivalenttohumanF327) week old chow-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice. https://doi.org/10.1038/s41591-022-02106-5 Right panels: the regressionofO consumption CO productionandheatproductionwithbody week old chow-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice. Right panels: the regression of O2 consumption, CO2 production and heat production with body mass. Two way ANOVA followed by Sidak’s multiple comparisons were performed in b (left panel), c (left panel), d (left panel). P values in b (left panel), c (left panel), d (left panel) were determined by two-tailed unpaired t-test. Regression-based ANCOVA analysis was performed in b (right panel), c (right panel) and d (right panel). , p < 0.05. Data are presented as mean ± SEM., p < 0.05. week old chow-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice. Right panels: the regression of O2 consumption, CO2 production and heat production with body mass. Two way ANOVA followed by Sidak’s multiple comparisons were performed in b (left panel), c (left panel), d (left panel). P values in b (left panel), c (left panel), d (left panel) were determined by two-tailed unpaired t-test. Regression-based ANCOVA analysis was performed in b (right panel), c (right panel) and d (right panel). , p < 0.05. Data are presented as mean ± SEM., p < 0.05. Extended Data Fig. 3 \| Metabolic phenotype in chow-fed male Htr2cF327L/Y Extended Data Fig. 3 \| Metabolic phenotype in chow-fed male Htr2cF327L/Y mice. a, The PCR products around the mouse F328 (equivalent to human F327) were amplified from genomic DNA extracts and incubated with NlaIV. A WT allele was cut into 153, 79 and 34 bp bands. A F327L mutant allele resulted in 232 and 34 bp bands. Heterozygous alleles (a WT allele + a F327L mutant allele) resulted in 232, 153, 79 and 34 bp bands (repeated 3 times independently for all groups). b-d, Left panels: temporal levels of O2 consumption (b), CO2 production (c) and heat production (d) during a 2-day period measured by the TSE PhenoMaster in 18- were amplified from genomic DNA extracts and incubated with NlaIV. A WT allele was cut into 153, 79 and 34 bp bands. A F327L mutant allele resulted in 232 and 34 bp bands. Heterozygous alleles (a WT allele + a F327L mutant allele) resulted in 232, 153, 79 and 34 bp bands (repeated 3 times independently for all groups). https://doi.org/10.1038/s41591-022-02106-5 https://doi.org/10.1038/s41591-022-02106-5 Article Extended Data Fig. 4 \| Metabolic phenotype in HFD-fed male Htr2cF327L/Y mice. a, Body weight curves of male WT (N = 7) and Htr2cF327L/Y mice (N = 9) fed on HFD. b-c, Blood glucose levels in a glucose tolerance test (GTT) (b) or in an insulin tolerance test (ITT) (c) in male WT (N = 7) and Htr2cF327L/Y (N = 9) mice fed on HFD. N = 7 or 9 mice per group. d-e, Left panels: temporal levels of XY axis activity (d), Z axis activity measured by the TSE PhenoMaster in 27-week old HFD-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice (e), Right panels: averaged values during the dark cycle, light cycle or 24 hours. f-h, Temporal levels of O2 consumption (f), CO2 production (g) and heat production (h) during a 2-day period measured by the TSE PhenoMaster in 27-week old HFD-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice. Right panels: the regression of O2 consumption, CO2 production and heat production with body mass. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a, b, c, d (left panel), e (left panel), g (left panel) and h (left panel). P values in d (left panel), e (left panel), f (left panel), g (left panel) and h (left panel) were determined by two-tailed unpaired t-test. Two-tailed unpaired t-tests were performed in d (right panel) and e (right panel). Regression-based ANCOVA analysis was performed in f (right panel), g (right panel) and h (right panel). , p < 0.05. Data are presented as mean ± SEM. Extended Data Fig. 4 \| Metabolic phenotype in HFD-fed male Htr2cF327L/Y by the TSE PhenoMaster in 27-week old HFD-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice. Right panels: the regression of O2 consumption, CO2 production and heat production with body mass. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a, b, c, d (left panel), e (left panel), g (left panel) and h (left panel). P values in d (left panel), e (left panel), f (left panel), g (left panel) and h (left panel) were determined by two-tailed unpaired t-test. Two-tailed unpaired t-tests were performed in d (right panel) and e (right panel). Regression-based ANCOVA analysis was performed in f (right panel), g (right panel) and h (right panel). , p < 0.05. https://doi.org/10.1038/s41591-022-02106-5 b-d, Left panels: temporal levels of O2 consumption (b), CO2 production (c) and heat production (d) during a 2-day period measured by the TSE PhenoMaster in 18- Nature Medicine Nature Medicine https://doi.org/10.1038/s41591-022-02106-5 Data are presented as mean ± SEM. by the TSE PhenoMaster in 27-week old HFD-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice. Right panels: the regression of O2 consumption, CO2 production and heat production with body mass. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a, b, c, d (left panel), e (left panel), g (left panel) and h (left panel). P values in d (left panel), e (left panel), f (left panel), g (left panel) and h (left panel) were determined by two-tailed unpaired t-test. Two-tailed unpaired t-tests were performed in d (right panel) and e (right panel). Regression-based ANCOVA analysis was performed in f (right panel), g (right panel) and h (right panel). , p < 0.05. Data are presented as mean ± SEM. Extended Data Fig. 4 \| Metabolic phenotype in HFD-fed male Htr2cF327L/Y mice. a, Body weight curves of male WT (N = 7) and Htr2cF327L/Y mice (N = 9) fed on HFD. b-c, Blood glucose levels in a glucose tolerance test (GTT) (b) or in an insulin tolerance test (ITT) (c) in male WT (N = 7) and Htr2cF327L/Y (N = 9) mice fed on HFD. N = 7 or 9 mice per group. d-e, Left panels: temporal levels of XY axis activity (d), Z axis activity measured by the TSE PhenoMaster in 27-week old HFD-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice (e), Right panels: averaged values during the dark cycle, light cycle or 24 hours. f-h, Temporal levels of O2 consumption (f), CO2 production (g) and heat production (h) during a 2-day period measured mice. a, Body weight curves of male WT (N = 7) and Htr2cF327L/Y mice (N = 9) fed on HFD. b-c, Blood glucose levels in a glucose tolerance test (GTT) (b) or in an insulin tolerance test (ITT) (c) in male WT (N = 7) and Htr2cF327L/Y (N = 9) mice fed on HFD. N = 7 or 9 mice per group. d-e, Left panels: temporal levels of XY axis activity (d), Z axis activity measured by the TSE PhenoMaster in 27-week old HFD-fed male WT (N = 7) and Htr2cF327L/Y (N = 9) mice (e), Right panels: averaged values during the dark cycle, light cycle or 24 hours. Article Article https://doi.org/10.1038/s41591 022 02106 5 Extended Data Fig. 5 \| Metabolic phenotype in HFD-fed female Htr2cF327L/+ mice. a-b, Body weight curves (a) and weekly chow intake (b) of female WT (N = 8) and Htr2cF327L/+ (N = 8) mice fed on regular chow. c-f, Body weight curves (c), body weight gain (d), body composition (e) and weekly HFD intake (f) of female WT (N = 6) and Htr2cF327L/+ (N = 8) mice fed on HFD. g, Left panel: cumulative HFD intake during a 2-day period measured by the TSE PhenoMaster in female WT (N = 6) and Htr2cF327L/+ (N = 8) mice. Right panel: averaged HFD intake during the dark cycle, light cycle or 24 hours. h-i, Blood glucose levels in a glucose tolerance test (GTT) in female WT (N = 7) and Htr2cF327L/+ (N = 6) mice (h) or in an insulin tolerance test (ITT) in female WT (N = 8) and Htr2cF327L/+ (N = 6) mice (i). j-n, Left panels: temporal levels of XY axis activity (j), Z axis activity (k), O2 consumption (l), CO2 production (m) and heat production (n) during a 2-day period measured by the TSE PhenoMaster in HFD-fed female WT (N = 6) and Htr2cF327L/+ (N = 8) mice. Right panels: averaged values during the dark cycle, light cycle in j and k and the regression of O2 consumption, CO2 production and heat production with body mass in l, m and n. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a-d, f, g, h, i, j (left panel), k (left panel), l (left panel), m (left panel), n (left panel). P values in c, f, g (left panel), i, j (left panel), k (left panel), l (left panel), m (left panel), n (left panel) were determined by two-tailed unpaired t-test. Two-tailed unpaired t-tests were performed in e, g (right panel), j (right panel) and k (right panel). Regression-based ANCOVA analysis was performed in l (right panel), m (right panel) and n (right panel). , p < 0.05, , p < 0.01, *, p < 0.001. Data are presented as mean ± SEM. Extended Data Fig. 5 \| Metabolic phenotype in HFD-fed female Htr2cF327L/+ by the TSE PhenoMaster in HFD-fed female WT (N = 6) and Htr2cF327L/+ (N = 8) mice. https://doi.org/10.1038/s41591-022-02106-5 f-h, Temporal levels of O2 consumption (f), CO2 production (g) and heat production (h) during a 2-day period measured Nature Medicine Nature Medicine https://doi.org/10.1038/s41591-022-02106-5 Article Article Article injections of saline or leptin (5 mg perkg) c Representativecurrentclampbaseline(withoutanytreatment) by β-endorphin) expressing c-fos in chow-fed male WT (N = 3) and Htr2cF327L/Y mice (N = 3) (7 months of age) (f) and chow-fed female WT (N = 4) and Htr2cF327L/+ mice (N = 4) (4 months of age) (g) after i.p. injections of saline or lorcaserin (3 mg per kg). h-i, Representative microscopic images showing immunoreactivity of β-endorphin (left green), c-fos (middle red) and merge in male (repeated 3 times independentlyforallgroups)(h)andfemale(repeated4timesindependentlyfor Extended Data Fig. 6 \| Activity of Pomc neurons in male Htr2cF327L/Y mice. a, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. injections of saline or lorcaserin (3 mg per kg). b, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. injections of saline or leptin (5 mg per kg). c, Representative current clamp baseline (without any treatment) traces in Pomc neurons from WT and Htr2cF327L/Y mice. d-e, Baseline resting membrane potential (d) and firing frequency (e) in Pomc neurons from WT (N = 31) and Htr2cF327L/Y (N = 35) mice. f-g, Percentage of Pomc neurons (labeled by β-endorphin) expressing c-fos in chow-fed male WT (N = 3) and Htr2cF327L/Y mice (N = 3) (7 months of age) (f) and chow-fed female WT (N = 4) and Htr2cF327L/+ mice (N = 4) (4 months of age) (g) after i.p. injections of saline or lorcaserin (3 mg per kg). h-i, Representative microscopic images showing immunoreactivity of β-endorphin (left green), c-fos (middle red) and merge in male (repeated 3 times independently for all groups) (h) and female (repeated 4 times independently for all groups) (i) mice. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a, b, f, g. Two-tailed unpaired t-tests were performed in d and e. Data are presented as mean ± SEM. Extended Data Fig. 6 \| Activity of Pomc neurons in male Htr2cF327L/Y mice. a, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. injections of saline or lorcaserin (3 mg per kg). b, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. Article Right panels: averaged values during the dark cycle, light cycle in j and k and the regression of O2 consumption, CO2 production and heat production with body mass in l, m and n. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a-d, f, g, h, i, j (left panel), k (left panel), l (left panel), m (left panel), n (left panel). P values in c, f, g (left panel), i, j (left panel), k (left panel), l (left panel), m (left panel), n (left panel) were determined by two-tailed unpaired t-test. Two-tailed unpaired t-tests were performed in e, g (right panel), j (right panel) and k (right panel). Regression-based ANCOVA analysis was performed in l (right panel), m (right panel) and n (right panel). , p < 0.05, , p < 0.01, *, p < 0.001. Data are presented as mean ± SEM. mice. a-b, Body weight curves (a) and weekly chow intake (b) of female WT (N = 8) and Htr2cF327L/+ (N = 8) mice fed on regular chow. c-f, Body weight curves (c), body weight gain (d), body composition (e) and weekly HFD intake (f) of female WT (N = 6) and Htr2cF327L/+ (N = 8) mice fed on HFD. g, Left panel: cumulative HFD intake during a 2-day period measured by the TSE PhenoMaster in female WT (N = 6) and Htr2cF327L/+ (N = 8) mice. Right panel: averaged HFD intake during the dark cycle, light cycle or 24 hours. h-i, Blood glucose levels in a glucose tolerance test (GTT) in female WT (N = 7) and Htr2cF327L/+ (N = 6) mice (h) or in an insulin tolerance test (ITT) in female WT (N = 8) and Htr2cF327L/+ (N = 6) mice (i). j-n, Left panels: temporal levels of XY axis activity (j), Z axis activity (k), O2 consumption (l), CO2 production (m) and heat production (n) during a 2-day period measured Nature Medicine Nature Medicine https://doi.org/10.1038/s41591-022-02106-5 Article Extended Data Fig. 6 \| Activity of Pomc neurons in male Htr2cF327L/Y mice. a, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. injections of saline or lorcaserin (3 mg per kg). b, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. Article injections of saline or leptin (5 mg per kg). c, Representative current clamp baseline (without any treatment) traces in Pomc neurons from WT and Htr2cF327L/Y mice. d-e, Baseline resting membrane potential (d) and firing frequency (e) in Pomc neurons from WT (N = 31) and Htr2cF327L/Y (N = 35) mice. f-g, Percentage of Pomc neurons (labeled by β-endorphin) expressing c-fos in chow-fed male WT (N = 3) and Htr2cF327L/Y mice (N = 3) (7 months of age) (f) and chow-fed female WT (N = 4) and Htr2cF327L/+ mice (N = 4) (4 months of age) (g) after i.p. injections of saline or lorcaserin (3 mg per kg). h-i, Representative microscopic images showing immunoreactivity of β-endorphin (left green), c-fos (middle red) and merge in male (repeated 3 times independently for all groups) (h) and female (repeated 4 times independently for all groups) (i) mice. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a, b, f, g. Two-tailed unpaired t-tests were performed in d and e. Data are presented as mean ± SEM. by β-endorphin) expressing c-fos in chow-fed male WT (N = 3) and Htr2cF327L/Y mice (N = 3) (7 months of age) (f) and chow-fed female WT (N = 4) and Htr2cF327L/+ mice (N = 4) (4 months of age) (g) after i.p. injections of saline or lorcaserin (3 mg per kg). h-i, Representative microscopic images showing immunoreactivity of β-endorphin (left green), c-fos (middle red) and merge in male (repeated 3 times independently for all groups) (h) and female (repeated 4 times independently for all groups) (i) mice. Two way ANOVA followed by Sidak’s multiple comparisons were performed in a, b, f, g. Two-tailed unpaired t-tests were performed in d and e. Data are presented as mean ± SEM. byβ-endorphin)expressingc-fosinchow-fedmaleWT(N = 3)andHtr2cF327L/Y Extended Data Fig. 6 \| Activity of Pomc neurons in male Htr2cF327L/Y mice. a, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. injections of saline or lorcaserin (3 mg per kg). b, One hour food intake in chow-fed female WT (N = 10) and Htr2cF327L/+ mice (2 months of age) (N = 13) after i.p. Article Extended Data Fig. 7 \| Abnormal behaviour in knock-in female Htr2cF327L/+ mice. a-c, Social exploration time (a), non-social exploration time (b) and defensive time (c), spent by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice (4 months of age) in the resident-intruder test. d, Percentage/number of female WT and Htr2cF327L/+ mice that exhibited offensive behaviors towards the intruder. e-g, Offensive latency, number and time quantified in female WT (N = 10) and Htr2cF327L/+ (N = 12) mice in the intruder test. h, Ratio of time spent sniffing a mouse when interacting with a mouse and an object in the three-chamber social interaction test. N = 9 mice in WT group and 12 mice Htr2cF327L/+ group. i, Ratio of time spent in sniffing a novel mouse when interacting with a new mouse and a familiar mouse in the three-chamber social interaction test, N = 9 mice in WT group and 12 mice Htr2cF327L/+ group. j-k, Number (j) and time (k) of risk assessment behaviour by female WT (N = 10) and Htr2cF327L/+ mice (N = 12) (4 months of age) in the EPM test. l-m, Number of entries to (l) and time spent on (m) the open arms of the EPM apparatus by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice. n-o, Number of (n) and time (o) spent in head dipping on the open arms of the EPM apparatus by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice. Two-tailed unpaired t-tests. Data are presented as mean ± SEM. of time spent in sniffing a novel mouse when interacting with a new mouse Extended Data Fig. 7 \| Abnormal behaviour in knock-in female Htr2cF327L/+ of time spent in sniffing a novel mouse when interacting with a new mouse and a familiar mouse in the three-chamber social interaction test, N = 9 mice in WT group and 12 mice Htr2cF327L/+ group. j-k, Number (j) and time (k) of risk assessment behaviour by female WT (N = 10) and Htr2cF327L/+ mice (N = 12) (4 months of age) in the EPM test. l-m, Number of entries to (l) and time spent on (m) the open arms of the EPM apparatus by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice. Article injections of saline or leptin (5 mg per kg). c, Representative current clamp baseline (without any treatment) traces in Pomc neurons from WT and Htr2cF327L/Y mice. d-e, Baseline resting membrane potential (d) and firing frequency (e) in Pomc neurons from WT (N = 31) and Htr2cF327L/Y (N = 35) mice. f-g, Percentage of Pomc neurons (labeled Nature Medicine Nature Medicine https://doi.org/10.1038/s41591-022-02106-5 Article Article Article n-o, Number of (n) and time (o) spent in head dipping on the open arms of the EPM apparatus by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice. Two-tailed unpaired t-tests. Data are presented as mean ± SEM. of time spent in sniffing a novel mouse when interacting with a new mouse and a familiar mouse in the three-chamber social interaction test, N = 9 mice in WT group and 12 mice Htr2cF327L/+ group. j-k, Number (j) and time (k) of risk assessment behaviour by female WT (N = 10) and Htr2cF327L/+ mice (N = 12) (4 months of age) in the EPM test. l-m, Number of entries to (l) and time spent on (m) the open arms of the EPM apparatus by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice. n-o, Number of (n) and time (o) spent in head dipping on the open arms of the EPM apparatus by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice. Two-tailed unpaired t-tests. Data are presented as mean ± SEM. ed Data Fig. 7 \| Abnormal behaviour in knock-in female Htr2c Extended Data Fig. 7 \| Abnormal behaviour in knock in female Htr2cF327L/ mice. a-c, Social exploration time (a), non-social exploration time (b) and defensive time (c), spent by female WT (N = 10) and Htr2cF327L/+ (N = 12) mice (4 months of age) in the resident-intruder test. d, Percentage/number of female WT and Htr2cF327L/+ mice that exhibited offensive behaviors towards the intruder. e-g, Offensive latency, number and time quantified in female WT (N = 10) and Htr2cF327L/+ (N = 12) mice in the intruder test. h, Ratio of time spent sniffing a mouse when interacting with a mouse and an object in the three-chamber social interaction test. N = 9 mice in WT group and 12 mice Htr2cF327L/+ group. i, Ratio Nature Medicine Nature Medicine β β β β ≥
https://openalex.org/W2932906561	https://hal.science/hal-02904857/document	English	null	Climate change in Africa: costs of mitigating heat stress	Climatic change	2,019	cc-by	9,375	To cite this version: Ben Parkes, Jennifer Cronin, Olivier Dessens, Benjamin Sultan. Climate change in Africa: costs of mitigating heat stress. Climatic Change, 2019, 154 (3-4), pp.461-476. 10.1007/s10584-019-02405-w. hal-02904857 Distributed under a Creative Commons Attribution 4.0 International License Ben Parkes ben.parkes@manchester.ac.uk HAL Id: hal-02904857 https://hal.science/hal-02904857v1 Submitted on 3 Jul 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License Climatic Change (2019) 154:461–476 https://doi.org/10.1007/s10584-019-02405-w Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10584-019-02405-w) contains supplementary material, which is available to authorized users. 1 Sorbonne Universit´es, (UPMC, Univ. Paris 06)-CNRS-IRD-MNHN LOCEAN/IPSL, 4 Place Jussieu, F-75005, Paris, France 4 ESPACE-DEV, Univ Montpellier, IRD, Univ Guyane, Univ R´eunion, Univ Antilles, Univ Avignon, Montpellier, France 2 Present address: School of Mechanical, Aerospace and Civil Engineering, University of Manchester, Manchester, M13 9PL, UK 3 Energy Institute, University College London, London, WC1H 0NN, UK 1 Introduction Heat stress is the inability of a body to cool sufficiently to maintain a stable internal temper- ature. Methods of mitigating heat stress include evaporative cooling, seeking shade, utilising additional space cooling or altering schedules to allow high intensity work in cooler temper- atures (Suzuki-Parker and Kusaka 2016). An increase in the number of heat-related deaths as a result of climate change has been observed (Field et al. 2014). With further increasing temperatures, the number of heat-related deaths is likely to increase (Dunne et al. 2013). Heat stress is affected by the ambient conditions and vulnerability or personal perception. There are several different heat stress metrics (Buzan et al. 2015), which are dependent on a number of variables including temperature, relative humidity, wind speed and radiation. Tropical regions with high temperatures and humidity are more vulnerable to heat stress than temperate or polar regions (Zhao et al. 2015). Coastal cities are particularly susceptible to changes in heat stress as evaporation from the sea increases the water vapour pressure which is a key component of heat stress (Diffenbaugh et al. 2007). Increases in heat stress reduce a person’s ability to work; manual labour is increasingly difficult as the body cannot disperse heat quickly enough (Kjellstrom et al. 2009). In envi- ronments where body heat is not the limiting factor, it is still possible for heat to have an effect, as hotter environments require more cooling for comfort and food requires more energy to refrigerate. The economic costs of climate change impacts have been investigated in Burke et al. (2015), which indicates that almost every country in Africa is vulnerable to increasing temperatures causing a reduction in gross domestic product (GDP). Northern Africa is vulnerable to urban heat stress amplification, where similar weather conditions when combined with the urban heat island leads to significant heat stress. In future climates, this vulnerability is expected to worsen, leading to higher levels of heat stress (Fischer et al. 2012). Heat stress incidence is expected to increase with climate change (Zhao et al. 2015); in tropical regions, this will reduce industrial activity as workplaces become unsuitable. The lost working time will lead to reduced economic productivity. This is of particular impor- tance in the tropics which makes up a large portion of the developing world (Lundgren et al. 2013). Abstract We applied two metrics, apparent temperature and humidex, to calculate heat stress in both present and future climates. We use an ensemble of CORDEX-Africa simulations to esti- mate heat stress during a baseline period and at two specific warming levels, 2 and 4 ◦C above pre-industrial. The increase in temperatures and changes to the precipitation distri- bution under climate change are projected to increase the intensity of heat stress events in Sahelian Africa and introduce new heat stress events in Northern and Central Africa. As the intensity of heat stress increases, it is expected that the use of energy-intensive cooling will increase. The energy system, therefore, will need to be able to supply more energy to power fans or air conditioning units. The cooling demand to turn a heat stress event into a non- heat stress event is computed. This value is then weighted by the population to find the total cooling required to prevent heat stress across the continent. Country-level results indicate that the greatest increases in cooling demand will occur in countries with densely populated regions, most notably Nigeria. Supplying this additional cooling demand will present the greatest challenge to less developed countries like Somalia. We find the least-cost future energy system that meets the projected increase in demand and derive the increase in energy system costs with the TIAM-UCL model. The total increase in energy costs to prevent heat stress is found to be $51bn by 2035 and $487bn by 2076. Ben Parkes ben.parkes@manchester.ac.uk Climatic Change (2019) 154:461–476 462 1 Introduction Climate change is also pro- jected to impact several parts of the energy supply system, and it is vital to understand the potential for compound impacts, particularly in regions of higher vulnerability (Cronin et al. 2018); therefore, improved demand studies focused on health outcomes are valuable. increase in residential energy demand is largely a result of demand in Asia for air condi- tioning. There is a similar signal in Africa, however, it is smaller as Africa is projected to be poorer and therefore unable to afford air conditioning (Isaac and van Vuuren 2009). Energy consumption is expected to increase with the intensity of climate change and more intense warming levels lead to higher demand. Global energy demand will increase by 17% in a high emissions scenario which is driven by tropical changes of 32% in contrast to 7 and 18% for a less intense scenario (De Cian and Sue Wing 2017). Climate change is also pro- jected to impact several parts of the energy supply system, and it is vital to understand the potential for compound impacts, particularly in regions of higher vulnerability (Cronin et al. 2018); therefore, improved demand studies focused on health outcomes are valuable. While reductions in heating energy demand may partially balance out the increases in cooling energy demand, especially at the global scale, the two are not directly equivalent in the energy system, as heating services are largely provided by a variety of different fuels and cooling by electricity (Isaac and van Vuuren 2009). As these increased electricity demands could necessitate increases in generation and transmission capacity, mitigating heat stress with increased space cooling could thus have significant impacts on regional energy sys- tems. These impacts include but are not limited to investment requirements, operating costs, energy prices and thus the wider economy. Previous studies have examined the impacts of climate change on heating and cooling demands using degree days with set target temperatures and project significant impacts at the regional scale (Isaac and van Vuuren 2009; Labriet et al. 2015). To the authors’ knowl- edge, no studies have so far examined the costs of mitigating heat stress through increased space cooling specifically. This study was conducted as part of the multinational high-end climate impacts and extremes (HELIX) project and is part of a specific focus on Africa. 1 Introduction Large-scale heatwaves can cause fatalities in Western European nations; coun- tries with fewer resources need to plan ahead to cope with heatwaves (Mitchell et al. 2018). Within Africa the Great Lakes region will experience an increase in heat stress as a result of climate change. Within this region Burundi, the Democratic Republic of the Congo, Uganda and Mozambique are most vulnerable to heat stress (Asefi-Najafabady et al. 2018). The urban fraction of the total global population is predicted to increase into the twenty- first century. The urban heat island effect is expected to exacerbate the effects of increasing temperatures increasing the frequency of heat stress (Fischer et al. 2012). The primary response to heat stress is the use of space cooling devices such as fans and air conditioning units. Both are heavily reliant on electricity, and therefore with an increase in potential heat stress, there is likely to be an increase in demand for electricity. Heatwaves are common in Sahelian Africa, and up to a third of the year is susceptible to temperatures that are highly hazardous to health (Barbier et al. 2018). Globally increas- ing temperatures as a result of climate change are expected to increase demand for energy to cool residential and commercial properties (Davis and Gertler 2015). In a high emissions scenario, Africa is projected to suffer significant heatwaves (Mora et al. 2017). As the devel- oping world continues to develop, there is expected to be an increase in energy demand for both industrial and residential uses (Medlock and Soligo 2001). In the period 2000–2011, the global residential energy demand increased by 14%, and this increase was primarily driven by urbanisation, population growth and economic growth (Nejat et al. 2015). The Climatic Change (2019) 154:461–476 463 increase in residential energy demand is largely a result of demand in Asia for air condi- tioning. There is a similar signal in Africa, however, it is smaller as Africa is projected to be poorer and therefore unable to afford air conditioning (Isaac and van Vuuren 2009). Energy consumption is expected to increase with the intensity of climate change and more intense warming levels lead to higher demand. Global energy demand will increase by 17% in a high emissions scenario which is driven by tropical changes of 32% in contrast to 7 and 18% for a less intense scenario (De Cian and Sue Wing 2017). 1 Introduction It addresses the research gap explained above by projecting heat stress across Africa on a gridded basis with a climate model ensemble, then uses a least-cost energy system optimi- sation model to project the change in energy demand for space cooling and the associated costs of mitigating heat stress. Policymakers and energy system planners at national and regional levels should consider the results, to feed into the consideration of climate change mitigation and adaptation strategies. 2 Input data and model The inputs for the analysis are the bias-corrected versions of data produced as part of the Coordinated Regional Climate Downscaling Experiment (CORDEX) Africa project (Nikulin et al. 2012). The CORDEX simulations were developed by driving regional climate models (RCMs) with data from general circulation models (GCMs) used during the CMIP5 project (Taylor et al. 2011). The bias correction was performed as part of the HELIX project with a multisegment statistical bias correction method (Grillakis et al. 2013; Papadimitriou et al. 2015). The input data used were eleven actualisations of CORDEX-Africa with a temporal res- olution of 1 day. In line with the HELIX focus on high-end climate change scenarios, for a model to be used in the analysis, it was required that the GCM reach a global average tem- perature change of + 4 ◦C for 30 years; six of the climate models which were used to drive four regional models over Africa met this criteria. Supplementary Information (SI) Table 1 shows the combinations of GCMs and RCMs used in this study. The baseline data time period was 1986–2005 which corresponds to the final 20 years of the CMIP5 historic record. The CMIP5 outputs from this time period values were used Climatic Change (2019) 154:461–476 464 as inputs for the CORDEX-Africa project. By requiring the temperatures reach + 4 ◦C, the future simulations were all from RCP8.5 which is a high emissions scenario (for details on RCPs see Meinshausen et al. (2011)). The mean of the ensemble of future climates reaches + 2 and + 4 ◦C during the 30-year periods centred on 2035 and 2076 respectively. The + 2 and + 4 ◦C time slices correspond to the specific warming levels (SWL) where the climate is + 2 and + 4 ◦C warmer than the pre-industrial control. SI Table 2 shows the time slices for the GCM results used for analysis. Due to the warming between the pre-industrial period (1870–1899) and the historic period (1986–2005) of 0.7 ◦C, the + 2 and + 4 ◦C warming levels are equivalent to + 1.3 and + 3.3 ◦C when compared with the 1986–2005 baseline period. The Shared Socioeconomic Pathway SSP3 was selected for this analysis, as it is consis- tent with a future of continued high GHG emissions and high levels of climate change (Riahi et al. 2017). 2 Input data and model This scenario corresponds to a world where policies allow the use of high emission energy sources, and there is little technological development in the energy sector. Furthermore, the focus is on regional rather than global trade and competition, and thus eco- nomic interaction between Africa and other regions is minimal. Population data for SSP3 were downscaled to a 0.5-degree grid by Murakami and Yamagata (2016), and the outputs were used to estimate the number of people affected by heat stress. Country-level GDP data was obtained from the IIASA SSP database (O‘Neill et al. 2012) and used as a driver for the energy system demand projections. The SSP3 OECD GDP projections were used in this study (Jones and O’Neill 2016). y TIAM-UCL is a technology-rich global optimisation model, which derives the least- cost future energy system within given technological, economic and policy constraints. It models the flows of energy carriers from primary resources to final service demands via stages of extraction, transformation and transportation. The model is calibrated with data describing the global energy system for the base year 2005, taken from the IEA World Energy Balances. The model uses projections of data—including availabilities and costs of primary resources and technologies—to design the optimal energy system transformation from 2005 to the end of the modelling period. Details of the input data are provided in the model documentation (Anandarajah et al. 2011). With perfect foresight over the chosen time period, the model selects energy technologies based on their investment and running costs and operational parameters, so as to meet service demands while minimising the total cost of the system. TIAM-UCL is chosen for this study, as it is a partial equilibrium, bottom-up model. These characteristics allow modification of the demand as an exogenous input, while the model has sufficient technological detail to describe the technologies suited to meet that demand. Scenario analysis can be used to test the effects of features such as policies, technology changes, socio-economic pathways or climate change impacts. The world is represented as 16 regions, one of which is Africa (Anandarajah et al. 2011). 3.1 Heat stress indices There are multiple heat stress metrics available for analysis (Buzan et al. 2015), two of which have been selected for further analysis: apparent temperature (AT) (Steadman 1984) and humidex (HD) (Masterton and Richardson 1979; Buzan et al. 2015). Both indices are dependant on temperature and vapour pressure, the latter of which is dependent on surface pressure and specific humidity. The derivation of the vapour pressure and the two indices are Climatic Change (2019) 154:461–476 465 shown in SI Table 3. The heat stress indices are calculated on each grid cell (0.44◦× 0.44◦) with a temporal resolution of 1 day. Each of the heat stress indices has different ranges for slight, moderate, strong and extreme stress: these values were specified in Zhao et al. (2015) and are shown in SI Table 4. As discussed in SM DM1 of Zhao et al. (2015), HD is a comfort index which is used for both indoors and outdoors conditions. The AT indoor version is used in this work and the ranges of values span from just above indoor comfort (slight) to the point at which heat stroke can occur (extreme). For each GCM-RCM combination, the number of days with heat stress in each level were calculated and then averaged to produce the ensemble mean. 3.2 Heat stress-related cooling A reduction in ambient temperature reduces the value of the heat stress indices for a given water vapour pressure, and it is therefore possible with sufficient cooling to prevent a heat stress event. It is possible to calculate the difference in temperature between a heat stress day and a no stress day by rearranging the formulae in SI Table 3. The equations are combined with the lowest limit in SI Table 4 (28 for AT and 35 for HD) to produce an equation that defines the maximum possible temperature that combined with the fixed vapour pressure would not cause heat stress. The rearranged temperature limit calculations are shown in SI Table 5, where Ttarget is the target temperature at which any increase would cause heat stress for that value of vapour pressure. The difference between the simulated temperature and the target temperature is the cooling requirement. As part of this initial sensitivity study, the impact of changing temperature on the water vapour capacity of the air parcel has been neglected. This equation is solved for each day of the simulations and the results are summed to find the annual cooling requirement in cooling degree days (CDDs). The cooling required to mitigate heat stress under the baseline and future climate and population are calculated on a gridded basis across Africa by multiplying the corresponding gridded CDD results by the gridded SSP3 population data. 3.3 Energy demand for high stress mitigation Energy service demands are an exogenous input to the TIAM-UCL model over the mod- elled period. The 2005 energy demands for space cooling in the Africa commercial and residential sectors is 21.8 and 27.8 PJ respectively, representing some use of fans and air conditioning units. These figures are derived from the IEA World Energy Balances and expert judgement (Anandarajah et al. 2011). Subsequent energy service demands are calcu- lated for each modelled time slice (t) based on the previous timestep value (t-1) and growth in various socio-economic drivers modulated by corresponding elasticity parameters. These elasticities represent how closely the demands are coupled to the drivers and vary over the modelled time horizon. Their values are taken from the ETSAP-TIAM model and are cal- ibrated across the energy demands in the model to represent saturation of energy service use and convergence between developing and industrialized countries. The energy demands for residential cooling (RC) and commercial cooling (CC) are calculated with Eqs. 1 and 2, where N is the number of households, GDPH is the GDP per household, k is an elastic- ity parameter and PSER is the service industry activity (in trillion USD). These drivers are derived from the country-level population and GDP projections for SSP3, summed over Africa. RCDt = RCDt−1 × N × GDPHk (1) CCDt = CCDt−1 × PSERk (2) RCDt = RCDt−1 × N × GDPHk (1) CCDt = CCDt−1 × PSERk (2) (1) (2) ( ) (2) Climatic Change (2019) 154:461–476 466 For this analysis, it is assumed that all space cooling services in the model are used to miti- gate heat stress and so the total cooling demand is the sum of the cooling required to mitigate each heat stress level. As described by Brown et al. (2016), the relationship between temper- ature and cooling demand may be represented by different functions depending on factors including building properties, diffusion of cooling technologies and cultural preferences. Following the method of Labriet et al. (2015), this study assumes that the usage of cooling appliances changes but the diffusion of appliances does not change as a response to climate change. Based on this, the cooling demands are assumed to scale proportionally with the change in cooling degree days. 3.3 Energy demand for high stress mitigation Given these assumptions, to simulate the impact of heat stress mitigation on energy demand, the CDDs required to mitigate heat stress are calculated by summing the population-weighted apparent temperature CDD results (Section 3.2) over Africa. These are used to derive the scaling factor A (Eq. 3), which is the ratio of the population-weighted cooling degree days required to prevent all heat stress at the future and historic warming levels: CDDfuture and CDDhistoric. A = CDDfuture CDDhistoric (3) A = CDDfuture CDDhistoric (3) (3) 3.4 Country-level cooling demands The cooling demand required to mitigate heat stress for the whole of Africa is downscaled to give the cooling demand for each country. This is done by apportioning the continen- tal increase in energy demand between the countries according to the fraction of additional population-weighted CDDs occurring in each country. See Eqs. 4 and 5, where CDclimi,t is the additional cooling demand for country i at time t due to climate change, CDclimA,t is the additional cooling demand for Africa at time t due to climate change. PopCDDi,t−T is the change in population-weighted CDDs for country i between times T (2005) and t (2035 or 2076) and PopCDDA,t−T is the change in population-weighted CDDs for Africa between times T (2005) and t (2035 or 2076). CDA,t,CC and CDA,t,noCC are the cooling demands for Africa at time t with and without climate change respectively. As an initial high-level indication of the challenge these additional electricity demands present to each country, the results are also presented per unit of GDP per capita. A schematic of the approach linking TIAM-UCL to the heat stress data is shown in SI Fig. 1. CDclimi,t = CDclimA,t × PopCDDi,t−T PopCDDA,t−T (4) CDclimA,t = CDA,t,CC −CDA,t,noCC (5) CDclimi,t = CDclimA,t × PopCDDi,t−T PopCDDA,t−T (4) CDclimA,t = CDA,t,CC −CDA,t,noCC (5) (4) (5) 4.1 Heat stress results Figure 1 shows the impact of a two- and four-degree warming on apparent temperature heat stress occurrence across Africa; the humidex results are shown in SI Figure 2. For both indices, the same pattern emerges: either the number of slight heat stress events increases, indicating that regions previously unaffected are now vulnerable, or where the number of slight events decreases the number of moderate, and in the case of apparent temperature strong, heat stress events increases. When global temperatures are 4 ◦C above the historic average (Fig. 1, SI Figure 2), the heat stress changes follow a similar geo- graphic pattern to a + 2 ◦C change but with a higher intensity. This higher intensity is also accompanied by an increase in the number of strong heat stress events using the humidex index. The standard deviation of the number of days in each heat stress level were calculated and the difference between the historic and future standard deviations are shown in SI Figures 3 and 4. The largest increase in variability is around central Africa where the number of heat stress days moving from slight to moderate varies more across the GCM-RCM pairings. The number of grid cells exposed to extreme heat stress increases significantly when temperatures are 4 ◦C above the pre-industrial limits. In general, the distribution of heat stress events increases in severity with climate change. The increase in moderate and stronger events affects highly populated coastlines and West African Sahelian nations. The same response is found in Central Africa over much of the Central African Republic and the Democratic Republic of the Congo. The equatorial and tropical regions show higher instances of heat stress with changes in strong heat stress events almost exclusively in the Northern Hemisphere. The highly populated West African region experiences the largest changes in moderate and strong heat stress events. 3.5 Costs of heat stress mitigation To calculate the costs of transforming the energy system to meet these increased demands, the TIAM-UCL model is run for the period 2005–2100, the extended time horizon reflecting the fact that policymakers and industry agents make decisions to prepare for, and which affect, the years and decades to come. In TIAM-UCL, the Africa region is split into rural (1) and urban (2) sub-regions for which the residential demand drivers and elasticities are defined separately. This represents the significant differences in access to energy supply and service technologies between the urban and rural populations. The commercial demand is not split into urban and rural regions. The model is run with two scenarios: the base case and the climate change case. In the base case, the cooling demands are only functions of 467 Climatic Change (2019) 154:461–476 the growth in the socio-economic drivers, as described above (Eqs. 1 and 2) and are not affected by any change from the present climate condition. In the climate change case, they are also scaled by the factor A (Eq. 3) to account for heat stress under a warmer climate in the future. Thus, the cooling demands in the two scenarios are equal for the base year 2005, then diverge through future years. the growth in the socio-economic drivers, as described above (Eqs. 1 and 2) and are not affected by any change from the present climate condition. In the climate change case, they are also scaled by the factor A (Eq. 3) to account for heat stress under a warmer climate in the future. Thus, the cooling demands in the two scenarios are equal for the base year 2005, then diverge through future years. 4.2 Increase in cooling demand The difference in the cooling requirement to prevent heat stress between the historic and future simulations are shown in Fig. 2. The variation in results from the 11 model combinations was examined. The standard deviation in the cooling required to prevent heat stress is shown in SI Figure 5. The spread in the cooling required is greatest in central Africa; this is the result of the variability in the monsoon systems in the input meteorology. The cooling demands taking into account future climate change are cal- culated for the time slices at + 2 and + 4 ◦C and are linearly interpolated to define the demand pathways. SI Figure 6 shows the cooling demands for the base and cli- mate change cases. In 2005, space cooling accounts for 0.4% of the total final energy demand in Africa. In the base case, this drops to 0.2% by 2100 but in the climate change case space cooling rises to account for 1.2% of the total final energy demand by 2100. Climatic Change (2019) 154:461–476 468 468 Climatic Change (2019) 154:461 476 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 Fig. 4.2 Increase in cooling demand 1 Difference in apparent temperature events per year between the + 2 ◦C (left) and + 4 ◦C (right) and baseline time slices for slight (first row), moderate (second row), strong (third row) and extreme (fourth row) levels 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 Longitude 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 Longitude g 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 0 20 40 Longitude -20 0 20 -300 -200 -100 0 100 200 300 Longitude Fig. 1 Difference in apparent temperature events per year between the + 2 ◦C (left) and + 4 ◦C (right) and baseline time slices for slight (first row), moderate (second row), strong (third row) and extreme (fourth row) levels Climatic Change (2019) 154:461–476 469 Longitude Latitude 0 10 20 30 40 50 0 -10 -10 -20 -30 10 20 30 Degree Days 0 500 -500 -1000 1000 Longitude Latitude 0 10 20 30 40 50 0 -10 -10 Longitude 0 10 20 30 40 50 -10 -20 -30 10 20 30 0 -10 -20 -30 10 20 30 Degree Days 0 500 -500 -1000 1000 Latitude Degree Days 0 500 -500 -1000 -1500 -2000 -2500 1000 1500 2000 2500 Fig. 2 Difference in cooling degree days required to prevent all apparent temperature heat stress between + 2 ◦C (left), + 4 ◦C (right) and baseline conditions Longitude 0 10 20 30 40 50 -10 0 -10 -20 -30 10 20 30 Latitude Degree Days 0 500 -500 -1000 -1500 -2000 -2500 1000 1500 2000 2500 Fig. 2 Difference in cooling degree days required to prevent all apparent temperature heat stress between + 2 ◦C (left), + 4 ◦C (right) and baseline conditions 4.3 Country-level demand Figure 3 shows the additional cooling demand that would occur in each country under + 2 and + 4 ◦C warming. Due to the significant increase in heat stress and high population density, the greatest increases in cooling demand occurs in Nigeria, followed by the Demo- cratic Republic of Congo, Egypt and Sudan. However, supplying these additional electricity demands will present different challenges to each country, depending on factors such as the 0 20 40 60 80 100 120 140 160 180 200 PJ Zimbabwe Zambia United Republic of Tanzania Uganda Tunisia Togo Swaziland Sudan South Africa Somalia Sierra Leone Senegal Rwanda Republic of Congo Nigeria Niger Namibia Mozambique Morocco Mauritania Mali Malawi Madagascar Libya Liberia Lesotho Kenya Ivory Coast Guinea Bissau Guinea Ghana Gambia Gabon Ethiopia Eritrea Equatorial Guinea Egypt Djibouti Democratic Republic of the Congo Chad Central African Republic Cameroon Burundi Burkina Faso Botswana Benin Angola Algeria +2C +4C Fig. 3 Country-level additional cooling demand due to heat stress mitigation under RCP8.5 for two timeslices Fig. 3 Country-level additional cooling demand due to heat stress mitigation under RCP8.5 for two timeslices 470 Climatic Change (2019) 154:461–476 existing energy system, the country’s wealth, and the style of governance and system plan- ning. As an initial high-level indicator of this, Fig. 4 shows the additional cooling demands per unit of GDP per capita. This suggests the additional cooling demand will present the greatest challenges to Somalia, Guinea Bissau, Djibouti and Central African Republic, par- ticularly at the lower warming level (2 ◦C in 2035), as this occurs before the large increases in GDP and population. 4.4 Impact of increased cooling energy demand TIAM-UCL results for the base case and the case with heat stress cooling demand (climate change case) are compared. SI Figure 7 shows the impact of the increased residential and commercial cooling demand on the final energy consumption of all the economic sectors. The changes in the residential and commercial sectors induce changes in the generation technology and fuel mix in other sectors; for example, final energy consumption in the trans- port sector decreases slightly in the middle of the century when gas is partially replaced with electricity, and energy consumption in the upstream sector decreases around 2060 and 2080 as the use of electricity and heat respectively are reduced. The total energy consumption of the Africa region is increased by 0.1% in 2035 and 0.3% in 2076 due to the additional cooling to mitigate heat stress. The increased cooling demands in the climate change case do not affect the technologies selected by TIAM-UCL to fulfil these end demands. From 2020 onwards in both scenarios, commercial cooling is entirely provided by rooftop chiller units and residential cooling is entirely provided by room air conditioning units. SI Figure 8 shows the mix of electricity generation technologies in the base case. The share of renewable energy sources—solar PV and hydroelectricity—grows rapidly from the 0 10 20 30 40 50 60 70 80 J/$/person Zimbabwe Zambia United Republic of Tanzania Uganda Tunisia Togo Swaziland Sudan South Africa Somalia Sierra Leone Senegal Rwanda Republic of Congo Nigeria Niger Namibia Mozambique Morocco Mauritania Mali Malawi Madagascar Libya Liberia Lesotho Kenya Ivory Coast Guinea Bissau Guinea Ghana Gambia Gabon Ethiopia Eritrea Equatorial Guinea Egypt Djibouti Democratic Republic of the Congo Chad Central African Republic Cameroon Burundi Burkina Faso Botswana Benin Angola Algeria +2C +4C Fig. 4 Additional energy demand per unit of GDP per capita under RCP8.5 for two timeslices Fig. 4 Additional energy demand per unit of GDP per capita under RCP8.5 for two timeslices Climatic Change (2019) 154:461–476 471 middle of the century. To provide the additional electricity required in the climate change case, approximately 3% additional generation is required from 2040 onwards. Of this addi- tional generation, solar PV provides an increasing share rising from 5% in 2045 to 15% in 2070. This mainly replaces coal in the technology mix, while the share of natural gas remains approximately constant around 3%. middle of the century. 4.4 Impact of increased cooling energy demand To provide the additional electricity required in the climate change case, approximately 3% additional generation is required from 2040 onwards. Of this addi- tional generation, solar PV provides an increasing share rising from 5% in 2045 to 15% in 2070. This mainly replaces coal in the technology mix, while the share of natural gas remains approximately constant around 3%. TIAM-UCL models the transformation of the energy system through the modelling period. The total cost of the regional energy system over that period is composed of expenditures (such as capital required for building and decommissioning technologies, oper- ations and maintenance, fuel costs, commodity imports, commodity delivery and taxes) and revenues (from commodity exports, subsidies and the salvage values of decommissioned technologies). In reality, these costs are paid and recouped by a number of different par- ties over various timescales. As several of these finance streams occur earlier or later than the energy generation which they facilitate, TIAM-UCL includes mechanisms to annualise them over the period of generation (Loulou et al. 2016). As the energy supply in a given year is made possible by the construction and operation of supply technologies in previous years, the energy system cost (cost (ES)) relevant for this analysis is defined as the sum of the annual investment and operations costs from the model base year up to the year in question. These cumulative costs are all reported in year 2005 US $ and are undiscounted. Table 1 shows the Africa cumulative energy system costs for the base case and climate change case. Due to the increased cooling demand required to mitigate heat stress resulting from the + 2 ◦C global average temperature rise, the integrated energy system cost from 2005 to 2035 increases by 0.26%, which is equivalent to approximately $51.3 billion. With + 4 ◦C global average temperature rise, the integrated energy system cost from 2005 to 2076 increases by 0.60%, which is equivalent to approximately $486.5 billion. 5 Discussion Over the 10-year periods centred on 2035 (+ 2 ◦C) and 2076 (+ 4 ◦C), the additional cost required to mitigate heat stress corresponds to an average additional cost of $2.29 and $5.52 per person per year respectively. To put this in context, between 2000 and 2013 investment in the power sector in Africa accounted for on average $12 billion per year (International Energy Agency 2014), which is equivalent to approximately $12 per person per year. The estimated cumulative cost of $51.3 billion to prevent heat stress in Africa between 2005 and 2035 can be compared with the total infrastructure costs estimated in Parry et al. (2009), where the removal of infrastructure deficit and adapting infrastructure cost is esti- mated to be between $64.7 billion and $73.9 billion. The results in Parry et al. (2009) employed the A1B emissions scenarios from IPCC Assessment Report 4, which is a mod- erate emissions scenario and analogous to RCP4.5. This is a lower emissions scenario than used in our study; however, the differences are not large by 2030. Therefore, the results herein, which focus solely on the effect of increasing cooling demand on the energy system, indicate that the total costs in Parry et al. (2009) may be underestimated. The input data used in this study were based on bias corrected CORDEX-Africa simu- lations that were driven by RCP 8.5 simulations from CMIP5. RCP8.5 is a high emission scenario with limited adaptation and mitigation of climate change (Meinshausen et al. 2011). In the cases of RCP2.6, RCP4.5 or RCP6.0, the expected temperature change is lower; this would results in a lower value for the heat stress metric and fewer events that exceed the predefined limits. A lower number of heat stress events will result in lower costs for adapting the energy system to provide sufficient electricity to prevent heat stress. The highly non-linear relationship between climate scenario and costs however make it impossible to provide a numeric value for the energy system costs in lower emission scenarios. This study has some key limitations, which should be addressed in further work. First, in the heat stress calculation, the impact of changing temperature on the water capacity of the air parcel has been neglected. In the economic modelling, the climate change effect of heat stress is applied to the Africa region and cooling demand only. 5 Discussion Across the entire African continent, there is a projected increase in the total number of heat stress events. In regions where the number of slight or moderate heat stress events falls, there is an increase in the number of more severe events. The increase in overall intensity of heat stress happens in highly populated regions such as the Nigerian coast, the Great Lakes and along several major rivers including the Nile, Volta, Niger and Zambezi. Heat stress has a more profound effect on the vulnerable members of society such as those living alone, the young or the elderly (Uejio et al. 2011) or those will chronic health problems such as heart conditions or mental health issues (Zhang et al. 2017); therefore, expected increases in heat stress should be noted by social and health services in these regions. The cooling demand is relatively consistent across the 11 GCM-RCM combinations; however, as the GCMs attain the SWLs at different times, the variability in population is large. Table 1 Energy system cost changes associated with preventing heat stress between the historic period and the + 2 and + 4 ◦C timeslices Period 2005–2035 2005–2076 Temperature increase + 2 ◦C + 4 ◦C Cumulative additional cost (billion USD) $51.327 $486.526 Energy system cost rise over base case 0.26% 0.60% Cost rise as fraction of GDP 0.03% 0.06% Energy system cost changes associated with preventing heat stress between the historic period and and + 4 ◦C timeslices 472 Climatic Change (2019) 154:461–476 The apparent temperature heat stress results were used to scale the cooling service energy demand for the Africa region in the TIAM-UCL model. Country-level analysis showed the greatest demand increases in the latter part of this century in countries with the highest warming levels and population density. However, it is expected that the greatest challenges will be presented earlier in the century to countries where the increase in heat stress is high relative to the level of economic development. Meeting the additional energy demand had the effect of increasing the energy system cost over the 2005–2076 period by 0.6%, which corresponds to 0.06% of the cumulative GDP for Africa over the same period. The additional yearly cost increases with time (as global warming increases). Although the population of Africa is projected to increase with time, the cost per person also increases. 5 Discussion Clearly, climate change will also affect other regions and elements of the energy system, such as heating demand and renewable resources, whose combined impacts could affect the technology choices made by TIAM-UCL in individual regions, commodity trading between regions, and thus the regional and global mitigation costs. The urban heat island (UHI) effect can raise the temperature of cities by several degrees in comparison with surrounding rural areas (Tran et al. 2006; Oke 1982). The analysis conducted in TIAM-UCL is performed using the multi-model ensemble mean levels of cooling required to prevent heat stress. The range of values used to generate this ensemble mean is dependent on the climate sensitivity of the GCM as this determines when the SWLs are reached. The timing of reaching the SWLs contributes to the total CDD Climatic Change (2019) 154:461–476 473 requirement primarily as a result of population change with lower climate sensitivity leading to a later SWL and therefore a higher population. As the feedbacks within TIAM-UCL are non-linear, further research into the range of cooling requirements is an area of future study. requirement primarily as a result of population change with lower climate sensitivity leading to a later SWL and therefore a higher population. As the feedbacks within TIAM-UCL are non-linear, further research into the range of cooling requirements is an area of future study. The CMIP5 GCMs and CORDEX-Africa RCMs do not simulate the UHI effect; there- fore, it is not possible to analyse it directly. Analysis of the UHI is often performed using high-resolution models or nested domains such as in Vahmani et al. (2016). With the increas- ing urbanisation of Africa (Cohen 2006; Parnell and Walawege 2011), the fraction of the population exposed to the urban heat island effect will increase. Therefore, the actual cool- ing demand and the associated costs are likely to be higher than the results presented here. This analysis should therefore be considered a sensitivity study and a demonstration of the use of technical (rather than economic) climate change feedback in the TIAM-UCL model to evaluate adaptation costs. Further work should focus on including further impacts of climate change on energy demand and supply in the TIAM-UCL model and adopting a global approach. The country-level analysis of heat stress and cooling demand also suggest energy system modelling and infrastructure planning for several countries should account for climate change impacts on demand. References Anandarajah G, Pye S, Usher W, Kesicki F, Mcglade C (2011) TIAM-UCL global model documentation. Tech. rep., University College London Asefi-Najafabady S, Vandecar KL, Seimon A, Lawrence P, Lawrence D (2018) Climate change, population, and poverty: vulnerability and exposure to heat stress in countries bordering the Great Lakes of Africa. Clim Chang 148(4):561–573. https://doi.org/10.1007/s10584-018-2211-5 Barbier J, Guichard F, Bouniol D, Couvreux F, Roehrig R (2018) Detection of intraseasonal large-scale heat waves: characteristics and historical trends during the Sahelian spring. J Clim 31(1):61–80. https://doi.org/10.1175/JCLI-D-17-0244.1 Brown MA, Cox M, Staver B, Baer P (2016) Modeling climate-driven changes in U.S. buildings energy demand. Clim Chang 134(1):29–44. https://doi.org/10.1007/s10584-015-1527-7 Burke M, Hsiang SM, Miguel E (2015) Global non-linear effect of temperature on economic production. Nature 527(7577):235–239 Buzan JR, Oleson K, Huber M (2015) Implementation and comparison of a suite of heat stress metrics within the Community Land Model version 4.5. Geosci Model Dev 8(2):151–170. https://doi.org/10.5194/gmd-8-151-2015 Cohen B (2006) Urbanization in developing countries: current trends, future projections, and key challenges for sustainability. Technol Soc 28(1):63–80. https://doi.org/10.1016/j.techsoc.2005.10.005. http://www. sciencedirect.com/science/article/pii/S0160791X05000588, sustainable Cities p , Cronin J, Anandarajah G, Dessens O (2018) Climate change impacts on the energy system: a review of trends and gaps. Climatic Change. https://doi.org/10.1007/s10584-018-2265-4 Davis LW, Gertler PJ (2015) Contribution of air conditioning adoption to future energy use under global warming. Proc Natl Acad Sci 112(19):5962–5967. https://doi.org/10.1073/pnas.1423558112. http:// www.pnas.org/content/112/19/5962 p g De Cian E, Sue Wing I (2017) Global energy consumption in a warming climate. Environmental and Resource Economics. https://doi.org/10.1007/s10640-017-0198-4 Diffenbaugh NS, Pal JS, Giorgi F, Gao X (2007) Heat stress intensification in the Mediterranean climate change hotspot. Geophys Res Lett 34(11):n/a–n/a. https://doi.org/10.1029/2007GL030000, l11706 D JP St ff RJ J h JG (2013) R d ti i l b it f h t t d li t i g , , g , ( ) change hotspot. Geophys Res Lett 34(11):n/a–n/a. https://doi.org/10.1029/2007GL030000, l11706 Dunne JP, Stouffer RJ, John JG (2013) Reductions in labour capacity from heat stress under climate warming. 6 Conclusion Heat stress is an existing problem that affects many countries in Africa, with the high- est intensities present in the tropics. Exposure to heat stress is projected to increase into the twenty-first century with increasing frequency of high-intensity events. In the highly populated Sahelian region, strong and extreme events are projected to become much more regular. The energy system will face extra demand as populations increase and urbanise. The combination of extra heat stress and larger populations mean that the energy system will need significant investment to prevent electricity grid failures. The results herein indicate that the Democratic Republic of the Congo, Egypt, Nigeria and Sudan are projected to have the highest costs incurred in adapting the energy system to prevent heat stress. The financial impacts are different when spread over the populations, Nigeria is relatively rich and highly populous and therefore faces a small individual cost. In contrast, the nations of the Central African Republic, Djibouti, Guinea Bissau and Somalia have the highest cost per unit of GDP per capita. The energy system however cannot be treated in isolation and further research should be conducted at a national or sub-national scale to asses the impact of heat stress. Heat stress is associated with heatwaves; there is potential that there will be competition between electricity generation, drinking water demand and irrigation for limited water. With the signing of the Paris accord (United Nations 2015), the global community agreed to work to prevent the catastrophic climate changes that are projected in RCP8.5. However, strong action to reduce carbon dioxide emissions is necessary to reach the agreed goals. This study indicates that under a high climate change scenario, infrastructure costs for adaptation may be larger than previous estimated. This should be considered by energy system planners and policymakers when balancing mitigation and adaptation options. With the long lead time of energy systems, often measured in decades, decision-makers will be faced with choosing between increasing capacity to prevent potential heat stress or being unable to provide energy when the need is critical. Acknowledgements The authors would also like to thank the reviewers for their insightful and constructive comments. 474 Climatic Change (2019) 154:461–476 Funding information The research leading to these results has received funding from the European Union Seventh Framework Programme FP7/2007-2013 under grant agreement n◦603864 (HELIX: High-End cLi- mate Impacts and eXtremes; http://www.helixclimate.eu). 6 Conclusion This research was made possible by support from the EPSRC as a Standard Research Studentship (DTP), grant no. EP/M507970/1. Dr Ben Parkes is the recip- ient of the Ekpe Research Impacts Fellowship at the School of Mechanical, Aerospace and Civil Engineering at The University of Manchester. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 Inter- national License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. References http://www.iea.org/publicatio freepublications/publication/WEIO2014.pdf p p p Isaac M, van Vuuren DP (2009) Modeling global residential sector energy demand for heating and air condi- tioning in the context of climate change. Energy Policy 37(2):507–521. https://doi.org/10.1016/j.enpol. 2008.09.051 Jones B, O’Neill BC (2016) Spatially explicit global population scenarios consistent with the shared soci conomic pathways. Environ Res Lett 11(8):084,003. http://stacks.iop.org/1748-9326/11/i=8/a=0840 Kjellstrom T, Holmer I, Lemke B (2009) Workplace heat stress, health and productivity - an increasing challenge for low and middle-income countries during climate change. Glob Health Action 2(1):2047. https://doi.org/10.3402/gha.v2i0.2047 p g g Labriet M, Joshi SR, Vielle M, Holden PB, Edwards NR, Kanudia A, Loulou R, Babonneau F (2015) World- wide impacts of climate change on energy for heating and cooling. Mitig Adapt Strateg Glob Chang 20(7):1111–1136. https://doi.org/10.1007/s11027-013-9522-7 ( ) p g Loulou R, Goldstein G, Kamudia A, Lettila A, Remme U (2016) Documentation for the TIMES Model Part I, ETSAP https://iea-ieaetsap.org/docs/Documentation for the TIMES Model-Part-I July-2016.pdf Loulou R, Goldstein G, Kamudia A, Lettila A, Remme U (2016) Documentation for the TIMES Model Part I, ETSAP https://iea-ieaetsap.org/docs/Documentation for the TIMES Model-Part-I July-2016.pdf Lundgren K, Kuklane K, Chuansi GAO, Holm´er I (2013) Effects of heat stress on working populatio facing climate change. Ind Health 51(1):3–15. https://doi.org/10.2486/indhealth.2012-0089 g g p g Masterton J, Richardson F (1979) Humidex: a method of quantifying human discomfort due to exces heat and humidity. 28cm. cli,1, Ministere de l’Environnement Medlock K, Soligo R (2001) Economic development and end-use energy demand. Energy J 22(Number 2):77–105. https://EconPapers.repec.org/RePEc:aen:journl:2001v22-02-a04 Meinshausen M, Smith S, Calvin K, Daniel J, Kainuma M, Lamarque JF, Matsumoto K, Montzka S, Raper S, Riahi K, Thomson A, Velders G, Vuuren D (2011) The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Clim Chang 109(1-2):213–241. https://doi.org/10.1007/s10584-011- 0156-z Mitchell D, Heaviside C, Schaller N, Allen M, Ebi KL, Fischer EM, Gasparrini A, Harrington L, Kharin V, Shiogama H, Sillmann J, Sippel S, Vardoulakis S (2018) Extreme heat-related mortality avoided under Paris agreement goals. Nat Clim Chang 8(7):551–553. https://doi.org/10.1038/s41558-018-0210-1 Mora C, Dousset B, Caldwell IR, Powell FE, Geronimo RC, Bielecki C, Counsell CWW, Dietrich BS, Johnston ET, Louis LV, Lucas MP, McKenzie MM, Shea AG, Tseng H, Giambelluca TW, Leon LR, Hawkins E, Trauernicht C (2017) Global risk of deadly heat. Nat Clim Chang 7:501. https://doi.org/10.1038/nclimate3322 Murakami D, Yamagata Y (2016) Estimation of gridded population and gdp scenarios with spatially explicit statistical downscaling. References Nature Clim Change 3(6):563–566 Dunne JP, Stouffer RJ, John JG (2013) Reductions in labour capacity from heat stress under climate warmi Nature Clim Change 3(6):563–566 Field CB, Barros VR, Mach KJ, Mastrandrea MD, Mv Aalst, Adger WN, Arent DJ, Barnett J, Betts R, Bilir TE, Birkmann J, Carmin J, Chadee DD, Challinor AJ, Chatterjee M, Cramer W, Davidson DJ, Estrada YO, Gattuso JP, Hijioka Y, Hoegh-Guldberg O, Huang HQ, Insarov GE, Jones RN, Kovats RS, Lankao PR, Larsen JN, Losada IJ, Marengo JA, McLean RF, Mearns LO, Mechler R, Morton JF, Niang I, Oki T, Olwoch JM, Opondo M, Poloczanska ES, P¨ortner HO, Redsteer MH, Reisinger A, Revi A, Schmidt DN, Shaw MR, Solecki W, Stone DA, Stone JMR, Strzepek KM, Suarez AG, Tschakert P, Valentini R, Vicu˜na S, Villamizar A, Vincent KE, Warren R, White LL, Wilbanks TJ, Wong PP, Yohe GW (2014) Technical summary. In: Field CB, Barros VR, Dokken DJ, Mach KJ, Mastrandrea MD, Bilir TE, Chatterjee M, Ebi KL, Estrada YO, Genova RC, Girma B, Kissel ES, Levy AN, MacCracken S, Mastrandrea PR, White LL (eds) Climate change 2014: impacts, adaptation, and vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, pp 35–94 Lankao PR, Larsen JN, Losada IJ, Marengo JA, McLean RF, Mearns LO, Mechler R, Morton JF, Niang I, Oki T, Olwoch JM, Opondo M, Poloczanska ES, P¨ortner HO, Redsteer MH, Reisinger A, Revi A, Schmidt DN, Shaw MR, Solecki W, Stone DA, Stone JMR, Strzepek KM, Suarez AG, Tschakert P, , j , , , , , , y , , Mastrandrea PR, White LL (eds) Climate change 2014: impacts, adaptation, and vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the p g p p Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, pp 35–94 her EM, Oleson KW, Lawrence DM (2012) Contrasting urban and rural heat stress responses to climate change. Geophys Res Lett 39(3):n/a–n/a. https://doi.org/10.1029/2011GL050576, l03705 Climatic Change (2019) 154:461–476 475 Grillakis MG, Koutroulis AG, Tsanis IK (2013) Multisegment statistical bias correction of daily GC precipitation output. J Geophys Res Atmos 118(8):3150–3162. https://doi.org/10.1002/jgrd.50323 International Energy Agency (2014) World energy investment outlook. References arXiv:1610.09041 Nejat P, Jomehzadeh F, Taheri MM, Gohari M, Abd Majid MZ (2015) A global review of energy con- sumption, CO2 emissions and policy in the residential sector (with an overview of the top ten CO2 emitting countries). Renew Sust Energ Rev 43(C):843–862. https://doi.org/10.1016/j.rser.2014.11.09. https://ideas.repec.org/a/eee/rensus/v43y2015icp843-862.html p p g y p Nikulin G, Jones C, Giorgi F, Asrar G, B¨uchner M, Cerezo-Mota R, Christensen OB, D´equ´e M, Fernandez J, H¨ansler A, van Meijgaard E, Samuelsson P, Sylla MB, Sushama L (2012) Precipitation climatol- ogy in an ensemble of CORDEX-Africa regional climate simulations. J Climate 25(18):6057–6078. https://doi.org/10.1175/JCLI-D-11-00375.1 p g Oke TR (1982) The energetic basis of the urban heat island. Q J R Meteorol Soc 108(455):1–24. https://doi.org/10.1002/qj.49710845502 p g qj O‘Neill BC, Carter TR, Ebi KL, Edmonds J, Hallegatte S, Kemp-Benedict E, Kriegler E, Mearns L, Moss R, Riahi K, van Ruijven B, van Vuuren D (2012) Meeting report of the workshop on the nature and use of new socioeconomic pathways for climate change research, Boulder, CO, November 2-4, 2011. Available at: http://www.isp.ucar.edu/socio-economic-pathways Papadimitriou LV, Koutroulis AG, Grillakis MG, Tsanis IK (2015) High-end climate change impact on European water availability and stress: exploring the presence of biases. Hydrol Earth Syst Sci Discuss 12(7):7267–7325. https://doi.org/10.5194/hessd-12-7267-2015 Parnell S, Walawege R (2011) Sub-saharan African urbanisation and global environmental change. Glob Environ Chang 21(Supplement 1):S12–S20. https://doi.org/10.1016/j.gloenvcha.2011.09.014, migration and Global Environmental Change - Review of Drivers of Migration Parry M, Arnell N, Berry P, Dodman D, Fankhauser S, Hope C, Kovats S, Nicholls R, Satterthwaite D, Tiffin R, Wheeler T (2009) Assessing the costs of adaptation to climate change: a review of the UNFCCC and other recent estimates. Tech rep. References International Institute for Environment and Development and Grantham Institute for Climate Change, London Climatic Change (2019) 154:461–476 476 Riahi K, van Vuuren DP, Kriegler E, Edmonds J, O‘Neill BC, Fujimori S, Bauer N, Calvin K, Dellink R, Fricko O, Lutz W, Popp A, Cuaresma JC, Samir KC, Leimbach M, Jiang L, Kram T, Rao S, Emmerling J, Ebi K, Hasegawa T, Havlik P, Humpen¨oder F, Silva LAD, Smith S, Stehfest E, Bosetti V, Eom J, Gernaat D, Masui T, Rogelj J, Strefler J, Drouet L, Krey V, Luderer G, Harmsen M, Takahashi K, Baumstark L, Doelman JC, Kainuma M, Klimont Z, Marangoni G, Lotze-Campen H, Obersteiner M, Tabeau A, Tavoni M (2017) The shared socioeconomic pathways and their energy, land use, and greenhouse gas emissions implications: an overview. Glob Environ Chang 42:153–168. https://doi.org/10.1016/j.gloenvcha.2016.05.009 Steadman RG (1984) A universal scale of apparent temperature. J Climate Appl Meteor 23(12):1674–1687. https://doi.org/10.1175/1520-0450(1984)023<1674:AUSOAT>2.0.CO;2 p g Suzuki-Parker A, Kusaka H (2016) Future projections of labor hours based on WBGT for Tokyo and Osaka, Japan, using multi-period ensemble dynamical downscale simulations. Int J Biometeorol 60(2):307–310 p g Suzuki-Parker A, Kusaka H (2016) Future projections of labor hours based on WBGT for Tokyo and Osaka, Japan, using multi-period ensemble dynamical downscale simulations. Int J Biometeorol 60(2):307–310 Taylor KE, Stouffer RJ, Meehl GA (2011) An overview of CMIP5 and the experiment design. Bull Amer M t S 93(4) 485 498 htt //d i /10 1175/BAMS D 11 00094 1 Taylor KE, Stouffer RJ, Meehl GA (2011) An overview of CMIP5 and the experiment design. Bull Am Meteor Soc 93(4):485–498. https://doi.org/10.1175/BAMS-D-11-00094.1 Tran H, Uchihama D, Ochi S, Yasuoka Y (2006) Assessment with satellite data of the urban heat island effects in Asian mega cities. Int J Appl Earth Obs Geoinf 8(1):34–48. https://doi.org/10.1016/j.jag.2005.05.003 Uejio CK, Wilhelmi OV, Golden JS, Mills DM, Gulino SP, Samenow JP (2011) Intra-urban societal vul- nerability to extreme heat: the role of heat exposure and the built environment, socioeconomics, and neighborhood stability. Health Place 17(2):498–507. https://doi.org/10.1016/j.healthplace.2010.12.005, geographies of Care g g p United Nations (2015) Adoption of the Paris agreement. In: 21st Conference of the Parties, Paris, United Nations Vahmani P, Sun F, Hall A, Ban-Weiss G (2016) Investigating the climate impacts of urbanization and the potential for cool roofs to counter future climate change in Southern California. Environ Res Lett 11(12):124,027. References http://stacks.iop.org/1748-9326/11/i=12/a=124027 Zhang Y, Nitschke M, Krackowizer A, Dear K, Pisaniello D, Weinstein P, Tucker G, Shakib S, Bi P (2017) Risk factors for deaths during the 2009 heat wave in Adelaide, Australia: a matched case-control study. Int J Biometeorol 61(1):35–47. https://doi.org/10.1007/s00484-016-1189-9 Zhao Y, Ducharne A, Sultan B, Braconnot P, Vautard R (2015) Estimating heat stress from climate-based indicators: present-day biases and future spreads in the CMIP5 global climate model ensemble. Environ Res Lett 10(8):084–013 Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
https://openalex.org/W4320070352	https://ejurnal.seminar-id.com/index.php/bits/article/download/1639/1149	Indonesian	null	Analisis Perbandingan Algoritma NGG dan GGHN pada Frekuensi Hasil Enkripsi	Building of Informatics, Technology and Science	2,022	cc-by	6,422	Analisis Perbandingan Algoritma NGG dan GGHN pada Frekuensi Hasil Enkripsi Farid Akbar Siregar1, Ade Rizka2, Annisa Fadillah Siregar3,* g g 1Ilmu Komputer dan Teknologi Informasi, Teknologi Informasi, Universitas Muhammadiyah Sumatera Utara, Medan, Indonesia 2Sains dan Teknologi, Sistem Komputer, Universitas Pembangunan Panca Budi, Medan, Indonesia 3Ilmu Komputer dan Teknologi Informasi, Teknik Informatika, Universitas Budi Dharma, Medan, Indonesia Email: 1faridakbar@umsu.ac.id, 2aderizka@dosen.pancabudi.ac.id, 3,*annisaf@univ-bd.ac.id Email Penulis Korespondensi: annisaf@univ-bd.ac.id Submitted: 06/06/2022; Accepted: 28/06/2022; Published: 30/06/2022 Abstrak−Data dan informasi dalam perkembangan teknologi digital memiliki peranan penting. Setiap kegiatan maupun aktifitas yang menggunakan teknologi digital berkaitan dengan data dan informasi, sehingga keamanan informasi maupun kerahasiaan data sangat penting. Untuk menjaga keamanan informasi dan kerahasiaan data dibutuhkan perlindungan dengan teknik kriptografi. Teknik kriptografi berhubungan dengan enkripsi yaitu dimana dilakukan proses pengacakan data dan menyembunyikan data dengan sistem kunci sedangkan dekripsi yaitu proses perubahan kondisi data ke bentuk aslinya agar mudah dipahami. Terdapat kendala maupun masalah dalam komunikasi digital, sehingga diperlukan teknik kriptografi yang memiliki tingkat keamanan yang lebih dan dapat diterapkan dalam komunikasi digital. Untuk mengetahui tingkat keamanan dalam teknik kriptografi diperlukan analisis frekuensi. Analisis frekuensi pada algoritma NGG dan GGHN dilakukan untuk mengetahui tingkat keamanan informasi berdasarkan hasil enkripsi data. Berdasarkan proses pengujian pada algoritma NGG dan GGHN akan diketahui frekuensi karakter pada teks yang bervariasi. Semakin banyak karakter yang digunakan pada kunci akan mempengaruhi tingkat keamanan informasi. Algoritma NGG memiliki tingkat keamanan yang lebih tinggi dibandingkan dengan algortima GGHN dengan selisih presentasi sebesar 0.000299967%. Jika frekuensi kemunculan karakter pada teks pesan yang telah di enkripsi semakin sering atau semakin tinggi, maka tingkat keamanan informasi pada pesan lebih rendah dan kata kunci lebih mudah dipecahkan. Kata Kunci: Kriptografi; NGG; GGHN; Enkripsi; Dekripsi Kata Kunci: Kriptografi; NGG; GGHN; Enkripsi; Dekripsi Abstract−Data and information in the development of digital technology have an important role. Every activity or activity that uses digital technology is related to data and information, so information security and data confidentiality are very important. To maintain information security and data confidentiality, protection with cryptographic techniques is needed. Cryptographic techniques are related to encryption, which is where the process of scrambling data is carried out and hiding data with a key system, while decryption is the process of changing the condition of the data to its original form so that it is easy to understand. There are obstacles and problems ini digital communication, so cryptographic techniques are needed that have higher level of security and can be applied in digital communications. Analisis Perbandingan Algoritma NGG dan GGHN pada Frekuensi Hasil Enkripsi To determine the level of security in cryptographic techniques required frequency anlysis. Frequency analysis on the NGG and GGHN algorithms is carried out to determine the level of information security based on the results of data encryption. BAse on the testing process on the NGG and GGHN algorithms, it will be known that the frequency of characters in the text varies. The more characters used in the key will affect the level of information security. The NGG algorithm has a higher level of security than the GGHN algorithm with a precentage difference of 0.000299967%. If the frequency of occurrence of characters in the message text that has been encrypted is more frequent or higher, then the level of information security in the message is lower and the password is easier to crack. Keywords: Cryptography; NGG; GGHN; Encryption; Decryption Keywords: Cryptography; NGG; GGHN; Encryption; Decryption Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Copyright © 2022 Farid Akbar Siregar, Page 303 BITS is licensed under a Creative Commons Attribution 4.0 International License 1. PENDAHULUAN Analisis frekuensi pada kedua algoritma akan menghasilkan nilai kinerja, sehingga dalam proses transaksi dan distribusi data maupun informasi dapat ditentukan algoritma yang memiliki kemampuan dan tingkat keamanan yang lebih dibutuhkan. Pada penelitian ini akan menganalisis frekuensi kedua algoritma tersebut. Hasil proses pengujian dapat dilihat dari jumlah kemunculan karakter yang sama pada proses enkripsi. Diharapkan dapat meningkatkan keamanan informasi dalam teknologi digital. Hal pertama yang tercatat dari teknik kriptografi secara harfiah ditulis di batu hampir empat milenium yang lalu oleh seorang juru tulis Mesir dengan menggunakan subtitusi simbol hieroglyphic dalam tulisannya di dinding batu makam seorang bangsawan [3]. Analisis pada keamanan dapat diusulkan serangan yang dibedakan berdasarkan bias jangka pendek dan jangka panjang pada keystream. Pembeda dapat digunakan untuk membedakan keystream yang dihasilkan beberapa kunci dan untuk membedakan keystream yang dihasilkan oleh pasangan kunci [4]. Pada RC4 stream cipher, dalam stream awal yang mengeksploitasi fitur array dalam menghasilkan pseudorandom bit dengan menggunakan beberapa operasi sederhana [5]. Karena struktur RC4 cukup aman jika cipher digunakan untuk tindakan pencegahan yang sesuai, perubahan rancangan yang tidak memiliki izin dapat mengakibatkan kerentanan yang potensial [6]. GGHN merupakan stream cipher yang mirip dengan RC4 dirancang untuk menggunakan prosesor 32-bit yaitu 3-5 kali lebih cepat dari RC4. Salah satu sumber keamanan tinggi GGHN yaitu ukuran besar keadaan internal rahasianya dengan total 8240 bit [7]. GGHN yaitu stream cipher yang relatif lebih efisien yang terinspirasi dari desain atau rancangan RC4 [8]. Pada penelitian S. An-nissa, H. Mawengkang dan S. Efendi (2022), algoritma RC4 dan GGHN untuk pengamanan pesan, algortima RC4 mengolah unit dan memasukan data dalam satu waktu dengan kombinasi panjang karakter pesan dan kunci yang berbeda-beda. Hasil dalam proses menyatakan bahwa panjang kunci dalam proses enkripsi tidak mempengaruhi kemampuan dalam waktu proses. Panjang karater pesan memiliki pengaruh yang besar dalam lama waktu proses enkripsi dan dekripsi, jika semakin banyak jumlah karakter maka semakin lama waktu proses. Hal tersebut mempengaruhi proses dan tingkat keamanan informasi [9]. Pada penelitian F. Akbar, H. Mawengkang dan S. Efendi (2018), analisis perbandingan algoritma RC4+, RC4 NGG dan RC4 GGHN pada keamanan file gambar, berdasarkan tinjauan pustaka algoritma RC4+, RC4 NGG lebih efisien dan lebih cepat dibandingkan algoritam RC4. Dengan melakukan perbandingan waktu proses dan kompleksitasnya. Dilakukan dua proses pengujian pada setiap algoritma yaitu, mengenkripsi gambar dengan panjang kunci yang berbeda dan ukuran gambar sama serta mengenkripsi gambar dengan panjang kunci yang sama dan ukuran gambar yang berbeda. 1. PENDAHULUAN Data dan informasi dalam perkembangan teknologi digital memiliki peranan penting. Setiap kegiatan maupun aktifitas yang menggunakan teknologi digital berkaitan dengan data dan informasi, sehingga keamanan informasi maupun kerahasiaan data sangat penting. Dalam komunikasi digital terdapat proses transaksi dan distribusi data maupun informasi yang menjadi rutinitas. Kendala dan masalah muncul dalam komunikasi digital sehingga diperlukan teknik kriptografi yang memiliki tingkat keamanan yang lebih dan dapat diterapkan untuk menjaga keamanan informasi dan kerahasiaan data. Teknik kriptografi mampu melindungi hak akses keamanan informasi dan kerahasiaan data, proses transaksi dalam pengiriman dan penerimaan data terjamin tidak ada perubahan dan keaslian berdasarkan sumber data. Teknik kriptografi berhubungan dengan enkripsi yaitu dilakukan proses pengacakan data asli atau merubah pesan asli dan menyembunyikan data dengan sistem kunci sedangkan dekripsi yaitu proses perubahan kondisi data ke bentuk aslinya atau mengembalikan pesan asli agar mudah dipahami. Teknik kriptografi yang dapat digunakan yaitu simetri dan asimetris. Pada simetri, kunci kriptografi saat proses enkripsi dan dekripsi menggunakan kunci yang sama yaitu disebut Private Key. Pada asimetris, kunci kriptografi saat proses enkripsi dan dekripsi menggunakan kunci yang berbeda yaitu pada umumnya dalam proses dekripsi menggunakan Private Key dan proses enkripsi menggunakan Public Key. Algoritma NGG dan GGHN merupakan teknik kriptografi simetris yaitu menggunakan kunci yang sama dalam proses enkripsi dan dekripsi [1]. Pada dasarnya kriptografi bertujuan untuk menyediakan privasi dalam komunikasi dua entitas serta menyediakan otentikasi antara satu entitas dengan yang lainnya [2]. g y g y Untuk mengetahui tingkat keamanan dalam teknik kriptografi diperlukan analisis frekuensi. Analisis frekuensi pada kriptografi merupakan tentang memprediksi probabilitas huruf ataupun kelompok huruf ciphertext. Dalam kriptografi pada setiap rentang pesan berupa tulisan huruf atau kombinasi huruf memiliki frekuensi yang beragam. Copyright © 2022 Farid Akbar Siregar, Page 303 BITS is licensed under a Creative Commons Attribution 4.0 International License Copyright © 2022 Farid Akbar Siregar, Page 303 BITS is licensed under a Creative Commons Attribution 4.0 International License Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Pada proses distribusi terdapat kemunculan karakter yang sama untuk semua data maupun informasi yang digunakan. Analisis frekuensi akan menghitung jumlah karakter dari tulisan huruf atau kombinasi huruf dalam setiap proses enkripsi agar kemampuan kunci kriptografi dapat dianalisis dalam keamanan informasi. Teknik kriptografi dengan menggunakan algoritma NGG dan GGHN memiliki kemampuan dan tingkat keamanan yang berbeda. 1. PENDAHULUAN Hasilnya yaitu algoritma RC4 NGG merupakan algoritma yang lebih cepat sebesar 13,3% dari RC4+ dan sebesar 10,2 % lebih cepat dibandingkan RC4 GGHN. Panjang kunci tidak mempengaruhi waktu proses namun ukuran gambar mempengaruhi waktu proses [10]. g p g p Pada penelitian S. Suhandinata, R. A. Rizal, D. O. Wijaya, P. Warren dan S. Srinjiwi (2019), analisis performa kriptografi Hybrid, algoritma Blowfish dan algoritma RSA, dengan beberapa jenis data yaitu dokumen, gambar, audio dan video. Hasilnya yaitu algoritma Hybrid memiliki performa yang tidak terlalu berbeda dengan algoritma Blowfish dan dalam proses enkripsi dan dekripsi lebih terjamin keamanannya dengan kelebihan algoritma RSA. Untuk rata- rata nilai performa pada algoritma Hybrid yaitu proses enkripsi dokumen sebesar 0,85 detik, gambar sebesar 1,06 detik, audio sebesar 3,38 detik dan video sebesar 15,56 detik dan proses dekripsi dokumen sebesar 1,01 detik, gambar sebesar 1,38 detik, audio sebesar 4,3 detik dan video sebesar 27,56 detik. Dengan algoritma Hybrid memiliki proses enkripsi dan dekripsi yang lebih cepat namun tidak secepat algoritma pembentuknya [11]. Pada penelitian Quad -RC4: Merging Four RC4 States towards a 32 bit Stream Cipher oleh G. Paul dan S. Maitra (2012), dalam menggabungkan yang terbaik dari keduanya dengan mempertahakan struktur dasar RC4 dengan keamanan yang terjamin dan menggabungkan 4 status RC4 untuk membuat aliran stream cipher yaitu Quad-RC4 dengan hasil output 32-bit pada tiap putaran. Ketentuan penyimpanan dalam keadaan internal yaitu 1024 bit. Cipher memiliki kinerja yang lebih cepat dibandingkan RC4 normal dan sebanding dengan HC-128, stream cipher lebih cepat diantara eSTREAM [12]. Pada penelitian S. Banik, S. Maitra dan S. Sarkar (2011) tentang evolusi GGHN Cipher, cipher dimotivasi dari RC4 dengan tujuan untuk mendapatkan percepatan melalui pertimbangan output keystream berorientasi kata daripada berorientasi byte. Dapat dibuktikan bahwa terdapat sejumlah siklus pendek dengan panjang sama yang panjang array keadaan digunakan dalam cipher. Analisis secara teori, mengenai evolusi tipe GGHN, dipelajari model acak dari kata kunci. Menggunakan proses Markovian, ditunjukan bahwa model berkembang ke seluruh keadaan nol lebih cepat daripada rencana yang diharapkan [13]. Pada penelitian On the Weak State in GGHN-like Ciphers Oleh A. Kircanski, E. Al-Zaidy dan A. M. Youssef (2009), GGHN merupakan stream cipher yang relatif lebih efisien yang terinspirasi dari RC4. Terdapat aspek yang menunjukan tantangan dari prinsip desain terakhir. Secara khusus dinilai algoritma mirip GGHN dengan status yang lemah, dimana semua karakter status internal dan elemen output genap. Copyright © 2022 Farid Akbar Siregar, Page 304 BITS is licensed under a Creative Commons Attribution 4.0 International License Copyright © 2022 Farid Akbar Siregar, Page 304 BITS is licensed under a Creative Commons Attribution 4.0 International License 1. PENDAHULUAN GGHN diserap dalam keadaan lemah, bit tidak signifikan dari karakter plaintext tapi akan dapat terdeteksi dengan melihat ciphertext. Dengan model algoritma Markov chain dan menghitung waktu penyerapan, maka jumlah rata-rata langkah yang diperlukan algoritma untuk Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 menginput dalam keadaan lemah dapat lebih rendah dari yang diharapkan pada proses awal dan karena hal tersebut harus lebih hati-hati saat memperkirakan nilainya [14]. menginput dalam keadaan lemah dapat lebih rendah dari yang diharapkan pada proses awal dan karena hal tersebut harus lebih hati-hati saat memperkirakan nilainya [14]. Pada penelitian analisis keamanan RC4+ Stream Cipher oleh S. Banik, S. Sarkar dan R. Kacke (2013), diklaim bahwa RC4+ mengatasi sebagian besar kelemahan dari RC4 dan sedikit lebih lambat dari RC4 pada software. Pemasangan serangan agar membedakan RC4+ berdasarkan bias byte output awal. Pembeda membutuhkan kisaran 226 sampel yang merupakan hasil kesalahan diferensial pada RC4 yang diusulkan oleh peneliti sebelumnya di RC4+. Hasilnya menunjukan bahwa RC4+ rentan dengan serangan kesalahan diferensial dan kemungkinan untuk pemulihan semua keadaan internal cipher pada awal PRGA dengan memperbaiki kisaran 217,2 kesalahan [15]. Dalam penelitian ini, analisis frekuensi pada metode NGG dan GGHN dilakukan untuk mengetahui tingkat keamanan informasi pada proses transaksi maupun distribusi. Tujuan dari pengujian algoritma berdasarkan analisis frekuensi masing-masing agar dapat mengetahui kinerja maksimal masing-masing algoritma dengan hasil probabilitas huruf pada setiap pesan enkripsi sehingga dapat diketahui algoritma yang memiliki kemampuan dan tingkat keamanan yang lebih. 2.1 Tahapan Penelitian Tahapan yang akan dilakukan dalam pelaksanaan penelitian analisis frekuensi algoritma NGG dan GGHN memiliki langkah-langkah yang terdapat pada Gambar 1. Gambar 1. Kerangka Kerja Penelitian Studi Literatur Analisa Permasalahan Perancangan Sistem Implementasi Sistem Pengujian Kesimpulan Gambar 1. Kerangka Kerja Penelitian Studi Literatur Analisa Permasalahan Perancangan Sistem Implementasi Sistem Pengujian Kesimpulan Pengujian Gambar 1. Kerangka Kerja Penelitian a. Studi Literatur Studi literatur yang dilakukan pada penelitian ini yaitu mengumpulkan bahan refrensi mengenai keamanan informasi dengan menggunakan algoritma NGG dan GGHN dari berbagai sumber yang terpercaya untuk melengkapi informasi sehingga memiliki landasan teori dan ilmu yang sesuai. b. Analisa Permasalahan Pada tahap ini dilakukan analisis terhadap hasil studi literatur untuk mengetahui dan mendapatkan pemahaman mengenai algoritma NGG dan algoritma GGHN. b. Analisa Permasalahan Pada tahap ini dilakukan analisis terhadap hasil studi literatur untuk mengetahui dan mendapatkan pemahaman mengenai algoritma NGG dan algoritma GGHN. c Perancangan Sistem c. Perancangan Sistem Pada tahap perancangan sistem dilakukan perancangan arsitektur, pengumpulan data pelatihan, merancang antarmuka. Proses perancangan dilakukan berdasarkan hasil analisis studi literatur yang telah didapatkan. d. Implementasi Sistem Pada tahap implementasi sistem ini dilakukan penerapan algoritma NGG dan algoritma GGHN pada sistem untuk melakukan enkripsi dan dekripsi pada pesan teks. e. Pengujian Pada tahap ini dilakukan pengujian sistem terhadap hasil enkripsi pesan teks dengan menganalisis frekuensi kemunculan karakter pada pesan teks. e. Pengujian Pada tahap ini dilakukan pengujian sistem terhadap hasil enkripsi pesan teks dengan menganalisis frekuensi kemunculan karakter pada pesan teks. f Kesimp lan p p f. Kesimpulan Kesimpulan merupakan hasil pembahasan dari proses penerapan dan pengujian sistem berdasarkan analisa permasalahan yang ingin diselesaikan. f. Kesimpulan Kesimpulan merupakan hasil pembahasan dari proses penerapan dan pengujian sistem berdasarkan analisa permasalahan yang ingin diselesaikan. 2 2 Tahapan Metode Penelitian 2.2 Tahapan Metode Penelitian Tahapan metode penelitian pada analisis frekuensi algoritma NGG dan GGHN dapat dilihat pada Gambar 2. Copyright © 2022 Farid Akbar Siregar, Page 305 BITS is licensed under a Creative Commons Attribution 4.0 International License Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Gambar 2. Tahapan Metode Penelitian Mulai Teks Asli, Kunci Pesan Enkripsi Teks dengan Algoritma NGG Enkripsi Teks dengan Algoritma GGHN Cipher Text, Frekuensi Kemunculan Karakter Cipher Text, Frekuensi Kemunculan Karakter Selesai Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Mulai Teks Asli, Kunci Pesan Enkripsi Teks dengan Algoritma GGHN Enkripsi Teks dengan Algoritma NGG Cipher Text, Frekuensi Kemunculan Karakter Cipher Text, Frekuensi Kemunculan Karakter Selesai Gambar 2. Tahapan Metode Penelitian Dalam tahapan metode penelitian ini dimulai dari menginput teks asli dan kunci pesan yang merupakan random string, selanjutnya dilakukan proses enkripsi teks asli dengan menggunakan algoritma NGG dan kunci pesan yang sudah ditentukan, dan secara terpisah dilakukan proses enkripsi juga dengan menggunakan algoritma GGHN dan kunci pesan yang sudah ditentukan, hasil output teks dari masing-masing algoritma adalah berupa teks yang telah terenkripsi dan jumlah frekuensi kemunculan masing-masing karakter pada pesan teks. a. Secret key array K[0…N - 1]. b. Precomputed random array a[0…N - 1]. 2.3 Analisis Frekuensi Pada substitusi sederhana, huruf plaintext diganti dengan huruf berbeda dan huruf tertentu plaintext akan diubah menjadi huruf yang sama dalam ciphertext. Jikan munculnya huruf a menjadi huruf Y maka, pesan ciphertext yang terdapat banyak huruf Y akan disarankan ke cryptanalyst Y mewakili a. Dasar dari analisis frekuensi yaitu menghitung frekuensi huruf ciphertext dan dikaitkan dengan huruf plaintext yang diterka. Dimana lebih banyak Ys di ciphertext menunjukan bahwa Y berhubungan dengan a pada plaintext, namun tidak pasti karena huruf lain juga sangat umum, maka Y bisa menjadi salah satu bagian huruf. Perubahan menjadi plaintext lain juga tidak mungkin. Cryptanalyst harus mencoba kombinasi dalam pemetaan antara hururf ciphertext dan plaintext. Statistik yang kompleks dapat digunakan agar mudah dipahami, dengan mempertimbangkan jumlah pasangan huruf (bigram), tiga seragam (trigram) dan selanjutnya. Maka hal tersebut dapat dilakukan agar memberikan informasi yang lebih banyak ke cryptanalist [16]. Untuk mengukur tingkat fekuensi kemunculan karakter pada masing-masing algoritma yaitu dengan menggunakan persamaan berikut: BITS is licensed under a Creative Commons Attribution 4.0 International License Copyright © 2022 Farid Akbar Siregar, Page 306 BITS is licensed under a Creative Commons Attribution 4.0 International License Key Scheduling Algorithm (KSA) NGG yaitu : Input : Copyright © 2022 Farid Akbar Siregar, Page 306 BITS is licensed under a Creative Commons Attribution 4.0 International License Initialization: Initialization: for i = 0,…, N - 1 do S[i] = ai; j = 0; end Scrambling : for i = 0,…, N - 1 do j = (j + S[i] + K[i]) mod N; Swap(S[i], S[j]); S[i] = (S[i] + S[j]) mod M ; d S[i] = (S[i] + S[j]) mod M ; end Pada fase Pseudo Random Generation Algorithm (PRGA), elemen pseudo-random dikirim ke output dan setelah itu segera diubah oleh tambahan (mod M) dari dua elemen lain dari array S. Pseudo Random Generation Algorithm (PRGA) NGG yaitu : Pseudo Random Generation Algorithm (PRGA) NGG yaitu : Pseudo Random Generation Algorithm (PRGA) NGG yaitu : Input: Key-dependent scrambled array S[0…N - 1]. Output: Pseudo-random keystream bytes z. Output: Pseudo-random keystream bytes z. Initialization: Output Keystream Generation Loop: i = (i + 1) mod N; j = (j +S[i]) mod N; Swap(S[i], S[j]); Output z = S [(S[i] + S[j]) mod M ) mod N]; S [(S[i] + S[j]) mod M ) mod N] = (S[i] + S[j]) mod M ; j j GGHN cipher merupakan versi lain dari NGG yang diperkenalkan. Dalam algoritma GGHN j j GGHN cipher merupakan versi lain dari NGG yang diperkenalkan. Dalam algoritma GGHN terdapat tiga variabel yaitu i, j, dan k untuk meningkatkan keamanan cipher. k merupakan inisialisasi dari KSA dan bergantung pada kunci. Jumlah perulangan loop KSA r bergantung pada parameter n, m. GGHN cipher merupakan versi lain dari NGG GGHN cipher merupakan versi lain dari NGG yang diperkenalkan. D GGHN cipher merupakan versi lain dari NGG yang diperkenalkan. Dalam algoritma GGHN terdapat tiga variabel yaitu i, j, dan k untuk meningkatkan keamanan cipher. k merupakan inisialisasi dari KSA dan bergantung pada kunci. Jumlah perulangan loop KSA r bergantung pada parameter n, m. variabel yaitu i, j, dan k untuk meningkatkan keamanan cipher. k merupakan pada kunci. Jumlah perulangan loop KSA r bergantung pada parameter n, m. Key Scheduling Algorithm (KSA) GGHN yaitu : Key Scheduling Algorithm (KSA) GGHN yaitu : Input : 1. Secret key array K[0…N - 1]. 2. Precomputed random array a[0…N - 1]. 2. Precomputed random array a[0…N - 1]. Output : 1. Scrambled array S[0…N - 1]. 2. Key-dependent secret variable k. Initialization: Scrambling : 2.4 Algoritma NGG dan GGHN Terdapat cipher baru yang dikembangkan dengan cara memperluas RC4 menjadi 32 bit yaitu Sheet Bend dan Bowline. Generalisasi RC4 diusulkan bertujuan untuk memperluas RC4 menjadi 32 atau 64 bit dimana ukuran state menjadi lebih kecil yaitu 232 atau 264. Algoritma tersebut yaitu RC4 (n, m), dimana N=2n merupakan ukuran array state dalam kata-kata, m yaitu ukuran kata dalam bit, n ≤ m. NGG merupakan adopsi dari inisial para perancang algoritma. Pada NGG KSA dan PRGA diperbaharui indeks i, j dengan cara yang sama dengan RC4 KSA dan PRGA. p j g y g g Pada NGG KSA, larik S diinisialisasi ke larik acak yang ditentukan sebelumnya a. Selanjutnya, S [i] dan S[j] diganti dan jumlah dari kedua elemen ini ( mod M=2m) ditetapkan ke S [i] [1] [17]. Key Scheduling Algorithm (KSA) NGG yaitu : Input : Output : Scrambled array S[0…N - 1]. Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 BITS is licensed under a Creative Commons Attribution 4.0 International License Scrambling : g for i = 0,…, N - 1 do Dalam Pseudo Random Generation Algorithm (PRGA), k digunakan untuk memperbarui S serta untuk menyamarkan output. Pseudo Random Generation Algorithm (PRGA) GGHN yaitu : Input: Key-dependent scrambled array S[0…N - 1]. Input: Key-dependent scrambled array S[0…N - 1]. Output: Pseudo-random keystream bytes z. Initialization: Output: Pseudo-random keystream bytes z. j Output Keystream Generation Loop: i = (i + 1) mod N; j = (j +S[i]) mod N; k = (k +S[j]) mod M; Output z = S [(S[i] + S[j]) mod N + k) mod M]; S [(S[i] + S[j]) mod N] = (k + S[i]) mod M ; Output Keystream Generation Loop: Output Keystream Generation Loop: i = (i + 1) mod N; j = (j +S[i]) mod N; k = (k +S[j]) mod M; Output z = S [(S[i] + S[j]) mod N + k) mod M]; S [(S[i] + S[j]) mod N] = (k + S[i]) mod M ; i = (i + 1) mod N; Copyright © 2022 Farid Akbar Siregar, Page 307 BITS is licensed under a Creative Commons Attribution 4.0 International License Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Copyright © 2022 Farid Akbar Siregar, Page 308 BITS is licensed under a Creative Commons Attribution 4.0 International License 3.1 Penerapan Algoritma NGG dan GGHN 3.1 Penerapan Algoritma NGG dan GGHN Pada proses penerapan algoritma NGG dan algoritma GGHN dilakukan 10 percobaan terhadap 3000 karakter pesan teks untuk melihat hasil enkripsi dan jumlah kemunculan tiap karakter pada pesan teks dengan menggunakan panjang karakter kunci yang berbeda-beda. Kode karakter yang digunakan yaitu kode ASCII 8 bit. Informasi yang dapat diambil yaitu, jumlah karakter pesan teks yang digunakan, jumlah kemunculan masing-masing karakter pada pesan teks dan presentasi kemunculan masing-masing karakter. Pada percobaan ke-1 dapat dilihat pada Gambar 5, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 10 karakter kunci yaitu “analisakri”. Dapat dilihat hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “ ” kode ASCII 129 sebagai karakter yang kemunculannya paling banyak 26 kali dengan presentasi kemunculan sebesar 0.86666667%. Pada algoritma GGHN yaitu karakter “NULL” kode ASCII 0 sebagai karakter yang kemunculannya paling banyak 22 kali dengan presentasi kemunculan sebesar 0.7333% pada pesan teks. Gambar 2. Hasil Enkripsi Percobaan ke-1 Gambar 2. Hasil Enkripsi Percobaan ke-1 Pada percobaan ke-2, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 20 karakter kunci yaitu “analisakriptografial”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “ACK” kode ASCII 6 sebagai karakter yang kemunculannya paling banyak 22 kali dengan presentasi kemunculan sebesar 0.7333%. Sedangkan pada algoritma GGHN yaitu karakter “q” kode ASCII 113 sebagai karakter yang kemunculannya paling banyak 23 kali dengan presentasi kemunculan sebesar 0.76666667%. y g p Pada percobaan ke-3, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 30 karakter kunci yaitu “analisakriptografialgortimangg”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “Æ” kode ASCII 198 sebagai karakter yang kemunculannya paling banyak 25 kali dengan presentasi kemunculan sebesar 0.8333%. Sedangkan pada algoritma GGHN yaitu karakter “EM” kode ASCII 25 sebagai karakter yang kemunculannya paling banyak 22 kali dengan presentasi kemunculan sebesar 0.7333%. 3. HASIL DAN PEMBAHASAN Pada bab ini akan diuraikan hasil dan pembahasan yang dilakukan meliputi penerapan algoritma NGG dan algoritma GGHN dan hasil pengujian berdasarkan analisis frekuensi masing-masing algoritma. Data pada penelitian ini adalah pesan teks sejumlah 3000 karakter yang akan dienkripsi dengan menggunakan algoritma NGG dan GGHN. Analisis frekuensi dilakukan dengan menghitung kemunculan karakter pada teks pesan yang telah dienkripsi. Kunci yang digunakan yaitu teks random string seperti pada Tabel 1. Tabel 1. Tabel Data Kunci No Percobaan Ke- Panjang Kunci Isi Kunci 1. 1 10 karakter analisakri 2. 2 20 karakter analisakriptografial 3. 3 30 karakter analisakriptografialgortimangg 4. 4 40 karakter analisakriptografialgoritmanggdangghndal 5. 5 50 karakter analisakriptografialgoritmanggdangghndalamkeamanan 6. 6 60 karakter analisakriptografialgoritmanggdangghndalamkeamananpesanteksd 7. 7 70 karakter analisakriptografialgoritmanggdangghndalamkeamananpesanteksdeng ankarak 8. 8 80 karakter analisakriptografialgoritmanggdangghndalamkeamananpesanteksdeng ankarakterpadakun 9. 9 90 karakter analisakriptografialgoritmanggdangghndalamkeamananpesanteksdeng ankarakterpadakuncipesantek 10. 10 100 karakter analisakriptografialgoritmanggdangghndalamkeamananpesanteksdeng ankarakterpadakuncipesanteks1karakter 3 1 Penerapan Algoritma NGG dan GGHN Copyright © 2022 Farid Akbar Siregar, Page 308 BITS is licensed under a Creative Commons Attribution 4.0 International License Copyright © 2022 Farid Akbar Siregar, Page 308 BITS is licensed under a Creative Commons Attribution 4.0 International License Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Pada percobaan ke-4, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 40 karakter kunci yaitu “analisakriptografialgoritmanggdangghndal”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “BS” kode ASCII 8 sebagai karakter yang kemunculannya paling banyak 26 kali dengan presentasi kemunculan sebesar 0.86666667%. Sedangkan pada algoritma GGHN yaitu karakter “#” kode ASCII 35 sebagai karakter yang kemunculannya paling banyak 24 kali dengan presentasi kemunculan sebesar 0.8%. Pada percobaan ke-5, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 50 karakter kunci yaitu “analisakriptografialgoritmanggdangghndalamkeamanan”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “)” kode ASCII 41 sebagai karakter yang kemunculannya paling banyak 20 kali dengan presentasi kemunculan sebesar 0.66666667%. Sedangkan pada algoritma GGHN yaitu karakter “”” kode ASCII 132 sebagai karakter yang kemunculannya paling banyak 22 kali dengan presentasi kemunculan sebesar 0.7333%. Pada percobaan ke-6, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 60 karakter kunci yaitu “analisakriptografialgoritmanggdangghndalamkeamananpesanteksd”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “L” kode ASCII 76 sebagai karakter yang kemunculannya paling banyak 22 kali dengan presentasi kemunculan sebesar 0.7333%. Sedangkan pada algoritma GGHN yaitu karakter “a” kode ASCII 97 sebagai karakter yang kemunculannya paling banyak 28 kali dengan presentasi kemunculan sebesar 0.9333%. Pada percobaan ke-7, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 70 karakter kunci yaitu “analisakriptografialgoritmanggdangghndalamkeamananpesanteksdengankarak”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “CAN” kode ASCII 24 sebagai karakter yang kemunculannya paling banyak 22 kali dengan presentasi kemunculan sebesar 0.7333%. Sedangkan pada algoritma GGHN yaitu karakter “¿” kode ASCII 191 sebagai karakter yang kemunculannya paling banyak 23 kali dengan presentasi kemunculan sebesar 0.76666667%. Pada percobaan ke-8, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 80 karakter kunci yaitu “analisakriptografialgoritmanggdangghndalamkeamananpesanteksdengankarakterpadakun”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “ä” kode ASCII 228 sebagai karakter yang kemunculannya paling banyak 23 kali dengan presentasi kemunculan sebesar 0.76666667%. Sedangkan pada algoritma GGHN yaitu karakter “r” kode ASCII 114 sebagai karakter yang kemunculannya paling banyak 22 kali dengan presentasi kemunculan sebesar 0.7333%. Copyright © 2022 Farid Akbar Siregar, Page 308 BITS is licensed under a Creative Commons Attribution 4.0 International License Pada percobaan ke-9, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 90 karakter kunci yaitu “analisakriptografialgoritmanggdangghndalamkeamananpesanteksdengankarakterpadakuncipesantek”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “Ã” kode ASCII 195 sebagai karakter yang kemunculannya paling banyak 23 kali dengan presentasi kemunculan sebesar 0.76666667%. Sedangkan pada algoritma GGHN yaitu karakter “M” kode ASCII 77 sebagai karakter yang kemunculannya paling banyak 29 kali dengan presentasi kemunculan sebesar 0.96666667%. g p Pada percobaan ke-10 atau percobaan terakhir, dilakukan penerapan algortima NGG dan GGHN dengan menggunakan 100 karakter kunci yaitu “analisakriptografialgoritmanggdangghndalamkeamananpesanteksdengankar akterpadakuncipesanteks1karakter”. Hasil enkripsi dan analisis frekuensi pada algoritma NGG yaitu karakter “GS” kode ASCII 29 sebagai karakter yang kemunculannya paling banyak 24 kali dengan presentasi kemunculan sebesar 0.8%. Sedangkan pada algoritma GGHN yaitu karakter “i” kode ASCII 105 sebagai karakter yang kemunculannya paling banyak 27 kali dengan presentasi kemunculan sebesar 0.9%. p g y g p Hasil dari 10 percobaan dengan menggunakan 10 panjang kunci yang berbeda-beda memiliki hasil frekuensi yang bervariasi. Rincian frekuensi tertinggi pada setiap percobaan dapat dilihat pada Tabel 2. Semakin banyak jumlah karakter yang digunakan pada kunci maka akan mempengaruhi frekuensi kemunculan karakter pada proses enkripsi. Tabel 2. Tabel Frekuensi No Percobaan Ke- Panjang Kunci Frekuensi Tertinggi NGG Frekuensi Tertinggi GGHN 1. 1 10 karakter 26 kali 22 kali 2. 2 20 karakter 22 kali 23 kali 3. 3 30 karakter 25 kali 22 kali 4. 4 40 karakter 26 kali 24 kali 5. 5 50 karakter 20 kali 22 kali 6. 6 60 karakter 22 kali 28 kali 7. 7 70 karakter 22 kali 23 kali 8. 8 80 karakter 23 kali 22 kali 9. 9 90 karakter 23 kali 29 kali 10. 10 100 karakter 24 kali 27 kali 3.2 Hasil Pengujian BITS is licensed under a Creative Commons Attribution 4.0 International License 3.2 Hasil Pengujian Tabel Presentasi Frekuensi No Percobaan Ke- Panjang Kunci Isi Kunci Presentasi Kemuncualan Tertinggi NGG Presentasi Kemuncualan Tertinggi GGHN 1. 1 10 karakter analisakri 0.86666667% 0.7333% 2. 2 20 karakter analisakriptografial 0.7333% 0.76666667% 3. 3 30 karakter analisakriptografialgo rtimangg 0.8333% 0.7333% 4. 4 40 karakter analisakriptografialgo ritmanggdangghndal 0.86666667% 0.8% 5. 5 50 karakter analisakriptografialgo ritmanggdangghndala mkeamanan 0.66666667% 0.7333% 6. 6 60 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks d 0.7333% 0.9333% 7. 7 70 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarak 0.7333% 0.76666667% 8. 8 80 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku n 0.76666667% 0.7333% 9. 9 90 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku ncipesantek 0.76666667% 0.96666667% 10. 10 100 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku ncipesanteks1karakter 0.8% 0.9% Gambar 3. Grafik Hasil Presentasi Frekuensi Berdasarkan dari Tabel 3 dan Gambar 3 dapat dilihat pada grafik hasil presentasi frekuensi kemunculan karakter terhadap panjang kunci yang berbeda-beda pada algoritma NGG dan GGHN memiliki hasil yang berbeda beda pada setiap percobaan. Pada 10 percobaan analisis frekuensi algoritma NGG didapat rata-rata presentasi kemunculan sebesar 0.007766533 % dan algoritma GGHN didapat rata-rata presentasi kemunculan sebesar 0.0080665 %. 0,00% 0,20% 0,40% 0,60% 0,80% 1,00% 1,20% Ke-1 Ke-2 Ke-3 Ke-4 Ke-5 Ke-6 Ke-7 Ke-8 Ke-9 Ke-10 Presentasi Frekuensi NGG GGHN Tabel 3. Tabel Presentasi Frekuensi Tabel 3. Tabel Presentasi Frekuensi Tabel 3. Tabel Presentasi Frekuensi No Percobaan Ke- Panjang Kunci Isi Kunci Presentasi Kemuncualan Tertinggi NGG Presentasi Kemuncualan Tertinggi GGHN 1. 1 10 karakter analisakri 0.86666667% 0.7333% 2. 2 20 karakter analisakriptografial 0.7333% 0.76666667% 3. 3 30 karakter analisakriptografialgo rtimangg 0.8333% 0.7333% 4. 4 40 karakter analisakriptografialgo ritmanggdangghndal 0.86666667% 0.8% 5. 5 50 karakter analisakriptografialgo ritmanggdangghndala mkeamanan 0.66666667% 0.7333% 6. 6 60 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks d 0.7333% 0.9333% 7. 7 70 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarak 0.7333% 0.76666667% 8. 8 80 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku n 0.76666667% 0.7333% 9. 9 90 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku ncipesantek 0.76666667% 0.96666667% 10. 10 100 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku ncipesanteks1karakter 0.8% 0.9% Presentasi Frekuensi 0.8% 0.9% Gambar 3. Grafik Hasil Presentasi Frekuensi 0,00% 0,20% 0,40% 0,60% 0,80% 1,00% 1,20% Ke-1 Ke-2 Ke-3 Ke-4 Ke-5 Ke-6 Ke-7 Ke-8 Ke-9 Ke-10 Presentasi Frekuensi NGG GGHN Gambar 3. Grafik Hasil Presentasi Frekuensi Berdasarkan dari Tabel 3 dan Gambar 3 dapat dilihat pada grafik hasil presentasi frekuensi kemunculan karakter terhadap panjang kunci yang berbeda-beda pada algoritma NGG dan GGHN memiliki hasil yang berbeda beda pada setiap percobaan. 3.2 Hasil Pengujian Pada bagian ini akan dilakukan pengujian terhadap analisis frekuensi dari hasil enkripsi pada algoritma NGG dan GGHN untuk mengetahui tingkat keamanan informasi. Proses pengujian kedua algoritma menggunakan pesan teks sejumlah 3000 karakter yang diinput ke sistem dan akan dienkripsi menggunakan masing-masing algoritma yaitu Copyright © 2022 Farid Akbar Siregar, Page 309 Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 Copyright © 2022 Farid Akbar Siregar, Page 310 algoritma NGG dan algortima GGHN. Selanjutnya berdasarkan hasil enkripsi akan dianalisis frekuensi karakter yang muncul memiliki presenasi yang dapat lihat pada tabel 3. Tabel 3. Tabel Presentasi Frekuensi No Percobaan Ke- Panjang Kunci Isi Kunci Presentasi Kemuncualan Tertinggi NGG Presentasi Kemuncualan Tertinggi GGHN 1. 1 10 karakter analisakri 0.86666667% 0.7333% 2. 2 20 karakter analisakriptografial 0.7333% 0.76666667% 3. 3 30 karakter analisakriptografialgo rtimangg 0.8333% 0.7333% 4. 4 40 karakter analisakriptografialgo ritmanggdangghndal 0.86666667% 0.8% 5. 5 50 karakter analisakriptografialgo ritmanggdangghndala mkeamanan 0.66666667% 0.7333% 6. 6 60 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks d 0.7333% 0.9333% 7. 7 70 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarak 0.7333% 0.76666667% 8. 8 80 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku n 0.76666667% 0.7333% 9. 9 90 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku ncipesantek 0.76666667% 0.96666667% 10. 10 100 karakter analisakriptografialgo ritmanggdangghndala mkeamananpesanteks dengankarakterpadaku ncipesanteks1karakter 0.8% 0.9% Gambar 3. Grafik Hasil Presentasi Frekuensi Berdasarkan dari Tabel 3 dan Gambar 3 dapat dilihat pada grafik hasil presentasi frekuensi kemunculan karakter terhadap panjang kunci yang berbeda-beda pada algoritma NGG dan GGHN memiliki hasil yang berbeda beda pada setiap percobaan. Pada 10 percobaan analisis frekuensi algoritma NGG didapat rata-rata presentasi kemunculan sebesar 0.007766533 % dan algoritma GGHN didapat rata-rata presentasi kemunculan sebesar 0.0080665 %. 0,00% 0,20% 0,40% 0,60% 0,80% 1,00% 1,20% Ke-1 Ke-2 Ke-3 Ke-4 Ke-5 Ke-6 Ke-7 Ke-8 Ke-9 Ke-10 Presentasi Frekuensi NGG GGHN algoritma NGG dan algortima GGHN. Selanjutnya berdasarkan hasil enkripsi akan dianalisis frekuensi karakter yang muncul memiliki presenasi yang dapat lihat pada tabel 3. algoritma NGG dan algortima GGHN. Selanjutnya berdasarkan hasil enkripsi akan dianalisis frekuensi karakter yang muncul memiliki presenasi yang dapat lihat pada tabel 3. p y g p p Tabel 3. BITS is licensed under a Creative Commons Attribution 4.0 International License 4. KESIMPULAN Pada penelitian ini, analisis frekuensi terhadap algoritma NGG dan GGHN untuk mengetahui tingkat keamanan informasi berdasarkan hasil enkripsi data. Dalam proses enkripsi pesan tergantung pada kunci yang telah ditentukan. Hal tersebut memiliki pengaruh yang signifikan terhadap hasil enkripsi. Berdasarkan proses pengujian pada algoritma NGG dan GGHN menghasilkan frekuensi karakter pada teks yang bervariasi. Pada 10 percobaan didapat rata-rata presentasi kemunculan tertinggi pada algoritma NGG yaitu 0.007766533% dan rata-rata presentasi kemunculan tertinggi pada algoritma GGHN 0.0080665%. Berdasarkan pada penelitian ini dapat diketahui bahwa algoritma NGG memiliki tingkat keamanan yang lebih tinggi dibandingkan algoritma GGHN dengan selisih presentasi sebesar 0.000299967%. Jika semakin banyak jumlah karakter yang digunakan pada kunci maka akan mempengaruhi tingkat keamanan informasi. Diharapkan dapat dilakukan pengujian terhadap lebih banyak pesan dan kunci agar menghasilkan akurasi dalam proses analisis frekuensi yang lebih spesifik. Jika frekuensi kemunculan karakter pada teks pesan yang telah dienkripsi semakin sering atau semakin tinggi, maka tingkat keamanan informasi pada pesan lebih rendah dan kata kunci lebih mudah dipecahkan. 3.2 Hasil Pengujian Pada 10 percobaan analisis frekuensi algoritma NGG didapat rata-rata presentasi kemunculan sebesar 0.007766533 % dan algoritma GGHN didapat rata-rata presentasi kemunculan sebesar 0.0080665 %. Copyright © 2022 Farid Akbar Siregar, Page 310 BITS is licensed under a Creative Commons Attribution 4.0 International License Building of Informatics, Technology and Science (BITS) Volume 4, No 1, Juni 2022 Page: 303−311 ISSN 2684-8910 (media cetak) ISSN 2685-3310 (media online) DOI 10.47065/bits.v4i1.1639 REFERENCES [1] A. Khalid, G. Paul, and A. Chattopadhyay, Domain specific high-level synthesis for cryptographic workloads. p y y [2] S. D. Galbraith, Mathematics of Public Key Cryptography, 1st ed. United State of America: Cambridge University Press, 2012. [3] D. Chaudhuri, G. Fink, and K. Liu, Discrete subaortic stenosis in a patient with a history of repaired AV canal defect, vol. 28, no. 9. 2011. doi: 10.1111/j.1540-8175.2011.01507.x. [3] D. Chaudhuri, G. Fink, and K. Liu, Discrete subaortic stenosis in a patient with a history of repaired AV canal defect, vol. 28, no. 9. 2011. doi: 10.1111/j.1540-8175.2011.01507.x. [4] S. Banik and T. Isobe, “Cryptanalysis of the full Spritz stream cipher,” in Lecture Notes in Computer Science (including b i L N i A ifi i l I lli d L N i Bi i f i ) 2016 l 9783 63 77 d i [4] S. Banik and T. Isobe, “Cryptanalysis of the full Spritz stream cipher,” in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2016, vol. 9783, pp. 63–77. doi: 10.1007/978-3-662-52993-5_4. [5] D. Liestyowati, Public Key Cryptography, vol. 1477, no. 5. 2020. doi: 10.1088/1742-6596/1477/5/052062. [6] G. Paul, S. Maitra, and A. Chattopadhyay, “Quad-RC4: Merging Four RC4 States towards a 32-bit Stream Cipher.,” IACR Cryptol. ePrint Arch., vol. 2013, no. May, p. 572, 2013, [Online]. Available: http://dblp.uni- trier.de/db/journals/iacr/iacr2013.html#PaulMC13 j [7] A. Kircanski and A. M. Youssef, “On the structural weakness of the GGHN stream cipher,” Cryptogr. Commun., vol. 2, no. 1, pp. 1–17, Sep. 2010, doi: 10.1007/s12095-009-0013-3. [8] A. Kircanski and A. M. Youssef, “On the weak state in GGHN-like ciphers,” in Proceedings - 2012 7th International Conference on Availability, Reliability and Security, ARES 2012, 2012, pp. 397–401. doi: 10.1109/ARES.2012.32. [9] I. Print, S. An-nissa, H. Mawengkang, and S. Efendi, “InfoTekJar : Jurnal Nasional Informatika dan Teknologi GGHN Cryptography Algorithm for Message Security,” vol. 2, pp. 6–10, 2022. yp g p y g g y pp [10] F. Akbar, H. Mawengkang, and S. Efendi, “Comparative analysis of RC4+ algorithm, RC4 NGG algorithm and RC4 GGHN algorithm on image file security,” IOP Conf. Ser. Mater. Sci. Eng., vol. 420, no. 1, 2018, doi: 10.1088/1757- 899X/420/1/012131. [11] S. Suhandinata, R. A. Rizal, D. O. Wijaya, P. Warren, and S. Srinjiwi, “Analisis Performa Kriptografi Hybrid Algoritma Blowfish Dan Algoritma Rsa,” JURTEKSI (Jurnal Teknol. dan Sist. Informasi), vol. 6, no. 1, pp. 1–10, 2019, doi: 10.33330/jurteksi.v6i1.395. j p 7] P. Singh, D. Mishra, and K. Seth, Transformation in healthcare with emerging technologies. REFERENCES [12] G. Paul and S. Maitra, RC4 stream cipher and its variants. CRC Press, 2012. p [13] S. Banik, S. Maitra, and S. Sarkar, “On the evolution of GGHN cipher,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 7107 LNCS, pp. 181–195, 2011, doi: 10.1007/978-3-642-25578-6_15. [14] A. Kircanski, R. Al-Zaidy, and A. M. Youssef, “A new distinguishing and key recovery attack on NGG st Cryptogr. Commun., vol. 1, no. 2, pp. 269–282, May 2009, doi: 10.1007/s12095-009-0012-4. [15] S. Banik, S. Sarkar, and R. Kacker, “Security analysis of the RC4+ stream cipher,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 8250 LNCS, pp. 297–307, 2013, doi: 10.1007/978-3-319- 03515-4_20. [16] F. Riza, N. Sridewi, A. M. Husein, and M. K. Harahap, “Analisa Frekuensi Hasil Enkripsi Pada Algoritma Kriptografi Blowfish Terhadap Keamanan Informasi,” J. Teknol. dan Ilmu Komput. Prima, vol. 1, no. 1, pp. 11–15, 2018, doi: 10.34012/jutikomp.v1i1.233. j p [17] P. Singh, D. Mishra, and K. Seth, Transformation in healthcare with emerging technologies.
https://openalex.org/W3115814459	https://www.researchsquare.com/article/rs-76086/v1.pdf?c=1600287048000	English	null	Identification of a Novel Epithelial-mesenchymal Transition Gene Signature Regulated by KEAP1-NRF2&nbsp;Pathway in Esophageal Carcinoma	Research Square (Research Square)	2,020	cc-by	6,669	Identification of a Novel Epithelial-mesenchymal Transition Gene Signature Regulated by KEAP1- NRF2 Pathway in Esophageal Carcinoma Mohamed Elshaer Zhejiang University,School of medicine Ahmed Hammad Zhejiang University, School of medicine Xiu Jun Wang Zhejiang University,School of medicine xiuwen Tang (  xiuwentang@zju.edu.cn ) Department of Biochemistry and Department of Tho Zhejiang University School of Medicine, Hangzhou 31 6601-1234 Mohamed Elshaer Primary research Keywords: TCGA, KEAP1, NRF2, epithelial-mesenchymal transition, esophageal cancer, biomarker Posted Date: September 16th, 2020 DOI: https://doi.org/10.21203/rs.3.rs-76086/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Methods In this study, 182 tumor samples from the Cancer Genome Atlas (TCGA)-ESCA RNA-Seq V2 level 3 data were segregated into two groups KEAP1-NRF2-mutated (22) and wild-type (160).The two groups were subjected to differential gene expression analysis and we performed Gene Set Enrichment Analysis (GSEA). Then, the enriched gene set was integrated with the differentially expressed genes (DEGs) to identify a gene signature regulated by the KEAP1-NRF2 pathway in ESCA. Furthermore, we validated the gene signature using mRNA expression data of ESCA cell lines provided by the Cancer Cell Line Encyclopedia (CCLE). The identified signature was tested in 3 independent ESCA datasets to assess its prognostic value. Background KEAP1-NRF2 pathway alterations were identified in many cancers including, esophageal cancer (ESCA). Identifying biomarkers that are associated with mutations in this pathway will aid in defining this cancer subset; and hence in supporting precision and personalized medicine. Results We identified 11 epithelial-mesenchymal transition (EMT) genes regulated by the KEAP1-NRF2 pathway in ESCA patients. Five of the 11 genes showed significant over-expression in KEAP1-NRF2-mutated ESCA cell lines. In addition, over-expression of these five genes was significantly associated with poor survival in 3 independent ESCA datasets, including the TCGA-ESCA dataset. Conclusion Altogether, we identified a novel EMT 5-gene signature regulated by the KEAP1-NRF2 axis and this signature is strongly associated with metastasis and drug resistance in ESCA. These 5-genes are potential biomarkers and therapeutic targets for ESCA patients in whom the KEAP1-NRF2 pathway is altered. DOI: https://doi.org/10.21203/rs.3.rs-76086/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/20 Page 1/20 Background Esophageal cancer is the sixth most common cause of cancer death and the eighth in incidence worldwide. In fact, it accounts for 4% of cancer diagnoses and for 6% of cancer deaths. The prognosis for esophageal carcinoma is poor, with a 5-year survival rate of 19% and only 0.9% for advanced esophageal carcinoma (1). To maintain oxidative homeostasis, cancer cells increase the transcription of antioxidant genes by acquiring either stabilizing mutations in NFE2L2 (encoding NRF2, the master transcriptional regulator of the cellular antioxidant program) or by selecting for inactivating mutations in its negative regulator, Page 2/20 KEAP1 (2). KEAP1 is a substrate receptor of the Cul3-RING ubiquitin ligase (CRL3) that, under physiological conditions, constitutively binds and targets NRF2 for degradation. In response to oxidative stress, the KEAP1-NRF2 binding is inhibited and, consequently, NRF2 is stabilized (3). KEAP1 (2). KEAP1 is a substrate receptor of the Cul3-RING ubiquitin ligase (CRL3) that, under physiological conditions, constitutively binds and targets NRF2 for degradation. In response to oxidative stress, the KEAP1-NRF2 binding is inhibited and, consequently, NRF2 is stabilized (3). The TCGA network has revolutionized the cancer research by enriching the cancer research community with a huge amount of cancer-related data. This revolution has enabled researchers to identify cancer driver genes, cancer dependency, prognostic biomarkers and therapeutic targets. In addition, it has enabled researchers to segregate one cancer type into subgroups in order to assist personalized and precision medicine field. Identification of genes and biological processes that are regulated by the KEAP1- NRF2 pathway in different cancers may provide an effective approach for therapy of subset of cancers that harbor KEAP1-NRF2 pathway alterations. Moreover, these gene signatures can be used to predict survival of patients. In previous studies, we identified gene signatures that are regulated by the KEAP1-NRF2 pathway in lung adenocarcinoma (LUAD) and head and neck squamous cancer (HNSC) (4–6). In the current study, we used the genomics, and transcriptomics data of the ESCA cohort from TCGA to identify a gene signature regulated by the KEAP1-NRF2 pathway in ESCA. Overall database selection TCGA RNA-Seq gene expression version 2 level 3 data (Illumina HiSeq platform) for 182 ESCA tissues were downloaded from the Broad GDAC (Global Data Assembly Centers) Firehose website (http://gdac.broadinstitute.org/). All the mutation data used in the present study was obtained from CVCDAP (https://omics.bjcancer.org/cvcdap/home.do) and UCSC Xena Browser (https://xenabrowser.net/) (7, 8). Twelve percent (22 out of 185) of the TCGA-ESCA patients were found to harbor mutations in either KEAP1, NRF2 or both. Then, we segregated these patients into two groups, one had 22 patients with mutations in either KEAP1, NRF2 or both (mutated group) while the other had 160 patients with neither KEAP1 nor NRF2 mutations (wild-type group). RNA-Seq data analysis TCGA RNA-Seq gene expression version 2 level 3 data (Illumina Hiseq platform) for 182 ESCA tissues were subjected to differential gene expression analysis. Briefly, level 3 transcriptomic data was normalized by the FPKM (Fragments per Kilobase of transcript per Million mapped reads) method. All gene expression values were log-transformed to approximate the data to a normal distribution. In addition, genes with zero values in more than 25% of the patients were excluded. The differentially expressed genes (DEGs) were identified by applying the two-tailed t-test assuming unequal variance. Then, P values were adjusted using the FDR method. DEGs with FDR 0.05 were considered significant. Gene Set Enrichment Analysis (GSEA) Cancer Cell Line Encyclopedia (CCLE) data analysis RNA-seq gene expression data for 1019 cell lines were downloaded from CCLE https://portals.broadinstitute.org/ccle/data. We identified ESCA cell lines that harbor mutations in either KEAP1, NRF2 or both using the COSMIC data base https://cancer.sanger.ac.uk/cell_lines. Then, we divided the ESCA cell lines into two groups mutated and wild-type. Differential gene expression analysis was performed as described above. Identification of NRF2 binding sites by in silico analysis To identify the NRF2 binding sites within the promoter regions of the putative NRF2 regulated genes, we used the transcription factor-binding site finding tool ConTra V3 (9) with stringency core = 0.95 and similarity matrix = 0.85. The search was limited to the -1 kb upstream the transcription start site (TSS). Gene Set Enrichment Analysis (GSEA) Page 3/20 GSEA (https://omics.bjcancer.org/) was performed to determine all significantly affected biological pathways. GSEA is a computational method that determines whether a defined set of genes shows statistically significant, concordant differences between two biological states. The primary result of the GSEA is the enrichment score (ES), which reflects the degree to which a gene set is overrepresented at the top or bottom of a ranked list of genes. The gene list metric was built using ratio between classes (biological states). A positive ES indicates gene set enrichment at the top of the ranked list (up-regulated); a negative ES indicates gene set enrichment at the bottom of the ranked list (down-regulated). Survival analysis For the identification of prognostic biomarkers, Kaplan-Meier curves were generated by using the web- based patients survival analysis tool SurvExpress (10). Log rank test P<0.05 was used as the cutoff for significance. The method of analysis has been discussed in a previous study (5). Overview of the ESCA mutational landscape The lack of prognostic biormarkers for the ESCA subgroup with KEAP1-NRF2 mutations motivated us to focus on identifying prognostic biormakers that is regulated by KEAP1-NRF2 in ESCA. First, we investigated the different driver mutations in the TCGA-ESCA patients and we found that TP53 was the most frequently mutated gene in ESCA as it was found mutated in 86% of patients (Fig.2B). 42% of ESCA patients harbored TTN mutations which made TTN in the second place. MUC16, SYNE1, CSMD3, FLAG, DNAH5, HMCN1, LRP1B, PCLO and RYR2 were mutated in 23%, 20%, 20%, 18%, 16%, 16%, 16%, 15% and 12% of ESCA patients, respectively. KEAP1-NRF2 was found mutated in 12% of ESCA patients. The TCGA-ESCA cohort included 182 patient samples. We segregated the cohort into two groups: KEAP1- NRF2-mutated (22 patients) and wild-type (160 patients). In order to ensure that the differences between the two groups were due to KEAP1-NRF2 mutations, we first investigated the driver mutations in the KEAP1-NRF2-mutated group. We found that TP53 was mutated in 95% of patients in this group, NRF2 was found mutated in 85% patients, TTN, KMTD2, MUC13, KEAP1, PATCH, and SACS were found mutated in 41%, 32%, 27%, 23%, 23% and 23%, respectively (Fig.3A). Then, we performed differential gene mutation analysis between the two groups to investigate the percentages of mutations of these driver genes in the two groups. Only NRF2 and KEAP1 weren't mutated in wild-type group while the other driver genes were found mutated in both the KEAP1-NRF2-mutated and wild-type groups, with similar percentages (Fig.3B). Therefore, none of these driver genes can be considered as variables that contribute to differences between the two groups. In order to better understand the mutational landscape of KEAP1-NRF2 in ESCA, we used the USCS Xena browser to examine the types of mutations and their positions in the domain structure of KEAP1 and NRF2 proteins. As noted earlier, we found that 2.19% of TCGA-ESCA patient samples had KEAP1 mutations while NRF2 was mutated in 9.34% and both were mutated in 0.54%. All the detected KEAP1 mutations were missense mutations while 77.8% (14/18) of NRF2 mutation were missense mutations, 11.1% were intron (2/18), 5.6% (1/18) were in-frame-deletions and 5.6% (1/18) were in-frame-insertions. Overview of the mutational landscape of the KEAP1-NRF2 pathway in cancer First, we investigated the mutational landscape of the KEAP1-NRF2 pathway in cancer by analyzing 11,079 TCGA samples from 33 different cancer types using the CVCDAP database. As shown in (Fig.2A), we found that the KEAP1-NRF2 pathway was altered in many cancers, however it was altered with a percentage that was higher than 10% in only five cancer types. Lung Squamous Cell Cancer (LUSC) was the cancer type that harbor the highest percentage of KEAP1-NRF2 pathway alterations with mutations in 24.2% (KEAP1, 9.92%; NRF2, 13.69%; both, 0.59%) of samples. LUAD came in the second place with KEAP1-NRF2 pathway mutations in 20.82% (KEAP1, 17.68%; NRF2, 2.97%; both, 0.17%) of LUAD patients. In addition, the KEAP1-NRF2 pathway was altered in 12.07% ESCA patients (KEAP1, 2.19%; NRF2, 9.34%; both, 0.54%), in 11.7% of uterine corpus endometrial carcinoma (UCEC) patients (KEAP1, 3.19%; NRF2, 7.04%; both, 1.48%) and in 10.1% of HNSC (KEAP1, 3.95%; NRF2, 5.49%; both, 0.659%). Page 4/20 Page 4/20 Overview of the ESCA mutational landscape KEAP1 consists of 605 amino-acids, and 3 main domains with two mutations were detected in the BTB (broad-complex, tramtrack, and bric-a-brac) domain, three in the IVR (intervening region), and one in the Kelch domain, which is essential for the binding of NRF2 (Fig.3C). In the case of NRF2 structure, the majority of mutations (17) occurred in the crucial KEAP1-binding domain Neh2, and only one was found in the Neh1 domain. Identification of genes regulated by the KEAP1-NRF2 pathway in ESCA In order to identify genes that are regulated by the KEAP1-NRF2 pathway in ESCA, we subjected the TCGA-ESCA data set (22 KEAP1-NRF2-mutated versus 160 wild-type tumor samples) to differential gene expression analysis. We identified 896 DEGs with log FC >\|1\| (p < 0.05 with FDR adjustment) (Fig.4A). Of these DEGs, 403 were up-regulated and 493 were down-regulated (Additional file1: Table S1). Since the ultimate effect of changes in the KEAP1-NRF2 pathway is increased activity of NRF2; and hence the over- expression of its target genes, it was not surprising that several bonafideNRF2 target genes were among Page 5/20 the up-regulated genes including, AKR1C1, AKR1C2, AKR1B10, GSTM2, UGT1A6, AKR1C3, G6PD, GCLC, GCLM, GSTM3, GPX2, ABCC1, OSGIN1, SRXN1, and TXNRD1. The gene expression profiles of 22 KEAP1- NRF2-mutated and 160 wild-type ESCA patient samples were visualized on a heatmap produced by unsupervised hierarchical clustering, and major differences between the gene expression patterns enabled cluster analysis to discriminate between sample types. Significant differences or trends between KEAP1-NRF2-mutated and wild-type ESCA patient samples were detectable for DEGs with log FC> \|1\| (Fig.4B). CES1, AKR1C1, ADH7, ALDH3A, and CYP4F11 were the top five up-regulated genes in KEAP1- NRF2-mutated ESCA patient samples (Fig.4C), while PIGR, MUC13, TASPAN8, LGALS4, and OLFM4 were the top five down-regulated genes (Fig.4D). Epithelial-mesenchymal transition is regulated by the KEAP1-NRF2 pathway in ESCA The expression signatures of the hallmark gene sets, each containing 50 specific gene sets, were derived by concentrating multiple gene sets from the Molecular Signatures Database to represent well-defined biological statuses or courses. GSEA was performed to determine whether the identified gene sets showed statistically notable differences between the KEAP1-NRF2-mutated and their wild-type counterparts groups (Additional file 2: Table S2). Interestingly, four gene sets were up-regulated in the KEAP1-NRF2-mutated ESCA, namely, estrogen response late, hypoxia, reactive oxygen species pathway and EMT, the 4 gene sets were greatly enriched, with FDR < 0.05 (Fig.5). The gene set with the lowest FDR, namely, EMT (FDR = 0.001), which contained 194 genes was selected for further analysis. In order to specifically identify EMT genes that is associated with the KEAP1-NRF2 pathway in ESCA, we integrated the DEGs between KEAP1-NRF2-mutated and wild-type ESCA patient samples (log FC> \|1.5\|, FDR < 0.05) with the set of 194 EMT-enriched genes (Fig.6A) using Venny 2.1 web-based tool (http://bioinfogp.cnb.csic.es/tools/venny/index.html). Intriguingly, we found 11 common genes: SPP1, PTHLH, WNT5A, COL11A1, COL7A1, GPC1, SNAI2, ADAM12, FBN2, PFN2, and IGFBP3 (Fig.6B). (http://bioinfogp.cnb.csic.es/tools/venny/index.html). Intriguingly, we found 11 common genes: SPP1, PTHLH, WNT5A, COL11A1, COL7A1, GPC1, SNAI2, ADAM12, FBN2, PFN2, and IGFBP3 (Fig.6B). Evaluation of prognostic power of EMT gene-signature regulated by the KE In order to evaluate the prognostic power of these 5 genes (SPP1, WNT5A, PTHLH, PFN2, and GPC1) as an EMT-derived signature for KEAP1-NRF2 pathway alterations in ESCA, we first analyzed overall survival in the TCGA-ESCA cohort using the SurvExpress database. A total of 184 patient samples were divided into high-risk (n = 127) and low-risk groups (n = 57) based on their expression patterns (Fig.7A). The separation of risk groups was optimized using the ‘maximize risk group’ option provided in the SurvExpress database. The survival probability estimates in the two risk groups were visualized as Kaplan-Meier plots. Strikingly, overall survival analysis revealed that the patients in the high-risk group had poorer survival (HR = 1.67 (CI: 1.01-2.78); Log-Rank p = 0.04443) than the low-risk group (Figure.7B). Moreover, the Rao Giddings (GSE11595) cohort (34 ESCA patient samples) showed that the expression of SPP1, WNT5A, PTHLH, PFN2, and GPC1 in the high-risk group (n=17) was associated with poorer survival (HR = 6.84 (CI: 2.36 - 19.8); Log-Rank p = 3.072×e-5) than the low-risk group (n=17) (Fig.7C). In addition, we analyzed the overall survival in the Peters C.Fitzgerald (GSE19417) cohort available in the SurvExpress database. After optimized risk group separation, a total of 70 ESCA patient samples were divided into high-risk (n = 27) and low-risk groups (n = 43) based on their expression patterns (Fig.7D). The survival probability estimates in the two risk groups were represented as Kaplan-Meier plots. Similarly, overall survival analysis showed that the patients in the high-risk group had poorer survival (HR = 2.25 (CI: 1.34 - 3.79); Log-Rank p = 0.001659) than the low-risk group. As shown in Fig.5, The 5 genes were significantly over-expressed in the high risk patients compared to the low risk group. Moreover, the expression of SPP1, WNT5A, PTHLH, PFN2, and GPC1 successfully discriminated the survival of the ESCA high risk group from that of the low risk group in three ESCA cohorts (288 patients). These findings indicated that over-expression of the 5 genes is associated with a poor prognosis of ESCA and presents SPP1, WNT5A, PTHLH, PFN2, and GPC1 as an EMT signature based on changes in the KEAP1-NRF2 pathway in ESCA. Epithelial-mesenchymal transition genes were validated using ESCA cell lines In order to validate these 11 EMT-related genes as potential NRF2 targets, we used CCLE to download RNA-seq mRNA expression data of 1019 cell lines. Then, using the COSMIC data base we identified human ESCA cell lines that harbors NRF2 and/or KEAP1 mutations. We selected three cell lines (TE6, TE11 and KYSE180) as the KEAP1-NRF2-mutated group. In addition, we selected another three human ESCA cell lines that have neither NRF2 nor KEAP1 mutations (TE5, TE9 and KYSE150) as the wild-type group. Then, we carried out differential gene expression analysis between the two groups. As shown in GSTM3, AKR1C1 and TXNRD1 (well-known NRF2 targets) showed significant up-regulation in the mutated group compared to the wild-type counterpart, which ensures KEAP1-NRF2 pathway alteration in the group (Fig.6C). Furthermore, we investigated the expression of these 11 EMT-related genes between the two groups. Interestingly, five of the 11 genes (SPP1, WNT5A, PTHLH, PNF2, and GPC1) were significantly up-regulated in the mutated ESCA cell lines (Fig.6D). This finding suggests that these five genes are potential NRF2 targets. For further evidence, we investigated the presence of the putative and known antioxidant responsive elements (AREs), the NRF2 binding site, (Fig.6E) in the promoter region of Page 6/20 Page 6/20 these five genes by using ConTra V3 web tool. We performed insilico analysis within the –1 kb upstream the transcription start site (TSS) of the 5 genes. Interestingly, we identified highly conserved NRF2 binding sites (AREs) in the promoter regions of human PTHLH (positions: -71 and -916), WNT5A (position: -434), SPP1 (positions: -545,-605 and -870), PFN2 (positions:-737 and -864) and GPC1 (position: -458) (Fig.6F). these five genes by using ConTra V3 web tool. We performed insilico analysis within the –1 kb upstream the transcription start site (TSS) of the 5 genes. Interestingly, we identified highly conserved NRF2 binding sites (AREs) in the promoter regions of human PTHLH (positions: -71 and -916), WNT5A (position: -434), SPP1 (positions: -545,-605 and -870), PFN2 (positions:-737 and -864) and GPC1 (position: -458) (Fig.6F). Discussion The key role of KEAP1-NRF2 pathway alterations in developing drug- and radio-resistance in ESCA is well- established. The lack of specific biomarkers for KEAP1-NRF2 pathway alterations in ESCA motivated us to analyze TCGA-ESCA data in order to identify different biological processes that are regulated by KEAP1-NRF2 in ESCA; hence identifying new therapeutic targets and biomarkers to predict prognosis of ESCA patients. We found that the KEAP1-NRF2 pathway was altered in 12% of TCGA-ESCA patients. Swada et al., performed whole-exome sequence analysis of tumor and nontumor esophageal tissues collected from 144 patients with ESCA (11). They found that NRF2 was mutated in 16.7% of the patients Page 7/20 Page 7/20 and it was one of the most frequently mutated gene in their cohort while KEAP1 was mutated in almost 5% of patients. We performed GSEA to detect gene sets that showed statistically-notable differences between KEAP1- NRF2-mutated samples and their wild-type counterparts groups. Interestingly, 4 gene sets, namely, estrogen response late, hypoxia, reactive oxygen species pathway and epithelial mesenchymal transition, were greatly enriched, with FDR < 0.05. Since the ultimate effect of KEAP1-NRF2 pathway alterations is the stabilization of NRF2, the master transcriptional regulator of the cells antioxidant program(12), it was not surprising to find the reactive oxygen species pathway among the top pathways that show statistically notable differences between the KEAP1- NRF2-mutated and wild- type groups. Surprisingly, pathways such as EMT and hypoxia were greatly enriched. As EMT was more enriched, we selected EMT pathway to identify a NRF2-KEAP1 pathway signature in ESCA. EMT, an evolutionarily conserved developmental program, has been implicated in carcinogenesis and confers metastatic properties upon cancer cells by enhancing mobility, invasion, and resistance to apoptotic stimuli. Furthermore, EMT-derived tumor cells acquire stem cell properties and exhibit marked therapeutic resistance (13). Given these attributes, the complex biological process of the EMT has been heralded as a key hallmark of carcinogenesis, and targeting EMT pathways constitutes an attractive strategy for cancer treatment (14). Recently, it has been suggested that NRF2 contributes to malignant transformation of pancreatic duct epithelium through distinct EMT-related mechanisms accounting for an invasive phenotype (15). Furthermore, the expression of NRF2 is correlated with the lymph node metastasis of esophageal squamous cell carcinoma and blockage of NRF2 enhances the expression of E-cadherin , the well-known marker of epithelial cell polarity (16). Discussion In order to identify an EMT-derived signature for the KEAP1-NRF2 pathway in ESCA, we integrated the EMT gene list obtained from GSEA with the DEGs between the mutated and wild-type groups (log FC> \|1.5\|, FDR < 0.05). Interestingly, 11 genes were identified (SPP1, PTHLH, WNT5A, COL11A1, COL7A1, GPC1, SNAI2, ADAM12, FBN2, PFN2, and IGFBP3). Then, we validated these 11 genes by subjecting the KEAP1-NRF2-muatated and wild-type ESAC cell lines to differential gene expression analysis. Intriguingly, only 5 of the 11 genes showed significant up- regulation between the two groups (SPP1, WNT5A, PTHLH, PFN2, and GPC1). Further, we evaluated the prognostic power of these 5 genes using three ESCA cohorts (288 patients), and we found that the over- expression of these five genes were associated with a poor prognosis in ESCA. In agreement with our analysis, SPP1 and EMT have been shown by bioinformatics analysis to have a close association in colorectal cancer (17). Moreover, Osteopontin, encoded by SPP1 promotes the EMT in hepatocellular carcinoma through regulating vimentin (18), and high SPP1 expression in hepatocellular carcinoma is associated with poor survival outcome (19). Additionally, up-regulation of WNT5A has been suggested to promote EMT and metastasis in pancreatic cancer models, which involves activation of β- catenin-dependent canonical Wnt signaling(20). Furthermore, it has been illustrated that WNT5A promotes EMT and metastasis in non-small-cell lung cancer (NSCLC), and high WNT5A expression is associated with poor prognosis in NSCLC patients (21). In addition, parathyroid hormone related-protein, encoded by PTHLH has been found to promote EMT in prostate and pancreatic cancers (22, 23). It has been indicated that PTHLH is a poor prognosis marker and promotes cell growth of HNSC. Besides, it has been illustrated that inhibition of PFN2 hinders cell invasion and migration, as well as induces an EMT Page 8/20 Page 8/20 phenotype, including increased expression of epithelial marker E-cadherin, decreased mesenchymal marker Vimentin, Snail, Slug and ZEB1, and morphological changes in ESCA cells in vitro (24). High PFN2 expression independently predicts poor overall survival in primary HNSC and ESCA (24,25). Moreover, Over-expression of GPC1 activates EMT which then increases invasion and migration in colorectal cancer and ESCA (26-28). Additionally, it has been suggested that GPC1 plays an important role in regulating TGF-β-mediated EMT and stemness, and could be a potential future therapeutic target to prevent progression of gastric cancer (29). Discussion It has been pointed out that GPC1 is over-expressed and implies a poor prognosis in several cancers including, ESCA, uterine cervical cancer, pancreatic cancer, and methothelioma (30-34). Altogether, the above evidence suggests an oncogenic role of the 5-gene signature in many cancers. Conclusions Our study identified an EMT-derived gene signature regulated by the KEAP1-NRF2 pathway that is strongly associated with tumorigenesis, metastasis, and drug resistance in ESCA. This 5-gene signature provides potential biomarkers and therapeutic targets for ESCA patients in whom KEAP1-NRF2 pathway is activated. List Of Abbreviations List Of Abbreviations Page 9/20 KEAP1 Kelch‑like ECH‑associated protein 1 NRF2 Nuclear factor erythroid 2‑related 2 LUAD Lung adenocarcinoma NSCLC Non-small-cell lung cancer TCGA The cancer genome atlas ARE Antioxidant response element ESCA ESE Esophageal cancer EMT Epithelial-mesenchymal transition CCLE Cancer cell line encyclopedia DEGs Differentially expressed genes PTHLH Parathyroid hormone like hormone FDR False discovery rate GSEA Gene set enrichment analysis Funding This work was supported by the National Natural Science Foundation of China (31571476, 31971188, and 31370772). Availability of data and materials The datasets used in this study are publicly available as noted in the text. Author information Author information Page 10/20 Consent for publication Not applicable Competing interests The authors declare that they have no competing interests. Affiliations 1Department of Biochemistry and Department of Thoracic Surgery of the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310003, PR China; 1Department of Biochemistry and Department of Thoracic Surgery of the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310003, PR China; 2Department of Pharmacology and Cancer Institute, The Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310009, PR China; 2Department of Pharmacology and Cancer Institute, The Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310009, PR China; 3Labeled Compounds Department, Hot Labs Center, Egyptian Atomic Energy Authority, Cairo 13759, Egypt; 4Radiation Biology Department, National Center for Radiation Research and Technology, Egyptian Atomic Energy Authority, Cairo 13759, Egypt; 4Radiation Biology Department, National Center for Radiation Research and Technology, Egyptian Atomic Energy Authority, Cairo 13759, Egypt; Contributions ME conducted the study. XT supervised the study. AH and XJW assisted in data interpretation. ME and XT wrote the manuscript with assistance from all authors. All authors read and approved the manuscript. References 1. Testa U, Castelli G, Pelosi E. Esophageal Cancer: Genomic and Molecular Characterization, Stem Cell Compartment and Clonal Evolution. Medicines (Basel). 2017;4(3):67. 1. Testa U, Castelli G, Pelosi E. Esophageal Cancer: Genomic and Molecular Characterization, Stem Cell Compartment and Clonal Evolution. Medicines (Basel). 2017;4(3):67. Page 11/20 Page 11/20 2. Lignitto L, LeBoeuf SE, Homer H, Jiang S, Askenazi M, Karakousi TR, et al. Nrf2 Activation Promotes Lung Cancer Metastasis by Inhibiting the Degradation of Bach1. Cell. 2019;178(2):316-29.e18. 3. Hammad A, Namani A, Elshaer M, Wang XJ, Tang X. “NRF2 addiction” in lung cancer cells and its impact on cancer therapy. Cancer Letters. 2019;467:40-9. 4. Namani A, Matiur Rahaman M, Chen M, Tang X. Gene-expression signature regulated by the KEAP1- NRF2-CUL3 axis is associated with a poor prognosis in head and neck squamous cell cancer. BMC Cancer. 2018;18(1):46. 5. Elshaer M, ElManawy AI, Hammad A, Namani A, Wang XJ, Tang X. Integrated data analysis reveals significant associations of KEAP1 mutations with DNA methylation alterations in lung adenocarcinomas. Aging (Albany NY). 2020;12(8):7183-206. 6. Namani A, Zheng Z, Wang XJ, Tang X. Systematic Identification of Multi Omics-based Biomarkers in <i>KEAP1</i> Mutated TCGA Lung Adenocarcinoma. Journal of Cancer. 2019;10(27):6813-21. 7. Goldman M, Craft B, Hastie M, Repečka K, Kamath A, McDade F, et al. The UCSC Xena platform for public and private cancer genomics data visualization and interpretation. bioRxiv. 2019:326470. 8. Guan X, Cai M, Du Y, Yang E, Ji J, Wu J. CVCDAP: an integrated platform for molecular and clinical analysis of cancer virtual cohorts. Nucleic Acids Research. 2020;48(W1):W463-W71. 9. Kreft Ł, Soete A, Hulpiau P, Botzki A, Saeys Y, De Bleser P. ConTra v3: a tool to identify transcription factor binding sites across species, update 2017. Nucleic Acids Research. 2017;45(W1):W490-W4. 10. Aguirre-Gamboa R, Gomez-Rueda H, Martínez-Ledesma E, Martínez-Torteya A, Chacolla-Huaringa R, Rodriguez-Barrientos A, et al. SurvExpress: an online biomarker validation tool and database for cancer gene expression data using survival analysis. PLoS One. 2013;8(9):e74250-e. 11. Sawada G, Niida A, Uchi R, Hirata H, Shimamura T, Suzuki Y, et al. Genomic Landscape of Esophageal Squamous Cell Carcinoma in a Japanese Population. Gastroenterology. 2016;150(5):1171-82. 12. Jaramillo MC, Zhang DD. The emerging role of the Nrf2-Keap1 signaling pathway in cancer. Genes & development. 2013;27(20):2179-91. 12. Jaramillo MC, Zhang DD. The emerging role of the Nrf2-Keap1 signaling pathway in cancer. Genes & development. 2013;27(20):2179-91. 13. Mittal V. Epithelial Mesenchymal Transition in Tumor Metastasis. Annual Review of Pathology: Mechanisms of Disease. 2018;13(1):395-412. 13. Mittal V. Epithelial Mesenchymal Transition in Tumor Metastasis. Annual Review of Pathology: Mechanisms of Disease. 2018;13(1):395-412. 14. Roche J. The Epithelial-to-Mesenchymal Transition in Cancer. Cancers (Basel). 2018;10(2):52. 14. Roche J. The Epithelial-to-Mesenchymal Transition in Cancer. Cancers (Basel). 2018;10(2):52. 15. Page 11/20 Arfmann-Knübel S, Struck B, Genrich G, Helm O, Sipos B, Sebens S, et al. The Crosstalk between Nrf2 and TGF-β1 in the Epithelial-Mesenchymal Transition of Pancreatic Duct Epithelial Cells. PLoS One. 2015;10(7):e0132978-e. 15. Arfmann-Knübel S, Struck B, Genrich G, Helm O, Sipos B, Sebens S, et al. The Crosstalk between Nrf2 and TGF-β1 in the Epithelial-Mesenchymal Transition of Pancreatic Duct Epithelial Cells. PLoS One. 2015;10(7):e0132978-e. 16. Shen H, Yang Y, Xia S, Rao B, Zhang J, Wang J. Blockage of Nrf2 suppresses the migration and invasion of esophageal squamous cell carcinoma cells in hypoxic microenvironment. Diseases of the esophagus : official journal of the International Society for Diseases of the Esophagus. 2014;27(7):685-92. Page 12/20 Page 12/20 17. Xu C, Sun L, Jiang C, Zhou H, Gu L, Liu Y, et al. SPP1, analyzed by bioinformatics methods, promotes the metastasis in colorectal cancer by activating EMT pathway. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie. 2017;91:1167-77. 18. Dong Q, Zhu X, Dai C, Zhang X, Gao X, Wei J, et al. Osteopontin promotes epithelial-mesenchymal transition of hepatocellular carcinoma through regulating vimentin. Oncotarget. 2016;7(11):12997- 3012. 19. Menyhárt O, Nagy Á, Győrffy B. Determining consistent prognostic biomarkers of overall survival and vascular invasion in hepatocellular carcinoma. Royal Society Open Science. 2018;5(12):181006. 20. Bo H, Zhang S, Gao L, Chen Y, Zhang J, Chang X, et al. Upregulation of WNT5A promotes epithelial- to-mesenchymal transition and metastasis of pancreatic cancer cells. BMC Cancer. 2013;13(1):496. 21. Wang B, Tang Z, Gong H, Zhu L, Liu X. WNT5A promotes epithelial-to-mesenchymal transition and metastasis in non-small-cell lung cancer. Biosci Rep. 2017;37(6):BSR20171092. 22. Ongkeko WM, Burton D, Kiang A, Abhold E, Kuo SZ, Rahimy E, et al. Parathyroid hormone related- protein promotes epithelial-to-mesenchymal transition in prostate cancer. PLoS One. 2014;9(1):e85803-e. 23. Pitarresi JR, Norgard RJ, Stanger BZ, Rustgi AK. Abstract B43: p120 catenin loss drives pancreatic cancer EMT and metastasis through activation of PTHrP-mediated calcium signaling. Cancer Research. 2019;79(24 Supplement):B43-B. 24. Cui X-B, Zhang S-M, Xu Y-X, Dang H-W, Liu C-X, Wang L-H, et al. PFN2, a novel marker of unfavorable prognosis, is a potential therapeutic target involved in esophageal squamous cell carcinoma. J Transl Med. 2016;14(1):137-. 25. Liu J, Wu Y, Wang Q, Liu X, Liao X, Pan J. Bioinformatic analysis of PFN2 dysregulation and its prognostic value in head and neck squamous carcinoma. Future Oncology. 2018;14(5):449-59. 26. Figures Figures Figure 1 Page 11/20 Li J, Li B, Ren C, Chen Y, Guo X, Zhou L, et al. The clinical significance of circulating GPC1 positive exosomes and its regulative miRNAs in colon cancer patients. Oncotarget. 2017;8(60):101189-202. 27. Li J, Chen Y, Zhan C, Zhu J, Weng S, Dong L, et al. Glypican-1 Promotes Tumorigenesis by Regulating the PTEN/Akt/β-Catenin Signaling Pathway in Esophageal Squamous Cell Carcinoma. Digestive Diseases and Sciences. 2019;64(6):1493-502. 28. Li Y, Li M, Shats I, Krahn JM, Flake GP, Umbach DM, et al. Glypican 6 is a putative biomarker for metastatic progression of cutaneous melanoma. PLoS One. 2019;14(6):e0218067. 28. Li Y, Li M, Shats I, Krahn JM, Flake GP, Umbach DM, et al. Glypican 6 is a putative biomarker for metastatic progression of cutaneous melanoma. PLoS One. 2019;14(6):e0218067. 29. Wang S, Wu Z, Zhou M, Liao W. Effect of GPC1 on epithelial-to-mesenchymal transition and stemness and interaction with ITGB1 in gastric cancer. Journal of Clinical Oncology. 2017;35(15_suppl):e15580-e. 30. Hara H, Takahashi T, Serada S, Fujimoto M, Ohkawara T, Nakatsuka R, et al. Over-expression of glypican-1 implicates poor prognosis and their chemoresistance in oesophageal squamous cell carcinoma. British journal of cancer. 2016;115(1):66-75. 30. Hara H, Takahashi T, Serada S, Fujimoto M, Ohkawara T, Nakatsuka R, et al. Over-expression of glypican-1 implicates poor prognosis and their chemoresistance in oesophageal squamous cell carcinoma. British journal of cancer. 2016;115(1):66-75. Page 13/20 31. Matsuzaki S, Serada S, Hiramatsu K, Nojima S, Matsuzaki S, Ueda Y, et al. Anti-glypican-1 antibody- drug conjugate exhibits potent preclinical antitumor activity against glypican-1 positive uterine Page 13/20 31. Matsuzaki S, Serada S, Hiramatsu K, Nojima S, Matsuzaki S, Ueda Y, et al. Anti-glypican-1 antibody- drug conjugate exhibits potent preclinical antitumor activity against glypican-1 positive uterine cervical cancer. 2018;142(5):1056-66. 32. Duan L, Hu XQ, Feng DY, Lei SY, Hu GH. GPC-1 may serve as a predictor of perineural invasion and a prognosticator of survival in pancreatic cancer. Asian journal of surgery. 2013;36(1):7-12. 32. Duan L, Hu XQ, Feng DY, Lei SY, Hu GH. GPC-1 may serve as a predictor of perineural invasion and a prognosticator of survival in pancreatic cancer. Asian journal of surgery. 2013;36(1):7-12. 33. Amatya VJ, Kushitani K, Kai Y, Suzuki R, Miyata Y, Okada M, et al. Glypican-1 immunohistochemistry is a novel marker to differentiate epithelioid mesothelioma from lung adenocarcinoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2018;31(5):809-15. 33. Page 11/20 Amatya VJ, Kushitani K, Kai Y, Suzuki R, Miyata Y, Okada M, et al. Glypican-1 immunohistochemistry is a novel marker to differentiate epithelioid mesothelioma from lung adenocarcinoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2018;31(5):809-15. 34. Kato D, Yaguchi T, Iwata T, Katoh Y, Morii K, Tsubota K, et al. GPC1 specific CAR-T cells eradicate established solid tumor without adverse effects and synergize with anti-PD-1 Ab. eLife. 2020;9:e49392. 34. Kato D, Yaguchi T, Iwata T, Katoh Y, Morii K, Tsubota K, et al. GPC1 specific CAR-T cells eradicate established solid tumor without adverse effects and synergize with anti-PD-1 Ab. eLife. 2020;9:e49392. Figure 1 Schematic diagram showing the analysis overflow that was followed in this study. Schematic diagram showing the analysis overflow that was followed in this study. Schematic diagram showing the analysis overflow that was followed in this study. Page 14/20 Page 14/20 Figure 2 General mutational landscape A) Bar chart representing TCGA-pan cancer analysis of KEAP1-NRF2 pathway alterations in diffrent cancers. B) Landscape of genetic alterations across the 185 ESCA samples. The samples are sorted by mutation rates (top bars), while genes are sorted by the proportion of altered samples (left bars) Figure 2 General mutational landscape A) Bar chart representing TCGA-pan cancer analysis of KEAP1-NRF2 pathway alterations in diffrent cancers. B) Landscape of genetic alterations across the 185 ESCA samples. The samples are sorted by mutation rates (top bars), while genes are sorted by the proportion of altered samples (left bars). General mutational landscape A) Bar chart representing TCGA-pan cancer analysis of KEAP1-NRF2 pathway alterations in diffrent cancers. B) Landscape of genetic alterations across the 185 ESCA samples. The samples are sorted by mutation rates (top bars), while genes are sorted by the proportion of altered samples (left bars). Page 15/20 Page 15/20 Figure 3 Mutational landscape of ESCA samples with KEAP1 and/or NRF2 mutations. A) Landscape of genetic alterations across the 22 ESCA samples with KEAP1 and/or NRF2 mutations. The samples are sorted by mutation rates (top bars), while genes are sorted by the proportion of altered samples (left bars). B) Differential mutational analysis between KEAP1-NRF2-mutated and wild-type ESCA samples. C) Lollipop plot showing the locations of mutations in the functional domains of NRF2 protein. D) Lollipop plot showing the locations of mutations in the functional domains of KEAP1 protein. The lollipops show the locations of the mutations as identified by whole-exon sequencing. Figure 3 Figure 3 Figure 3 Mutational landscape of ESCA samples with KEAP1 and/or NRF2 mutations. A) Landscape of genetic alterations across the 22 ESCA samples with KEAP1 and/or NRF2 mutations. The samples are sorted by mutation rates (top bars), while genes are sorted by the proportion of altered samples (left bars). B) Differential mutational analysis between KEAP1-NRF2-mutated and wild-type ESCA samples. C) Lollipop plot showing the locations of mutations in the functional domains of NRF2 protein. D) Lollipop plot showing the locations of mutations in the functional domains of KEAP1 protein. The lollipops show the locations of the mutations as identified by whole-exon sequencing. Mutational landscape of ESCA samples with KEAP1 and/or NRF2 mutations. A) Landscape of genetic alterations across the 22 ESCA samples with KEAP1 and/or NRF2 mutations. The samples are sorted by mutation rates (top bars), while genes are sorted by the proportion of altered samples (left bars). B) Differential mutational analysis between KEAP1-NRF2-mutated and wild-type ESCA samples. C) Lollipop plot showing the locations of mutations in the functional domains of NRF2 protein. D) Lollipop plot showing the locations of mutations in the functional domains of KEAP1 protein. The lollipops show the locations of the mutations as identified by whole-exon sequencing. Page 16/20 Figure 4 Differential gene expression analysis. A) Volcano plot showing the distribution of DEGs between KEAP1- NRF2-mutated and wild-type ESCA patient samples based on significance and fold change. B) Heatmap showing the top DEGs between KEAP1- NRF2 mutated and wild-type ESCA patient samples with Log FC>\|1\| and FDR <0.05. C) Box plots showing the top 5 overexpressed genes between KEAP1- NRF2- mutated and wild-type ESCA patient samples. D) Box plots showing the top 5 down-regulated genes between KEAP1-NRF2-mutated and wild-type ESCA patient samples. Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles; outliers are represented by dots. Figure 4 Differential gene expression analysis. A) Volcano plot showing the distribution of DEGs between KEAP1- NRF2-mutated and wild-type ESCA patient samples based on significance and fold change. B) Heatmap showing the top DEGs between KEAP1- NRF2 mutated and wild-type ESCA patient samples with Log FC>\|1\| and FDR <0.05. C) Box plots showing the top 5 overexpressed genes between KEAP1- NRF2- mutated and wild-type ESCA patient samples. D) Box plots showing the top 5 down-regulated genes between KEAP1-NRF2-mutated and wild-type ESCA patient samples. Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles; outliers are represented by dots. Page 17/20 Page 17/20 Figure 5 Gene set enrichment analysis. Enrichment plots of four gene sets that are importantly differentiated between KEAP1-NRF2-mutated and wild-type ESCA samples. Figure 5 Figure 5 Gene set enrichment analysis. Enrichment plots of four gene sets that are importantly differentiated between KEAP1-NRF2-mutated and wild-type ESCA samples. Gene set enrichment analysis. Enrichment plots of four gene sets that are importantly differentiated between KEAP1-NRF2-mutated and wild-type ESCA samples. Page 18/20 Page 18/20 Figure 6 Figure 6 Identification of the EMT signature of ESCA patients with altered KEAP1-NRF2 pathway A) Venny diagram showing the overlapping between the EMT-enriched gene set and DEGs between KEAP1-NRF2- mutated and wild-type ESCA patient samples. B) Box plots showing the differential expression of 11 overlapped EMT genes between KEAP1- NRF2-mutated and wild-type ESCA patient samples. C) Bar chart showing differential mRNA expression of some well-known NRF2 targets between ESCA cell lines with KEAP1 and/or NRF2 mutations and their wild-type counterparts. D) Bar chart showing 5 EMT genes that were significantly differentially expressed between ESCA cell lines with KEAP1 and/or NRF2 mutations and their wild-type counterparts. (p < 0.05, * p < 0.01, *** p < 0.001). E) The NRF2 binding motif as provided by JASPER. (F) Schematic representation of the locations of insilico-predicted NRF2 binding sites (AREs) in the promoter regions of the human SPP1, WNT5A, PTHLH, PFN2, and GPC1genes. Figure 7 Five-gene signature predicts poor survival in three independent cohorts. (A) Box plots showing the expression differences of the 5-gene signature in low (green) and high (red) risk groups of TCGA–ESCA patients (y-axis, gene expression value of each gene).(B) Kaplan-Meier survival plots showing that high expression of the 5-gene signature is associated with poor survival in TCGA–ESCA patients. (C) The Rao Giddings (GSE11595) cohort. D) The Peters C.Fitzgerald (GSE19417) cohort. Red, high-risk group; green, low-risk group; top right corner inset, numbers of high- and low-risk samples (x-axis, time; y-axis, overall survival probability; HR, hazard ratio; CI, confidence interval). Figure 7 Figure 8 Schematic diagram summarizing the findings of this study. Figure 7 Five-gene signature predicts poor survival in three independent cohorts. (A) Box plots showing the expression differences of the 5-gene signature in low (green) and high (red) risk groups of TCGA–ESCA patients (y-axis, gene expression value of each gene).(B) Kaplan-Meier survival plots showing that high expression of the 5-gene signature is associated with poor survival in TCGA–ESCA patients. (C) The Rao Giddings (GSE11595) cohort. D) The Peters C.Fitzgerald (GSE19417) cohort. Red, high-risk group; green, low-risk group; top right corner inset, numbers of high- and low-risk samples (x-axis, time; y-axis, overall survival probability; HR, hazard ratio; CI, confidence interval). Page 19/20 Page 19/20 Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. Additionalfile2.xlsx Additionalfile2.xlsx Additionalfile1.xlsx Additionalfile1.xlsx Page 20/20 Page 20/20
https://openalex.org/W4220724808	https://ora.ox.ac.uk/objects/uuid:2654aa37-b143-40d2-bf85-fedc1c730a35/files/srv042v24t	English	null	Unravelling the processes between phenotypic plasticity and population dynamics in migratory birds	Journal of animal ecology	2,022	cc-by	10,789	R E S E A R C H A R T I C L E R E S E A R C H A R T I C L E Unravelling the processes between phenotypic plasticity and population dynamics in migratory birds n Lei1 \| Xunqiang Mo3 \| Chris J. Hassell4 \| Zhengwang Zhang1 \| Jin Liu1,2 \| Weipan Lei1 \| Xunqiang Mo3 \| Chris J. Hassell4 \| Zhengwang Zhang1 \| Tim Coulson2 Jin Liu1,2 \| Weipan Lei1 \| Xunqiang Mo3 \| Chris J. Hassell4 \| Zhengwang Zhang1 \| Tim Coulson2 1Key Laboratory for Biodiversity Science and Ecological Engineering, College of Life Sciences, Beijing Normal University, Beijing, China 1Key Laboratory for Biodiversity Science and Ecological Engineering, College of Life Sciences, Beijing Normal University, Beijing, China 2Department of Zoology, University of Oxford, Oxford, UK 3School of Geographic and Environmental Sciences, Tianjin Normal University, Tianjin, China 4Global Flyway Network, Broome, WA, Australia Correspondence Zhengwang Zhang Email: zzw@bnu.edu.cn Funding information National Natural Science Foundation of China, Grant/Award Number: 31830089, 31572288 and 31801985; China Scholarship Council, Grant/ Award Number: 201806040101; Fundamental Research Funds for the Central Universities; Paulson Institute; SEE Foundation; Spinoza Premium of Netherlands Organisation Prize; Vogelbescherming Nederland; WWF Netherlands; MAVA Foundation Handling Editor: Jean-Michel Gaillard This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. Journal of Animal Ecology published by John Wiley & Sons Ltd on behalf of British Ecological Society. J Anim Ecol. 2022;91:983–995. Abstract 1. Populations can rapidly respond to environmental change via adaptive pheno- typic plasticity, which can also modify interactions between individuals and their environment, affecting population dynamics. Bird migration is a highly plastic resource-tracking tactic in seasonal environments. However, the link between the population dynamics of migratory birds and migration tactic plasticity is not well-understood. 2Department of Zoology, University of Oxford, Oxford, UK 4Global Flyway Network, Broome, WA, Australia 2. The quality of staging habitats affects individuals' migration timing and energy budgets in the course of migration and can consequently affect individuals' breeding and overwintering performance, and impact population dynamics. Given staging habitats being lost in many parts of the world, our goal is to in- vestigate responses of individual migration tactics and population dynamics in the face of loss of staging habitat and to identify the key processes connecting them. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Received: 20 August 2021 \| Accepted: 17 February 2022 DOI: 10.1111/1365-2656.13686 Received: 20 August 2021 \| Accepted: 17 February 2022 DOI: 10.1111/1365-2656.13686 Received: 20 August 2021 \| Accepted: 17 February 2022 DOI: 10.1111/1365-2656.13686 R E S E A R C H A R T I C L E 1 \| INTRODUCTION carry-over and density-dependent effects (Harrison et al., 2011; Newton, 2007). The individual state in one stage can influence indi- vidual performance in subsequent stages, and the change in popula- tion size in one stage can influence per capita rates and consequently regulate population size in later stages (Ratikainen et al., 2008; Ryan Norris & Marra, 2007; Studds & Marra, 2005). The tactic an individual follows while at the staging area can consequently affect breeding and overwintering performance, and impact population dynamics. 1 Populations can rapidly respond to environmental change via adaptive phenotypic plasticity, and this allows them to cope with profound environmental impacts (Coulson et al., 2017; Piersma & Drent, 2003; Pigliucci, 2001). Plasticity modifies interactions be- tween individuals and their environment, ultimately affecting pop- ulation dynamics (Miner et al., 2005). Migration can be an adaptive plastic tactic in seasonal environments (Lack, 1968; Newton, 2007) that allows individuals to increase reproductive output by avoid- ing unsuitable ecological conditions (Hedenström, 2008; Winkler et al., 2014). Plasticity of migration tactics enables migratory species to respond to environmental changes in multiple ways, such as by altering migratory routes (Dolman & Sutherland, 1995; Sutherland & Crockford, 1993), timing of migration (Balbontín et al., 2009; Gienapp et al., 2007) and through diet (Parrish, 2000). However, the link between the population dynamics of migratory species and mi- gration tactics plasticity is not well-understood. Staging habitat for migratory waterbirds in the Yellow Sea is being lost in significant quantities, primarily due to land reclama- tion for infrastructure development and aquaculture (Bi et al., 2012; Murray et al., 2014; Yang et al., 2011). Although illegal hunting, human disturbance, exotic Spartina invasion and climate change im- pacting other parts of the annual cycle also threaten migratory birds along the East Asian Australasian Flyway (EAAF), the loss of staging habitat has been suggested as a primary cause of population declines (Amano et al., 2012; Ma et al., 2014; Melville et al., 2016; Studds et al., 2017). It is presumably because staging habitat in this system is the stage of the annual cycle where density dependence is stron- gest (Sutherland, 1996b). Habitat loss in staging habitats can reduce food resources, decrease foraging and fat accumulation rates of mi- grants (Baker et al., 2004; Morrision, 2006; Verkuil et al., 2012), in- crease competition and interference in the population and can have significant consequences for population regulation (Newton, 2007; Sutherland, 1996b). 1 \| INTRODUCTION However, the way in which individuals respond to such changes, as well as the processes and mechanisms that cause population declines, are yet to be generally established. Migratory waterbirds along the EAAF provides a unique system to explore how individuals respond to changes in one life-history stage and how this response would influence their populations. Bird migration is a resource-tracking tactic that aims to optimize a bird's energy budget in the face of fluctuating resources in sea- sonal environments and in the face of strong competition (Alerstam et al., 2003; Cox, 1968; Somveille et al., 2018; Winger et al., 2019). Migration is energetically costly, so birds build up fat reserves. However, carrying a large energy reserve increases flight costs and can also attract predators (Alerstam & Lindström, 1990). One tactic to minimize such costs is to stop over several times during the jour- neys between breeding and wintering sites to refuel (Piersma, 1988). For individuals to remain in favourable environments across their migration route, they must carefully manage the timing of depar- ture and arrival (Alerstam et al., 2003; Alerstam & Lindström, 1990; Winkler et al., 2014). In general, individuals that arrive at breeding grounds earlier have higher reproductive success than those that arrive later (Marra et al., 1998; Norris et al., 2004), and selection favours individuals that minimize the time spent travelling during the northward migration (Lindstrom & Alerstam, 1992). Migratory birds usually spend much longer accumulating energy reserves in staging areas than in flying (Hedenström & Alerstam, 1997). Therefore, the total time spent on migration is consequently strongly influenced by the quality of, and an individual's behaviour at, staging areas (Erni et al., 2002; Hedenström & Alerstam, 1997). To examine the effects of habitat loss within staging habitat on individual migration tactics and population dynamics, we use the- oretical modelling to generate hypotheses that we next test with empirical data. First, we conducted an individual-based modelling exercise, building a stylized full-annual-cycle model in which individ- uals follow the same migrating rules. The model led us to hypothe- size that the loss of staging habitat generates plasticity in migration tactics, with individuals staying in the remaining staging site for longer to obtain food due to a reduction in per capita food avail- ability. Correspondence Zhengwang Zhang Funding information National Natural Science Foundation of China, Grant/Award Number: 31830089, 31572288 and 31801985; China Scholarship Council, Grant/ Award Number: 201806040101; Fundamental Research Funds for the Central Universities; Paulson Institute; SEE Foundation; Spinoza Premium of Netherlands Organisation Prize; Vogelbescherming Nederland; WWF Netherlands; MAVA Foundation 3. We started by constructing and analysing a general full-annual-cycle individual- based model with a stylized migratory population to generate hypotheses on how changes in the size of staging habitat might drive changes in individual stopover duration and population dynamics. Next, through the interrogation of survey data, we tested these hypotheses by analysing population trends and stopover duration of migratory waterbirds experiencing the loss of staging habitat. 4. Our modelling exercise led to us posing the following hypotheses: the loss of staging habitat generates plasticity in migration tactics, with individuals remain- ing on the staging habitat for longer to obtain food due to a reduction in per capita food availability. The subsequent increasing population density on the staging habitat has knock-on effects on population dynamics in the breeding and overwintering stage. Our empirical results were consistent with the model- ling predictions. wileyonlinelibrary.com/journal/jane Journal of Animal Ecology 984 LIU et al. LIU et al. 5. Our results demonstrate how environmental change that impacts one energeti- cally costly life-history stage in migratory birds can have population dynamic impacts across the entire annual cycle via phenotypic plasticity. K E Y W O R D S annual cycle, carrying capacity, density-dependent, individual-based model, loss of staging habitat, migratory birds, population dynamics, stopover duration K E Y W O R D S annual cycle, carrying capacity, density-dependent, individual-based model, loss of staging habitat, migratory birds, population dynamics, stopover duration 2.1.1 \| Model description Our IBMs include three types of habitats in the model landscape, which are wintering habitat (W), breeding habitat (B) and staging habitat (S). To examine the role of the remaining staging site on individual migratory tactics and population dynamics, we speci- fied two alternative sites in the staging habitat—the S1 and S2 sites (Figure 1a). The size of the S1 site remained constant across all simulations, while the habitat size of the S2 site can be adjusted (Figure 1b). The total staging habitat was defined as the smallest rectangle that could encompass S1 and S2 sites, the size of the rectangle has no impact on model results, it is simply a boundary that allows individuals to move between S1 and S2 sites easily. The rest of the grid cells in the landscape are non-habitat, where individuals pass by during migration but do not stop. Each grid cell within W, B, S1 and S2 sites contained renewing food resources, while the rest of the landscape did not include any food resources. Food resources at each grid cell renewed each time step at the habitat-specified food recovery rate after consumption. Each time step in this model represented 1 day such that 1 year was com- prised of 365 steps. 1 \| INTRODUCTION The increasing population density in the staging habitat has knock-on effects on breeding and overwintering stages that impact the population dynamics, via impacts on survival and reproduction For migratory species, all stages of the annual cycle are closely linked at both the individual and population levels, through LIU et al. Journal of Animal Ecology 985 rates. We used this model to make predictions about changes in tac- tics that might influence population dynamics in the staging, breed- ing and overwintering grounds. We did this by examining wherein the life cycle density dependence operated most strongly. Next, using 13 years of survey data on 148 waterbird species in total, we examined whether observed empirical trends were consistent with hypotheses we generated from our individual-based model. Our em- pirical analyses were consistent with the theoretical hypotheses our model suggested: we found that in the EAAF system where the total population size is declining along the whole flyway, population den- sity increases at the remaining staging site as the size of the staging habitat decreases, with prolonged stopover duration for individuals. We conclude that environmental change effects on one life-history stage in migratory birds can consequently have population dynamic impacts across the entire annual cycle via phenotypic plasticity. rules in our models; (c) individuals who meet the condition for repro- duction produce once each year; and (d) males are not limiting and can be ignored such that we construct a female-only model. 2.1 \| The individual-based model The individual-based model (IBM) we constructed is a stylized model. The basic assumptions of our model are as follows: (a) the stages of the annual cycle in our model includes northward migration, staging in the course of northward migration, breeding, southward migra- tion and overwintering; (b) all individuals followed the same set of FI G U R E 1 A schematic diagram of the landscape and movement rules of the individual-based model. (a) Model landscape. ① represents the case in which food is available in the neighbouring grid cells for individuals, and ② represents the case in which food is unavailable in the neighbouring cells. The movement rules of individuals for ① and ② are shown in (c1) and (c2) respectively. (b) Three selected scenarios for the landscape in which the size of the S2 site differed. (c) Movement rules, (c1) ‘move’ under the case of ①, (c2) ‘move’ under the case of ② and (c3) the order of ‘fly’ FI G U R E 1 A schematic diagram of the landscape and movement rules of the individual-based model. (a) Model landscape. ① represents the case in which food is available in the neighbouring grid cells for individuals, and ② represents the case in which food is unavailable in the neighbouring cells. The movement rules of individuals for ① and ② are shown in (c1) and (c2) respectively. (b) Three selected scenarios for the landscape in which the size of the S2 site differed. (c) Movement rules, (c1) ‘move’ under the case of ①, (c2) ‘move’ under the case of ② and (c3) the order of ‘fly’ LIU et al. 986 LIU et al. Journal of Animal Ecology We only considered the female component of the population and characterized each individual by identity, age, reproduction status and energy reserves. The behaviours of each individual in the model in- cluded: fly, move, search for food, eat, orient, mature, reproduce and die. ‘Fly’ was the movement across habitats during the course of migration, the speed of ‘fly’ was three grid cells per time step. ‘Move’ was the move- ment within each habitat during the course of overwintering, staging and breeding, when individuals reached the boundary of each habitat, the be- haviour of ‘move’ started, the speed of ‘move’ was one grid cell per time step. Either type of movement consumes part of an individual's energy reserves at each time step. Birds searched for food by following a ran- dom movement rule. They compared the food value on their neighbour- ing eight grid cells to the current one, if there was a highest food value, birds moved towards this cell; if food values were the same on more than one cell, birds randomly chose a direction to move (Figure 1c1). Birds ate and increased their energy reserves when food was available; if food was unavailable, then birds randomly moved to a neighbouring cell and did not eat (Figure 1c2). There was no randomness of food acquisition; as long as food resources were available, the behaviour of ‘eat’ hap- pened and energy was stored. When the energy reserve of an individual reached the energy threshold for departure, or the time reached for the latest possible departure arrived, the individual first oriented, adjusting its facing to the centre of the next destination habitat in the next time step, then flew towards it. Individuals whose energy reserves reached the threshold for reproduction matured and reproduced once each year. Reproduction also consumed individuals' energy reserves. Hatchlings were set with an initial value of age, reproduction status and energy re- serves. Individuals aged 15 years, or with zero energy reserves, died and were removed from the population. Details of events and decisions of the models are provided in Figure S1 and Appendix S2. carrying capacity at other life cycle stages (Table S2 in Appendix S1 and Appendix S3). In addition, since individuals with different breeding tactics (capital breeding and income breeding) have different energy budgets along the life cycle, we also tested the effects of breeding tactics on model outputs (Table S2 in Appendix S1 and Appendix S3). The average daily population density and the total number of individuals, individual stopover duration, individual energy reserves and energy accumulation rate during stopover were recorded at equilibrium and were examined by comparing the mean value of sim- ulation results. The model was run 10 times for each scenario and for 30 years in each simulation, which was sufficient to converge to stationary dynamics. Results were obtained from year 5 to year 30 of the simulation (Figure S2). The average daily population density was recorded at the S1 site; it was the mean of the abundance per day at the S1 site during the period of the stopover stage each year. The total number of individuals for the population and for the three types of habitats were recorded each year respectively. Individuals were recorded in age classes and reproduction status (juveniles, breeding adults and non-breeding adults). The per capita reproduction rate for adults, the reproduction rate among breeding adults and the survival rate were all calculated. 2.1.2 \| Model implementation To examine the impacts of parameter values on the model outputs and to test the robustness of the model results were to variation in parameter values, we conducted a local sensitivity analysis (see Methods and Results in Appendix S4). Our main goal is to examine the effects of habitat change in the stag- ing habitat on individual stopover duration in the course of northward migration and population dynamics across the entire life cycle and to identify the processes that connect them. To further examine whether our hypothesized processes only occur when the staging habitat be- came the stage with the lowest carrying capacity during the annual cycle, rather than the breeding or wintering stage, we tested impacts on individual stopover duration and population dynamics by reducing The per capita reproduction rate was the ratio of the number of juveniles and the number of all adults in the breeding stage; the reproduction rate among breeding adults was the ratio of the number of juveniles and the number of breeding adults; survival rate was split into two periods—one is the survival rate of the northward migration (the ratio between the number of adults at the breeding habitat and the number of individuals on the last day of wintering stage), and the other is the survival rate of the southward migration (the ratio between the num- ber of individuals at the wintering habitat and the number of individu- als at the breeding habitat). The energy reserves of each individual were assumed to be depen- dent on their initial energy, energy gained and energy expended. The ex- pected energy gained from food relied on both population density and food density (Goss-Custard et al., 2002). The stopover duration of a bird in the staging habitat was related to the energy requirement for migra- tion and the rate of energy acquisition (Hedenström & Alerstam, 1997). Parameter values in this model were consensus values, drawn from em- pirical studies of multiple migratory waterbirds (Table S1). These values were not generated from statistical analyses of individual data, due to the lack of available data. The calculation of individual energy reserves and stopover duration is provided in Appendix S2. Mean individual stopover duration in the S1 site was recorded each year. Individual departure energy reserves were recorded for three types of habitats respectively, as the energy reserves at the last day before the individual left the habitat range. Individual en- ergy reserves during the stopover period were recorded at each time step when the individual stayed in the staging habitat. The energy accumulation rate during stopover was calculated by dividing energy gained from the staging habitat by stopover duration. 2.2.1 \| Statistical analyses for bird number trends The distribution of Ti,y was normal distribution, so we fitted lin- ear model in program R to test the stopover duration trends across years for each common species, the regression equation was of the form: The abundance of all waterbirds and the abundance of the most common species were analysed in this study. Survey sites that were surveyed on less than 30% of survey dates were excluded from the analysis. To estimate total waterbird abundance, we summed the counts of all species i (including unidentified species) observed from all of the survey sites k in each day j, denoting it as Na. To identify the most common species, we used two methods, the ‘Frequency-based Method’ and the ‘Distribution-based Method’, to select 25 common species (S5). The number of these 25 species observed on each day j and survey site k was denoted as Nb. (3) Ti,y = 훼3 + 훽31 (y) + 훽32 (i) + 훽33 (y × i) + 휀3. (3) The correlation between stopover duration and bird abundance. The correlation between stopover duration and bird abundance. We fitted a regression between bird abundance and stopover duration to test the correlation between them. We calculated the mean abundance during the stopover period for each species i each year y, denoting it as Nc. Nc was used as the response variable, and y, i and Ti,y were used as explanatory variables. The distribution of Nc was an overdispersed Poisson, so we fitted GLM with a ‘quasi- Poisson’ error structure in program R. The regression equations were of the form: We calculated survey effort as the number of observers each day at each site as Ej,k. We transformed date to Julian date t, and calculated t2, because the temporal changes in bird abundance at staging habitat during the period of stopover is often quadratic (Thompson, 1993), and visual examination of our data also revealed a quadratic relation- ship. We treated ‘year’ (y) as a continuous variable in our models, and we treated ‘survey site’ (k) as a categorical variable. (4) Nc =exp(훼4 +훽41 (y) +훽42 (i) +훽43 (y ×i) +훽44 (Ti,y ) +훽45 (y ×Ti,y ) +훽46 (i×Ti,y ) +휀4). (4) Waterbird abundance (including Na, Nb) was used as response variables, and Ej,k, k, t, t2, i and y were used as explanatory variables. 2.2.1 \| Statistical analyses for bird number trends The distribution of Na and Nb was well-described as an overdis- persed Poisson (variance greater than mean), so we fitted GLM with a ‘quasi-Poisson’ error structure in program R. The regression equa- tions were of the form: We used the ANOVA command in R to assess the significance of each variable, used adjusted R2 to assess the goodness-of-fit of the linear model, and used 1 − (residual deviance/null deviance) to assess the goodness-of-fit of the GLMs. (1) Na =exp(훼1 +훽11 (y) +훽12 (Ej,k ) +훽13 (y ×Ej,k ) +훽14 (k) +훽15 (y ×k) +훽16 (t) +훽17 (t2) +훽18 (y ×t) +훽19 (y ×t2) +휀1), (1) ) 2.2 \| Empirical data Our empirical study was conducted in the wetlands in the north of Bohai Bay, between 38°36′-39°13′N and 117°11′-118.22′E, located Journal of Animal Ecology LIU et al. LIU et al. 987 in the northwest of the Yellow Sea (Figure S3). Despite habitat being lost in the surveyed area, the loss of favoured habitat was minimal, with the highest food density for migratory waterbirds compared to other staging sites in the Northern Yellow Sea (Peng et al., 2021; Wang et al., 2021; Yang et al., 2016). Our surveyed area is most equiv- alent to the S1 site in our IBMs. Survey data of migratory waterbirds were collected at boreal spring between 2004 and 2018 (details of the study area and data collection are provided in Appendix S5). All surveys were carried out under permits from Tianjin Municipal Bureau of Planning and Natural Resources and Luannan Forestry Bureau. This study did not require ethical approval. relationship between Julian date and bird abundance (Figure S4a), we first estimated normal distributions of bird abundance within the period of stopover for each species each year (Figure S4b). To do this, we extracted the mean date and variance in date for each species each year from the survey data, and scaled the curves by the bird number from the survey data. Then we esti- mated the date by which each quantile of the distribution of bird abundance is reached, date at 2.5% quantile and 97.5% quantile was the date of arrival at staging habitat and the date of depar- ture from staging habitat for each species each year within 95% confidential interval respectively (Figure S4c). We calculated the stopover duration based on the arrival date and departure date for each species each year, denoting it as Ti,y (details are provided in Appendix S5.4). 3.1 \| IBM: Population dynamics and individual stopover duration at the staging habitat Nb =exp (훼2 +훽21 (y) +훽22 (Ej,k ) +훽23 (y ×Ej,k ) +훽24 (k) +훽25 (y ×k) +훽26 (t) +훽27 (t2) +훽28 (y ×t) +훽29 (y ×t2) +훽210 (i) +훽211 (y ×i) +휀2) (2) As the size of the S2 site decreased through eight scenarios in the model 2 simulations, there was a decrease in the total number of individuals, and an increase in average daily population density at the S1 site (Figure 2a). The curve of daily population density at the S1 site became steeper in the early and late phases of the stopover period, and higher and wider when the population reached peak numbers (Figure 2b), showing that the population took less time to reach the peak number and remained at the peak number for a 2.2.2 \| Statistical analyses for stopover duration trends 3.3 \| IBM: Population structure and demography 3.3 longer period of time. Individual's annual stopover duration at the S1 site increased as the size of the S2 site decreased (Figure 2c). It suggests that as the habitat loss of the S2 site intensified, indi- viduals stayed longer in the S1 site where the habitat size remains constant. As the size of the S2 site decreased in model 2 simulations, the pro- portion of breeding adults and juveniles in the population decreased, while the proportion of non-breeding adults increased in each life cycle stage (Figure 4a). The per capita reproduction rate decreased because of an increased proportion of non-breeding adults, but the reproduction rate among breeding adults increased. The survival rate in both the northward migration and southward migration in- creased (Figure 4b). 3.2 \| IBM: Individual energy reserves As the size of the S2 site decreased in model 2 simulations, the indi- vidual energy accumulation rate when birds remained in the staging habitat decreased (Figure 3a), and the distribution shifted towards lower energy reserves with fewer individuals having reached the en- ergy threshold for departure from the staging habitat (Figure 3b). The energy reserves when individuals left the staging habitat were decreased; however, the departure energy reserves from the breed- ing habitat and wintering habitat were increased, both in adults and juveniles (Figure 3c). 2.2.2 \| Statistical analyses for stopover duration trends We estimated the stopover duration for each common species by using our survey data. Since our survey data revealed a quadratic Journal of Animal Ecology LIU et al. LIU et al. 988 3.4 \| Empirical data: Overall abundance The analysis of the abundance of all recorded species revealed an increase in the numbers of waterbirds at a rate of 61.85% per year (F1,809 = 13.60, p < 0.001) (Figure 5a). Even though effort affected abundance estimates (F1,808 = 11.04, p < 0.001) with more birds FI G U R E 2 Population dynamics and stopover duration as the size of the staging habitat decreased across the eight scenarios. (a) The relationship between average daily population density at the S1 site and the total number of individuals; (b) the curves of population density at the S1 site during the period of stopover; (c) individual annual stopover duration in relation to the size of staging habitat 20 30 40 50 60 90 120 150 180 210 total number of indivdiuals average population density at the S1 site (a) 0 25 50 75 100 50 100 150 200 date population density at the S1 site (b) 40 50 60 70 80 90 0% 18% 37% 50% 59% 68% 79% 100% the loss of S2 site individual stopover duration (day) the loss of S2 site 0% 18% 37% 50% 59% 68% 79% 100% (c) 20 30 40 50 60 90 120 150 180 210 total number of indivdiuals average population density at the S1 site (a) 20 30 40 50 60 90 120 150 180 210 total number of indivdiuals average population density at the S1 site (a) 40 50 60 70 80 90 0% 18% 37% 50% 59% 68% 79% 100% the loss of S2 site individual stopover duration (day) the loss of S2 site 0% 18% 37% 50% 59% 68% 79% 100% (c) 40 50 60 70 80 90 0% 18% 37% 50% 59% 68% 79% 100% the loss of S2 site individual stopover duration (day) the loss of S2 site 0% 18% 37% 50% 59% 68% 79% 100% (c) (a) (c) FI G U R E 2 Population dynamics and stopover duration as the size of the staging habitat decreased across the eight scenarios. (a) The relationship between average daily population density at the S1 site and the total number of individuals; (b) the curves of population density at the S1 site during the period of stopover; (c) individual annual stopover duration in relation to the size of staging habitat Journal of Animal Ecology 989 LIU et al. 3.4 \| Empirical data: Overall abundance FI G U R E 3 Individual energy reserves across the annual cycle. (a) The energy accumulating rate during the period of stopover when individuals were at the staging habitat. (b) The distribution of individual departure energy reserves from the staging habitat. (c) Individual departure energy reserves at each annual cycle stage 40 80 120 160 40 80 120 Julian date energy reserves (unit energy) the loss of S2 site 0% 18% 37% 50% 59% 68% 79% 100% (a) 0 50 100 150 0% 18% 37% 50% 59% 68% 79% 100% the loss of S2 site individual energy reserves (unit energy) the loss of S2 site 0% 18% 37% 50% 59% 68% 79% 100% (b) breeding staging wintering 0% 18% 37% 50% 59% 68% 79% 100% 50 75 100 125 150 50 75 100 125 150 50 75 100 125 150 the loss of S2 site individual energy reserves (unit energy) individual type adult juvenile stage of the annual cycle breeding staging wintering (c) 40 80 120 160 40 80 120 Julian date energy reserves (unit energy) the loss of S2 site 0% 18% 37% 50% 59% 68% 79% 100% (a) breeding staging wintering 0% 18% 37% 50% 59% 68% 79% 100% 50 75 100 125 150 50 75 100 125 150 50 75 100 125 150 the loss of S2 site individual energy reserves (unit energy) individual type adult juvenile stage of the annual cycle breeding staging wintering (c) 40 80 120 160 40 80 120 Julian date energy reserves (unit energy) (a) breeding staging wintering 0% 18% 37% 50% 59% 68% 79% 100% 50 75 100 125 150 50 75 100 125 150 50 75 100 125 150 the loss of S2 site individual energy reserves (unit energy) individual type adult juvenile stage of the annual cycle breeding staging wintering (c) the loss of S2 site 0 50 100 150 0% 18% 37% 50% 59% 68% 79% 100% the loss of S2 site individual energy reserves (unit energy) the loss of S2 site 0% 18% 37% 50% 59% 68% 79% 100% (b) FI G U R E 3 Individual energy reserves across the annual cycle. (a) The energy accumulating rate during the period of stopover when individuals were at the staging habitat. (b) The distribution of individual departure energy reserves from the staging habitat. 3.4 \| Empirical data: Overall abundance (c) Individual departure energy reserves at each annual cycle stage p = 0.009) (Figure 5c), after correcting for effects from observ- ers' effort (F1,18,417 = 47.63, p < 0.001) and the interaction be- tween year and effort (F1,18,372 = 6.94, p = 0.008), survey sites (F18,18,399 = 95.28, p < 0.001) and the interaction between year and survey sites (F18,18,354 = 5.96, p < 0.001) (Figure S9b), species (F24,18,373 = 80.14, p < 0.001) and the interaction between year and species (F24,18,328 = 6.96, p < 0.001). There was a quadratic association between Julian date and the number of birds (linear term, F1,18,398 = 13.21, p < 0.001; quadratic term F1,18,397 = 97.79, p < 0.001) (Figure 5d), but the interaction between year and Julian date was not statistically significant (linear term, F1,18,352 = 0.52, p = 0.470; quadratic term F1,18,353 = 2.57, p = 0.109). The GLM ex- plained 45.9% of the deviance. counted as effort increased, there was no interaction between year and effort (F1,787 = 0.094, p = 0.76). There was a positive relation- ship between waterbird abundance and Julian date (F1,789 = 5.98, p = 0.015). This slowed, and became negative with time, as the quadric term of date (F1,788 = 10.34, p = 0.0014) indicated a pa- rabola of waterbird abundance within the migration season. There was also an interaction between year and the quadric term of date (F1,768 = 17.96, p < 0.001), showing that the parabola shape within the migration season changed across years (Figure 5b). The average waterbird abundance (F18,790 = 30.81, p < 0.001) and the annual trend (F18,769 = 3.59, p < 0.001) differed between survey sites (Figure S9a). The GLM explained 51.1% of the deviance. 3.6 \| Empirical data: Stopover duration The analysis of the stopover duration of common species revealed an increasing trend at the rate of 1.39 days per year (F1,245 = 14.58, In the analysis of the most common species, the abundance of com- mon species increased at a rate of 35.13% per year (F1,18,418 = 6.92, LIU et al. Journal of Animal Ecology 990 FI G U R E 4 Population structure and demography as the size of the staging habitat decreased across the eight scenarios. (a) The proportion of individuals in different age classes and maturity status in three stages of the annual cycle. For illustration purposes, we only show three habitat scenarios here (0% loss of S2, 50% loss of S2 and 100% loss of S2) for ease of reading, the complete figure can be found in Figure S5. (b) Per capita rates of reproduction for all adults, reproduction for breeding adults, survival during southward migration and survival during northward migration 0% 50% 100% staging breeding wintering staging breeding wintering staging breeding wintering 0.00 0.25 0.50 0.75 1.00 stage of the annual cycle the proportion of each age class age classes juvenile breeding adult non−breeding adult (a) 0.4 0.8 1.2 0% 18% 37% 50% 59% 68% 79% 100% the loss of S2 site per capita rate per capita rate per capita reproduction rate reproduction rate for breeding adults survival rate during northward migration survival rate during southward migration (b) per capita rate w w w stage of the annual cycle the loss of S2 site FI G U R E 4 Population structure and demography as the size of the staging habitat decreased across the eight scenarios. (a) The proportion of individuals in different age classes and maturity status in three stages of the annual cycle. For illustration purposes, we only show three habitat scenarios here (0% loss of S2, 50% loss of S2 and 100% loss of S2) for ease of reading, the complete figure can be found in Figure S5. (b) Per capita rates of reproduction for all adults, reproduction for breeding adults, survival during southward migration and survival during northward migration plasticity in migration tactic via increased intraspecific competi- tion during migration stopovers. As the size of the staging area is reduced, individuals need to remain in the staging area for longer to obtain sufficient food to continue on their way due to an increase in the density of competitors. 3.6 \| Empirical data: Stopover duration Our model shows that the consequence of this is individuals depart later, often with poor condition, and fewer individuals make it to the breeding area. However, those that do make it fare well. The dynamics at the staging area can conse- quently have knock-on effects on populations in the overwintering and breeding areas that impact the population dynamics across the annual cycle, by altering the component of the life history where population dynamics are regulated. p < 0.001) (Figure 5e). The stopover duration (F24,221 = 7.07, p < 0.001) and its temporal trends (F24,197 = 1.87, p = 0.011) were different among species (Figure S9c). The adjust R-square of the lin- ear model was 0.423 (F49,197 = 4.67, p < 0.001). In the analysis of the correlation between bird abundance and the stopover duration, abundance was positively related to stopover duration (F1,220 = 18.26, p < 0.001) (Figure 5f), also pos- itively related to year (F1,245 = 56.80, p < 0.001). The abundance (F24,221 = 35.79, p < 0.001) was different across species. The re- lation between year and abundance (F24,196 = 3.27, p < 0.001) and the relation between stopover duration and abundance (F24,172 = 2.23, p = 0.002) were different across species. The inter- action between year and stopover duration was not statistically significant (F1,171 = 0.74, p = 0.390). The GLM explained 85.7% of the deviance. Habitat loss in the staging habitat reduces the carrying capacity of the flyway, leading to population declines along it. As a conse- quence, the part of the annual cycle that determines the carrying capacity has been switched from the breeding habitat and wintering habitat to the staging habitat, as the spatial extent of the staging habitat decreases. The staging habitat during northward migration becomes the stage where the strongest density dependence oper- ates, and the population becomes regulated by the carrying capac- ity of the staging habitat (Figure 6a). In contrast, competition in the breeding or wintering habitat is reduced, with processes operating in these areas no longer playing a major role in regulating the popu- lation dynamics. But any negative effects on the breeding or winter- ing grounds still can influence individual performance in the staging habitat through carry-over effects (Norris, 2005; Ryan Norris & Marra, 2007). The contributions of different life-history stages to 4 \| DISCUSSION By building a full-annual-cycle IBM of a stylized migratory popula- tion, we identify the critical role of stopover stage of northward mi- gration in the influence of migration tactics and population dynamics of migratory birds across the whole annual cycle, also identify the key processes linking individual migration tactic and population dy- namics (Figure 6). Our empirical data provides evidence to support the mechanisms shown from our theoretical model. Specifically, our results are consistent with the loss of staging habitat generating Journal of Animal Ecology 991 LIU et al. \| FI G U R E 5 Trends in waterbird abundance and stopover duration from the survey data. (a) Overall waterbird abundance as a function of year; (b) overall waterbird abundance as a function of Julian date in different years; (c) the abundance of the most common species as a function of year; (d) the abundance of the most common species as a function of Julian date in different years; (e) stopover duration of common species as a function of year; (f) the correlation between stopover duration and the abundance of common species FI G U R E 5 Trends in waterbird abundance and stopover duration from the survey data. (a) Overall waterbird abundance as a function of year; (b) overall waterbird abundance as a function of Julian date in different years; (c) the abundance of the most common species as a function of year; (d) the abundance of the most common species as a function of Julian date in different years; (e) stopover duration of common species as a function of year; (f) the correlation between stopover duration and the abundance of common species wintering or breeding areas (Figure 6b). How do we reach these conclusions? the population dynamics can consequently vary with the spatial ex- tent of the staging habitat. Such changes have the potential to alter selection on traits associated with competition for resources, and the entire life history. In our models, individuals adjust their migration tactic to prolong the stopover duration when food becomes scarce. Consequently, individuals often arrive at the breeding ground late and with low en- ergy reserves, or they fail to reach the breeding grounds. This results in fewer adults arriving in the breeding ground, fewer adults repro- ducing and a decline in total reproductive output. 4 \| DISCUSSION (b) The feedback process between individual stopover duration and population dynamics FI G U R E 6 The processes linking individual migration tactics and population dynamics. (a) The changes in carrying capacity and population density as a function of the size of the staging habitat. The circle sizes represent the carrying capacity of each habitat, with the grey rectangles representing the maximum population size across the annual cycle. In (a1), the size of the wintering and breeding habitats determine carrying capacity, while in (a2) the size of the staging habitat determines it. The line charts show how population density in the remaining staging site changes during the period of migration stopover for scenarios (a1) and (a2) respectively. (b) The feedback process between individual stopover duration and population dynamics reductions in total population-level reproductive output (Desprez et al., 2018; Newton, 2006). influence individuals' energy budgets across the annual cycle, in- dividual stopover duration increased for both capital breeders and income breeders when facing the loss of staging habitats (Figure S6). Therefore, whether the changes in the extent of the staging habi- tat alters the part of the life cycle where carrying capacity is lowest will determine the direction of change in life-history processes. Our proposed process can only occur when the staging habitat becomes the stage that determines the carrying capacity of the whole annual cycle. All of this is due to increased competition in the staging habitat, which reduces per capita food availability and generates behavioural plasticity in migratory tactics. As individual behaviour changes, so to do the population dynamics. The altered population dynamics, in turn, affect individual behaviour (Miner et al., 2005). This process continues until an equilibrium is reached, when the behaviour set- tles down to equilibrium, as do the population dynamics (Figure 6b). The connection between individual phenotypic plasticity and popu- lation dynamics is the result of feedback process across the annual cycle, which can result in eco-evolutionary feedbacks are argued by Coulson (2021). Our empirical and theoretical results align well, and are consis- tent, with patterns reported in the existing literature on the EAAF migration flyway, which are decreasing trends in total popula- tion size along the flyway, increasing trends in population density (Clemens et al., 2016; Piersma et al., 2016; Studds et al., 2017; Wilson et al., 2011; Yang et al., 2011) and increasing stopover duration at the remaining staging area (Conklin et al., 2021). 4 \| DISCUSSION However, those in- dividuals that do arrive experience less competition and have higher departure energy reserves at the breeding habitat, breeding adults have higher per capita reproductive rate. All individuals have higher survival rate during southward migration and northward migration. Previous studies have reported evidence that support parts of the processes we describe here including: longer stopover duration is related to habitat loss, scarce food or high density of competitors (Conklin et al., 2021; Kelly et al., 2002; Moore & Yong, 1991), and de- creased refuelling rate causes poor departure body mass in the stag- ing area (Baker et al., 2004), limited refuelling time reducing survival rates in the northward migration (Rakhimberdiev et al., 2018) and Both in our simulations and empirical analysis, population den- sity changes spatio-temporally following habitat loss in the staging habitat, with a positive correlation between bird numbers and stop- over duration (Figures 2b and 5b,d). This contrasts the ‘buffer effect’ hypothesis, which proposes that population density changes only in space (Brown, 1969; Gill et al., 2001; Sutherland, 1996a). Our re- sults show that in addition to birds becoming more concentrated in the remaining area, a decrease in the extent of the staging area can also result in an increase in time spent there, leading to intensified competition during the staging period. Therefore, during the time- limited northward migration, the reduced staging area not only leads to a higher population density in the staging habitat as individual birds stay longer, but also that the high population density is main- tained for a longer period during this part of the life cycle (Figure 6a). And this can alter the strength of population regulation in either the LIU et al. Journal of Animal Ecology 992 \| FI G U R E 6 The processes linking individual migration tactics and population dynamics. (a) The changes in carrying capacity and population density as a function of the size of the staging habitat. The circle sizes represent the carrying capacity of each habitat, with the grey rectangles representing the maximum population size across the annual cycle. In (a1), the size of the wintering and breeding habitats determine carrying capacity, while in (a2) the size of the staging habitat determines it. The line charts show how population density in the remaining staging site changes during the period of migration stopover for scenarios (a1) and (a2) respectively. REFERENCES Alerstam, T., Hedenström, A., & Åkesson, S. (2003). Long-distance migra- tion: Evolution and determinants. Oikos, 103, 247–260. https://doi. org/10.1034/j.1600-0706.2003.12559.x Alerstam, T., & Lindström, Å. (1990). Optimal bird migration: The relative importance of time, energy, and safety. In E. Gwinner (Ed.), Bird mi- gration (pp. 331–351). Springer. Amano, T., Székely, T., Koyama, K., Amano, H., & Sutherland, W. J. (2012). A framework for monitoring the status of populations: An exam- ple from wader populations in the East Asian–Australasian flyway. Biological Conservation, 145, 278–295. https://doi.org/10.1016/j. biocon.2010.06.010 Baker, A. J., González, P. M., Piersma, T., Niles, L. J., do Nascimento, I. d. L. S., Atkinson, P. W., Clark, N. A., Minton, C. D. T., Peck, M. K., & Aarts, G. (2004). Rapid population decline in red knots: Fitness con- sequences of decreased refuelling rates and late arrival in Delaware Bay. Proceedings of the Royal Society of London. Series B: Biological Sciences, 271, 875–882. https://doi.org/10.1098/rspb.2003.2663 Balbontín, J., Møller, A. P., Hermosell, I. G., Marzal, A., Reviriego, M., & De Lope, F. (2009). Individual responses in spring arrival date to ecological conditions during winter and migration in a migra- tory bird. Journal of Animal Ecology, 78, 981–989. https://doi. org/10.1111/j.1365-2656.2009.01573.x ORCID We call for further investigation on the connection between in- dividual timing, energy reserves and demography across the whole annual cycle. Migratory birds need to be studied in depth at their breeding, wintering and staging areas to allow us to fully understand their dynamics. Focusing on a single area in detail, with limited data elsewhere, makes dynamical inference challenging, and may also lead researchers to reach erroneous conclusions, such as: the mis- interpretation of increasing counts, and if the survey efforts were always put on the ‘best’ sites with increasing counts, this could lead to a situation where we keep seeing local increases in abundance right up until the point where species extinction is imminent, and lead us to miss the most opportune time for any conservation inter- vention. Migratory species play a key role for a number of other eco- logical processes by transporting energy and nutrients, or via their impacts on the ecological networks they form with other species in geographically separated areas (Bauer & Hoye, 2014), understanding their dynamics is also important to further extend our knowledge on their roles in ecosystem stability. Although logistically difficult, and costly, our work suggests that studies of migratory species across the entire annual cycle are necessary if we are going to understand the dynamics of species that exhibit one of the most remarkable be- haviours in the animal kingdom. ORCID Jin Liu https://orcid.org/0000-0001-7045-9827 Weipan Lei https://orcid.org/0000-0001-7616-7200 Zhengwang Zhang https://orcid.org/0000-0003-1063-7198 Tim Coulson https://orcid.org/0000-0001-9371-9003 Jin Liu https://orcid.org/0000-0001-7045-9827 Weipan Lei https://orcid.org/0000-0001-7616-7200 Zhengwang Zhang https://orcid.org/0000-0003-1063-7198 Tim Coulson https://orcid.org/0000-0001-9371-9003 CONFLICT OF INTEREST processes such as sex differences and a mixture distribution of migratory tactics. Males often migrate earlier than females in the northward migration to occupy the best breeding territories (Kokko et al., 2006; Newton, 2011), and some individuals might abort mi- gration without making any breeding attempt (Lisovski et al., 2016; Shaw & Levin, 2011). However, as the lack of individual-identified information on both timing, energy reserves and migrating trajec- tory, we chose to ignore them in this model. Nevertheless, individual differences in energy reserves persist throughout our model, and individuals with low energy reserves would not breed even they ar- rived at the breeding habitat. Therefore, individual differences might make our model more realistic, but are unlikely to change our main conclusions on the link between migratory tactics and population dynamics. None of the authors have a conflict of interest. DATA AVAILABILITY STATEMENT Code and data are available at Dryad Digital Repository https://doi. org/10.5061/dryad.gxd2547p6 (Liu et al., 2022). Code and data are available at Dryad Digital Repository https://doi. org/10.5061/dryad.gxd2547p6 (Liu et al., 2022). Code and data are available at Dryad Digital Repository https://doi. org/10.5061/dryad.gxd2547p6 (Liu et al., 2022). AUTHORS' CONTRIBUTIONS J.L., T.C. and Z.Z. conceived the ideas and designed the method- ology; J.L., W.L., X.M. and C.J.H. collected the data; J.L. built the model, analysed the data and led the writing of the manuscript with significant contributions from T.C. All authors contributed critically to the drafts and gave final approval for publication. 4 \| DISCUSSION Our study reveals the underlying mechanisms behind a seemingly positive phenomena ob- served in migratory birds: the increasing counts are likely driven by longer stopover duration, combined with refugees squeezed from other staging sites that have declined in quality, rather than an over- all population increase. Our study also highlights the crucial role of one life-history stage to population dynamics across the whole an- nual cycle. If the part of the annual cycle that determines carrying capacity is not operating in the staging habitat, individual tactics and popula- tion dynamics can show different patterns. The staging habitat was no longer the stage regulating population dynamics, and the stron- gest density-dependent effects at the breeding or wintering habi- tat can cause a decrease in reproduction or survival rates at either stage, leading to a lower population density at the staging habitat, and a shortened stopover duration (Figure S6). Evidence has been reported by Holmes et al. (1996) and Rockwell et al. (2017)), who observed reduced survival rates when the wintering habitat is lim- iting, and Marra and Holmes (2001), Rodenhouse et al. (2003) and Tomotani et al. (2018)), who reported poor physical condition of juveniles and decreased reproduction when the breeding habitat is limiting. In addition, even though different breeding tactics may When constructing our theoretical model, we focused on the general mechanism caused by the loss of staging habitat, there- fore we considered an average migratory tactic for all individuals, ignoring individual differences in migration behaviour driven by Journal of Animal Ecology LIU et al. LIU et al. 993 CONFLICT OF INTEREST None of the authors have a conflict of interest. ACKNOWLEDGEMENTS The authors thank Hongyan Yang, Pinjia Que, Yajing Chang, Bingrun Zhu, Jingsheng Ma, Jianmin Wang, Paul Holt, Adrian Boyle, the Global Flyway Network team and many volunteers for helping with the field- work. They also thank Jean-Michel Gaillard, Richard Fuller and an anon- ymous reviewer for their constructive feedback of our work. This work was supported by the grants from the Natural Science Foundation of China (31830089 and 31572288 to Z.Z., 31801985 to W.L.), the Paulson Institute grant and the SEE Foundation grant to Z.Z., the China Scholarship Council grant to J.L., the Fundamental Research Funds for the Central Universities to W.L. C.J.H. has been funded by GFN with grants secured by Theunis Piersma via BirdLife Netherlands (2007– 2012), WWF Netherlands (2010–2014, 2016, 2018–2021), Spinoza Premium of Netherlands Organization Prize for Scientific Research (2014–2017) and MAVA (foundation pour la nature) in 2018. Bauer, S., & Hoye, B. J. (2014). Migratory animals couple biodiversity and ecosystem functioning worldwide. Science, 344, 1242552. http:// doi.org/10.1126/science.1242552 Bi, X., Liu, F., & Pan, X. (2012). Coastal projects in China: From reclama- tion to restoration. Environmental Science & Technology, 46, 4691– 4692. https://doi.org/10.1021/es301286d 4692. https://doi.org/10.1021/es301286d Brown, J. L. (1969). The buffer effect and productivity in tit pop- ulations. The American Naturalist, 103, 347–354. https://doi. org/10.1086/282607 Clemens, R., Rogers, D. I., Hansen, B. D., Gosbell, K., Minton, C. D. T., Straw, P., Bamford, M., Woehler, E. J., Milton, D. A., Weston, M. A., Venables, B., Wellet, D., Hassell, C., Rutherford, B., Onton, K., Herrod, A., Studds, C. E., Choi, C.-Y., Dhanjal-Adams, K. L., … Fuller, R. A. (2016). Continental-scale decreases in shorebird populations in Australia. Emu - Austral Ornithology, 116(2), 119–135. https://doi. org/10.1071/MU15056 Journal of Animal Ecology 994 LIU et al. LIU et al. Conklin, J. R., Lisovski, S., & Battley, P. F. (2021). Advancement in long- distance bird migration through individual plasticity in departure. Nature Communications, 12, 1–9. https://doi.org/10.1038/s41467- 021-25022-7 Lisovski, S., Gosbell, K., Christie, M., Hoye, B. J., Klaassen, M., Stewart, I. D., Taysom, A. J., & Minton, C. (2016). Movement patterns of Sanderling (Calidris alba) in the East Asian-Australasian Flyway and a comparison of methods for identification of crucial areas for conservation. Emu-Austral Ornithology, 116, 168–177. https://doi. org/10.1071/MU15042 Coulson, T. (2021). Environmental perturbations and transitions between ecological and evolutionary equilibria: An eco-evolutionary feed- back framework. Peer Community Journal, 1. https://peercommun ityjournal.org/articles/10.24072/pcjournal.4 Liu, J., Lei, W., Mo, X., Hassell, C., Zhang, Z., & Coulson, T. tree.2005.08.002 Moore, F. R., & Yong, W. (1991). Evidence of food-based competition among passerine migrants during stopover. Behavioral Ecology and Sociobiology, 28, 85–90. https://doi.org/10.1007/BF001 80984 Gill, J. A., Norris, K., Potts, P. M., Gunnarsson, T. G., Atkinson, P. W., & Sutherland, W. J. (2001). The buffer effect and large-scale popula- tion regulation in migratory birds. Nature, 412, 436–438. https:// doi.org/10.1038/35086568 Morrision, R. I. G. (2006). Body transformations, condition, and survival in red knots Calidris canutus travelling to breed at alert, Ellesmere Island, Canada. Ardea, 94, 607–618. Goss-Custard, J. D., Stillman, R. A., West, A. D., Caldow, R. W. G., & McGrorty, S. (2002). Carrying capacity in overwintering migratory birds. Biological Conservation, 105, 27–41. https://doi.org/10.1016/ S0006-3207(01)00175-6 Murray, N. J., Clemens, R. S., Phinn, S. R., Possingham, H., & Fuller, R. A. (2014). Tracking the rapid loss of tidal wetlands in the Yellow Sea. Frontiers in Ecology and the Environment, 12, 267–272. https://doi. org/10.1890/130260 Harrison, X. A., Blount, J. D., Inger, R., Norris, D. R., & Bearhop, S. (2011). Carry-over effects as drivers of fitness differences in animals. Journal of Animal Ecology, 80, 4–18. https://doi. org/10.1111/j.1365-2656.2010.01740.x Newton, I. (2006). Can conditions experienced during migration limit the population levels of birds? Journal of Ornithology, 147, 146–166. https://doi.org/10.1007/s10336-006-0058-4 Hedenström, A. (2008). Adaptations to migration in birds: Behavioural strategies, morphology and scaling effects. Philosophical Transactions of the Royal Society B: Biological Sciences, 363, 287–299. https://doi.org/10.1098/rstb.2007.2140 Newton, I. (2007). The migration ecology of birds. Elsevier. ISBN: 978-0-12-517367-4. https://doi.org/10.1016/B978-0-12-51736 7-4.X5000-1 Hedenström, A., & Alerstam, T. (1997). Optimum fuel loads in migra- tory birds: Distinguishing between time and energy minimiza- tion. Journal of Theoretical Biology, 189, 227–234. https://doi. org/10.1006/jtbi.1997.0505 Newton, I. (2011). Migration within the annual cycle: Species, sex and age differences. Journal of Ornithology, 152, 169–185. https://doi. org/10.1007/s10336-011-0689-y Norris, D. R. (2005). Carry-over effects and habitat quality in mi- gratory populations. Oikos, 109, 178–186. https://doi. org/10.1111/j.0030-1299.2005.13671.x Holmes, R. T., Marra, P. P., & Sherry, T. W. (1996). Habitat-specific de- mography of breeding black-throated blue warblers (Dendroica caerulescens): Implications for population dynamics. Journal of Animal Ecology, 65, 183–195. https://doi.org/10.2307/5721 Norris, D. R., Marra, P. P., Kyser, T. K., Sherry, T. W., & Ratcliffe, L. M. (2004). Tropical winter habitat limits reproductive success on the temperate breeding grounds in a migratory bird. Proceedings of the Royal Society of London. Series B: Biological Sciences, 271, 59–64. https://doi.org/10.1098/rspb.2003.2569 Kelly, J. F., DeLay, L. S., & Finch, D. M. (2002). ACKNOWLEDGEMENTS (2022). Data from: Unravelling processes between phenotypic plasticity and population dynamics in migratory birds. Dryad Digital Repository, https://doi.org/10.5061/dryad.gxd2547p6 ityjournal.org/articles/10.24072/pcjournal.4 Coulson, T., Kendall, B. E., Barthold, J., Plard, F., Schindler, S., Ozgul, A., & Gaillard, J.-M. (2017). Modeling adaptive and nonadaptive responses of populations to environmental change. The American Naturalist, 190, 313–336. https://doi.org/10.1086/692542 https://doi.org/10.5061/dryad.gxd2547p6 Ma, Z., Melville, D. S., Liu, J., Chen, Y., Yang, H., Ren, W., Zhang, Z., Piersma, T., & Li, B. (2014). Rethinking China's new great wall. Science, 346, 912–914. https://doi.org/10.1126/science.1257258 Cox, G. W. (1968). The role of competition in the evolution of migration. Evolution, 22, 180–192. https://doi.org/10.1111/j.1558-5646.1968. tb03461.x Marra, P. P., Hobson, K. A., & Holmes, R. T. (1998). Linking winter and summer events in a migratory bird by using stable-carbon iso- topes. Science, 282, 1884–1886. https://doi.org/10.1126/scien ce.282.5395.1884 Desprez, M., Jenouvrier, S., Barbraud, C., Delord, K., & Weimerskirch, H. (2018). Linking oceanographic conditions, migratory schedules and foraging behaviour during the non-breeding season to repro- ductive performance in a long-lived seabird. Functional Ecology, 32, 2040–2053. https://doi.org/10.1111/1365-2435.13117 Marra, P. P., & Holmes, R. T. (2001). Consequences of dominance- mediated habitat segregation in American redstarts during the non- breeding season. The Auk, 118, 92–104. https://doi.org/10.1093/ auk/118.1.92 Dolman, P. M., & Sutherland, W. J. (1995). The response of bird popula- tions to habitat loss. Ibis, 137, S38–S46. https://doi.org/10.1111/ j.1474-919X.1995.tb08456.x Melville, D. S., Chen, Y., & Ma, Z. (2016). Shorebirds along the Yellow Sea coast of China face an uncertain future—A review of threats. Emu-Austral Ornithology, 116, 100–110. https://doi.org/10.1071/ MU15045 Erni, B., Liechti, F., & Bruderer, B. (2002). Stopover strategies in passer- ine bird migration: A simulation study. Journal of Theoretical Biology, 219, 479–493. https://doi.org/10.1006/jtbi.2002.3138 Miner, B. G., Sultan, S. E., Morgan, S. G., Padilla, D. K., & Relyea, R. A. (2005). Ecological consequences of phenotypic plasticity. Trends in Ecology and Evolution, 20, 685–692. https://doi.org/10.1016/j. tree.2005.08.002 Gienapp, P., Leimu, R., & Merilä, J. (2007). Responses to climate change in avian migration time—Microevolution versus phenotypic plas- ticity. Climate Research, 35, 25–35. https://doi.org/10.3354/ cr00712 tree.2005.08.002 Density-dependent mass gain by Wilson's warblers during stopover. The Auk, 119, 210–213. https://doi.org/10.1093/auk/119.1.210 Parrish, J. D. (2000). Behavioral, energetic, and conservation implica- tions of foraging plasticity during migration. Studies in Avian Biology, 20, 53–70. Kokko, H., Gunnarsson, T. G., Morrell, L. J., & Gill, J. A. (2006). Why do fe- male migratory birds arrive later than males? Journal of Animal Ecology, 75, 1293–1303. https://doi.org/10.1111/j.1365-2656.2006.01151.x Lack, D. (1968). Bird migration and natural selection. Oikos, 19, 1–9. https://doi.org/10.2307/3564725 Peng, H. B., Chan, Y. C., Compton, T. J., Cheng, X. F., Melville, D. S., Zhang, S. D., Zhang, Z., Lei, G., Ma, Z., & Piersma, T. (2021). Mollusc aqua- culture homogenizes intertidal soft-sediment communities along the 18,400 km long coastline of China. Diversity and Distributions, 27, 1553–1567. https://doi.org/10.1111/ddi.13302 Lindstrom, A., & Alerstam, T. (1992). Optimal fat loads in migrating birds: A test of the time-minimization hypothesis. The American Naturalist, 140, 477–491. https://doi.org/10.1086/285422 LIU et al. 995 Journal of Animal Ecology Piersma, T. (1988). Hop, skip, or jump? Constraints on migration of arctic waders by feeding, fattening, and flight speed. Wader Study Group Bulletin, 53, 6–8. Sutherland, W. J., & Crockford, N. J. (1993). Factors affecting the feed- ing distribution of red-breasted geese Branta ruficollis winter- ing in Romania. Biological Conservation, 63, 61–65. https://doi. org/10.1016/0006-3207(93)90074-B Piersma, T., & Drent, J. (2003). Phenotypic flexibility and the evolution of organismal design. Trends in Ecology & Evolution, 18, 228–233. https://doi.org/10.1016/S0169-5347(03)00036-3 Thompson, J. J. (1993). Modeling the local abundance of shorebirds staging on migration. Theoretical Population Biology, 44, 299–315. https://doi.org/10.1006/tpbi.1993.1030 https://doi.org/10.1016/S0169-5347(03)00036-3 Piersma, T., Lok, T., Chen, Y., Hassell, C. J., Yang, H. Y., Boyle, A., Slaymaker, M., Chan, Y. C., Melville, D. S., Zhang, Z. W., & Ma, Z. (2016). Simultaneous declines in summer survival of three shore- bird species signals a flyway at risk. Journal of Applied Ecology, 53, 479–490. https://doi.org/10.1111/1365-2664.12582 Tomotani, B. M., van der Jeugd, H., Gienapp, P., de la Hera, I., Pilzecker, J., Teichmann, C., & Visser, M. E. (2018). Climate change leads to differen- tial shifts in the timing of annual cycle stages in a migratory bird. Global Change Biology, 24, 823–835. https://doi.org/10.1111/gcb.14006 –490. https://doi.org/10.1111/1365-2664.12582 Pigliucci, M. (2001). Phenotypic plasticity: Beyond nature and nurture. The Johns Hopkins University Press. Verkuil, Y. I., Karlionova, N., Rakhimberdiev, E. N., Jukema, J., Wijmenga, J. J., Hooijmeijer, J. C., Pinchuk, P., Wymenga, E., Baker, A. J., & Piersma, T. (2012). tree.2005.08.002 Losing a staging area: Eastward redistribution of Afro-Eurasian ruffs is associated with deteriorating fuelling condi- tions along the western flyway. Biological Conservation, 149, 51–59. https://doi.org/10.1016/j.biocon.2012.01.059 Rakhimberdiev, E., Karagicheva, J., Camphuysen, C. J., Castricum, V. R. Rakhimberdiev, E., Karagicheva, J., Camphuysen, C. J., Castricum, V. R. S., Dekinga, A., Dekker, R., Gavrilov, A., ten Horn, J., Jukema, J., Saveliev, A., Soloviev, M., Tibbitts, T. L., van Gils, J. A., & Piersma, T. (2018). Fuelling conditions at staging sites can mitigate Arctic warming effects in a migratory bird. Nature Communications, 9, 4263. https://doi.org/10.1038/s41467-018-06673-5 Wang, X., Xiao, X., Xu, X., Zou, Z., Chen, B., Qin, Y., Zhang, X., Dong, J., Liu, D., Pan, L., & Li, B. (2021). Rebound in China's coastal wetlands fol- lowing conservation and restoration. Nature Sustainability, 4, 1076– 1083. https://doi.org/10.1038/s41893-021-00793-5 Ratikainen, I. I., Gill, J. A., Gunnarsson, T. G., Sutherland, W. J., & Kokko, H. (2008). When density dependence is not instantaneous: Theoretical developments and management implications. Ecology Letters, 11, 184–198. https://doi.org/10.1111/j.1461-0248.2007.01122.x Wilson, H. B., Kendall, B. E., Fuller, R. A., Milton, D. A., & Possingham, H. P. (2011). Analyzing variability and the rate of decline of migratory shorebirds in Moreton Bay, Australia. Conservation Biology, 25, 758– 766. https://doi.org/10.1111/j.1523-1739.2011.01670.x Rockwell, S. M., Wunderle, J. M., Sillett, T. S., Bocetti, C. I., Ewert, D. N., Currie, D., White, J. D., & Marra, P. P. (2017). Seasonal sur- vival estimation for a long-distance migratory bird and the influ- ence of winter precipitation. Oecologia, 183, 715–726. https://doi. org/10.1007/s00442-016-3788-x Winger, B. M., Auteri, G. G., Pegan, T. M., & Weeks, B. C. (2019). A long win- ter for the Red Queen: Rethinking the evolution of seasonal migration. Biological Reviews, 94, 737–752. https://doi.org/10.1111/brv.12476 Rodenhouse, N. L., Scott Sillett, T., Doran, P. J., & Holmes, R. T. (2003). Multiple density–dependence mechanisms regulate a migratory bird population during the breeding season. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270, 2105–2110. https://doi.org/10.1098/rspb.2003.2438 Winkler, D. W., Jørgensen, C., Both, C., Houston, A. I., McNamara, J. M., Levey, D. J., Partecke, J., Fudickar, A., Kacelnik, A., Roshier, D., & Piersma, T. (2014). Cues, strategies, and outcomes: How migrating vertebrates track environmental change. Movement Ecology, 2, 10. https://doi.org/10.1186/2051-3933-2-10 Yang, H.-Y., Chen, B., Barter, M., Piersma, T., Zhou, C.-F., Li, F.-S., & Zhang, Z.-W. (2011). Impacts of tidal land reclamation in Bohai Bay, China: Ongoing losses of critical Yellow Sea waterbird staging and winter- ing sites. tree.2005.08.002 Bird Conservation International, 21, 241–259. https://doi. org/10.1017/S0959270911000086 Ryan Norris, D., & Marra, P. P. (2007). Seasonal interactions, habitat qual- ity, and population dynamics in migratory birds. The Condor, 109, 535–547. https://doi.org/10.1093/condor/109.3.535 Shaw, A. K., & Levin, S. A. (2011). To breed or not to breed: A model of partial migration. Oikos, 120, 1871–1879. https://doi. org/10.1111/j.1600-0706.2011.19443.x Yang, H.-Y., Chen, B., Piersma, T., Zhang, Z., & Ding, C. (2016). Molluscs of an intertidal soft-sediment area in China: Does overfishing ex- plain a high density but low diversity community that benefits staging shorebirds? Journal of Sea Research, 109, 20–28. https://doi. org/10.1016/j.seares.2016.01.006 Somveille, M., Rodrigues, A. S., & Manica, A. (2018). Energy efficiency drives the global seasonal distribution of birds. Nature Ecology & Evolution, 2, 962–969. https://doi.org/10.1038/s41559-018-0556-9 Studds, C. E., Kendall, B. E., Murray, N. J., Wilson, H. B., Rogers, D. I., Clemens, R. S., Gosbell, K., Hassell, C. J., Jessop, R., Melville, D. S., Milton, D. A., Minton, C. D. T., Possingham, H. P., Riegen, A. C., Straw, P., Woehler, E. J., & Fuller, R. A. (2017). Rapid population decline in migratory shore- birds relying on Yellow Sea tidal mudflats as stopover sites. Nature Communications, 8, 14895. https://doi.org/10.1038/ncomms14895 SUPPORTING INFORMATION Additional supporting information may be found in the online version of the article at the publisher’s website. Studds, C. E., & Marra, P. P. (2005). Nonbreeding habitat occupancy and population processes: An upgrade experiment with a migratory bird. Ecology, 86, 2380–2385. https://doi.org/10.1890/04-1145 How to cite this article: Liu, J., Lei, W., Mo, X., Hassell, C. J., Zhang, Z. & Coulson, T. (2022). Unravelling the processes between phenotypic plasticity and population dynamics in migratory birds. Journal of Animal Ecology, 91, 983–995. https:// doi.org/10.1111/1365-2656.13686 Sutherland, W. J. (1996a). From individual behaviour to population ecology (R. M. May & P. H. Harvey [Eds.]). Oxford Series in Ecology and Evolution. Oxford University Press. Sutherland, W. J. (1996b). Predicting the consequences of habitat loss for migratory populations. Proceedings of the Royal Society of London. Series B: Biological Sciences, 263, 1325–1327. https://doi. org/10.1098/rspb.1996.0194

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.